# fearofgod.ai

> Private and niche data for those who want to simulate the world.

fearofgod.ai licenses private and niche data from companies, securely, so they generate additional, long-term revenue from data they already hold. It works with companies, not individuals, in any industry, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences. fearofgod.ai has paid $17M to partners and has 60+ partner companies. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner. A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. A partner company can keep ownership of its data. This is agreed case by case. The data goes to AI labs and teams building world models, robotics and simulation; fearofgod.ai never names its buyers.

## Facts about fearofgod.ai

fearofgod.ai licenses private and niche data from companies, securely, so they generate additional, long-term revenue from data they already hold. This page states the facts about fearofgod.ai in one place. They were last checked on Oct 25, 2026.

### What is fearofgod.ai?

fearofgod.ai: Private and niche data for those who want to simulate the world.

fearofgod.ai licenses the data companies already hold, securely, so they generate additional, long-term revenue from it. The data goes to AI labs and teams building world models, robotics and simulation, who are never named.

### Who does fearofgod.ai work with?

fearofgod.ai works with companies, not individuals. Partner companies are based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan. fearofgod.ai works with companies in any industry.

### What data does fearofgod.ai license?

Company data from any industry. The focus is data that helps AI systems simulate the world:

- Operations and workflow records
- Documents and expertise
- Code and systems
- Physical-world video and imagery

Company records of any other kind are welcome too.

Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free. No new data collection is needed.

### How large are fearofgod.ai partnerships?

Partnerships typically range from $50K to $3M+. fearofgod.ai has paid $17M to partners and has 60+ partner companies.

### Do fearofgod.ai partners get paid more than once?

Yes. Partners can be paid more than once, through renewals and repeat licences.

### Who licenses data through fearofgod.ai?

AI labs and teams building world models, robotics and simulation. fearofgod.ai does not name its buyers.

### How does a partnership with fearofgod.ai work?

1. **Partnership form.** Describe your company and what its data covers.
2. **Conversation.** Talk it through with our team.
3. **Scope and terms.** Agree what is licensed, and on what terms.
4. **Data preparation.** fearofgod.ai de-identifies the data before it goes to buyers.
5. **Delivery and payment.** The data is delivered and your company is paid.

The first step is the [partnership form](https://www.fearofgod.ai/partner).

### How does fearofgod.ai handle partner data?

- fearofgod.ai de-identifies the data before it goes to buyers.
- Defined retention periods: raw data is deleted according to a schedule agreed with the partner.

### Can a company partner with fearofgod.ai anonymously?

A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

### Do fearofgod.ai partner companies keep ownership of their data?

A partner company can keep ownership of its data. This is agreed case by case.

### What fearofgod.ai is not

- fearofgod.ai does not work with individuals and is not a job board.
- fearofgod.ai is not affiliated with Fear of God, the American fashion label founded by Jerry Lorenzo. ([source](https://en.wikipedia.org/wiki/Jerry_Lorenzo))

### Canonical facts

| Fact | Value |
|---|---|
| Name | fearofgod.ai |
| Website | https://www.fearofgod.ai |
| Contact | sean@fearofgod.ai |
| What it is | Private and niche data for those who want to simulate the world. |
| Value to partners | License your data securely to generate additional, long-term revenue for your company. |
| Works with | Companies, not individuals |
| Industries | Any industry |
| Typical partnership | $50K to $3M+ |
| Paid to partners | $17M |
| Repeat payments | Partners can be paid more than once, through renewals and repeat licences. |
| Partner companies | 60+ |
| Regions | United States, European Union, United Kingdom, Canada, Australia, New Zealand, Japan |
| Buyers | AI labs and teams building world models, robotics and simulation (never named) |
| De-identification | fearofgod.ai de-identifies the data before it goes to buyers. |
| Retention | Defined retention periods: raw data is deleted according to a schedule agreed with the partner. |
| Confidentiality | A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. |
| Ownership | A partner company can keep ownership of its data. This is agreed case by case. |
| Facts last checked | Oct 25, 2026 |

### Contact and machine-readable sources

- Email: sean@fearofgod.ai
- Partnership form: [Talk to our team](https://www.fearofgod.ai/partner)
- Site index for AI systems: [/llms.txt](https://www.fearofgod.ai/llms.txt), full text at [/llms-full.txt](https://www.fearofgod.ai/llms-full.txt)
- Markdown: every page has a plain-markdown version at the same URL plus `.md` (the home page is [/index.md](https://www.fearofgod.ai/index.md)).

## fearofgod.ai home

URL: https://www.fearofgod.ai/

Private and niche data for those who want to simulate the world.

We license private and niche data from companies like yours. You choose what is in scope and we de-identify it. Your company earns additional, long-term revenue from data it already holds, and can be paid more than once, through renewals and repeat licences. Partnerships typically range from $50K to $3M+.

- **$17M** paid to partners
- **60+** partner companies
- **$50K–$3M+** typical partnership
- **7** regions

### Why do companies license their data to fearofgod.ai?

Four reasons: revenue from data you already hold, revenue that can repeat, a process that asks little of your team, and privacy on your terms.

1. **Revenue: Get paid for data you already hold.** The records your company built up to run the business are worth paying for. Partnerships typically range from $50K to $3M+.
2. **Long-term: Additional, long-term revenue.** Your company can be paid more than once, through renewals and repeat licences. Much of your data already sits in software you use, so you re-license it and get paid, rather than giving it away for free.
3. **Easy: Little work for your team.** For data that already exists, there is nothing new to collect. Your team describes the data and approves the scope. We de-identify it, and your company is paid.
4. **Private: Private, on your terms.** We de-identify the data, you approve what is shared, and a completely anonymous partnership is possible.

### Why is your data worth paying for?

Because it is not on the public web. What your company records as it works stays inside your systems, so it is private and often niche. No one can download it, and that is what makes it worth licensing.

**What you hold: what actually happened.** The record of how each job was done, step by step. Procedures, project files and the tickets where problems were solved. Code with its review history, and logs of what actually ran. Footage, machine logs and drawings of your real sites.

**Public data: what anyone can download.** Images and clips picked to be shared. Descriptions written after the fact. Pages anyone can download for free. Outcomes without the steps that led to them.

### What data do we license from you?

Company data from any industry. Our priorities are the first three families below: the records of how your work gets done, what your teams know, and the code and systems behind it. We also license physical-world video, sensor and 3D data, and welcome company records of any other kind.

**Already in your software.** Much of your data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free.

1. **Operations and workflow records.** The record of how work gets done: what was ordered, scheduled, dispatched, inspected, repaired and signed off, in sequence. Examples: work orders and maintenance logs, inspection and quality reports, scheduling and dispatch, orders, inventory and transactions. Where it usually lives: ERP systems, maintenance and field-service apps, dispatch and scheduling tools, quality systems.
2. **Documents and expertise.** Written know-how: procedures, manuals, project files, recorded training and the tickets and chats where problems were actually solved. Examples: manuals and procedures, project files and reports, recorded training sessions, support tickets and chats. Where it usually lives: document management, shared drives, wikis and intranets, ticketing and chat.
3. **Code and systems.** Software and the systems around it: repositories with their review history, test and simulation results, and the logs that show what ran. Examples: code repositories and review history, simulation and test results, system and application logs. Where it usually lives: git hosting, CI and test systems, log platforms, simulation outputs.
4. **Physical-world video and imagery.** Footage of real sites, roads, shop floors and job sites, recorded as the work happened. It shows how places, machines and people actually move. Examples: helmet and body-camera video, dashcam and fleet video, drone and aerial footage, CCTV and site cameras, inspection imagery. Where it usually lives: video management systems, fleet camera platforms, drone and survey archives, cloud storage.
5. **Sensor, telemetry and robotics logs.** Readings from machines, vehicles and robots over time, including the faults, stoppages and odd days that never make it into a brochure. Examples: machine and IoT telemetry, PLC and SCADA logs, GPS and fleet traces, robot and teleoperation logs, environmental readings. Where it usually lives: telematics platforms, historians and SCADA, IoT dashboards, robot logs on disk.
6. **3D, CAD and spatial data.** The geometry of things that were designed or built: scans, point clouds, drawings and models, often with their revision history attached. Examples: LiDAR and point clouds, 3D scans and photogrammetry, CAD and BIM models, maps and geospatial layers. Where it usually lives: CAD and PLM vaults, BIM platforms, GIS servers, scan archives.

### How simple is it to license your data?

Five steps, from a short form to payment.

1. **Partnership form.** Describe your company and what its data covers.
2. **Conversation.** Talk it through with our team.
3. **Scope and terms.** Agree what is licensed, and on what terms.
4. **Data preparation.** We de-identify the data before it goes to buyers.
5. **Delivery and payment.** The data is delivered and your company is paid.

**Your team:** describes the company and its data on the partnership form; talks it through with our team; agrees the scope and terms, and approves them before anything is shared.

**Our team:** de-identifies the data; deletes raw data on the schedule agreed with you; pays your company.

### How is your data kept private?

You set the scope and approve it before anything is shared. We de-identify what is licensed, keep your company's name confidential until the agreement is fully signed, and delete raw data on the schedule agreed with you.

- **De-identification.** We de-identify the data before it goes to buyers.
- **Confidentiality and anonymity.** Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. We never name our buyers.
- **Approved scope.** Only the data named in the scope you agree is licensed. Whatever you leave out stays out.
- **Your approval.** You approve the scope and the terms before anything is shared.
- **Defined retention periods.** Raw data is deleted according to a schedule agreed with you.

### Which industries do we license data from?

Any industry. We work with companies wherever useful records have built up, as in these examples. If your company holds records like these, they can earn revenue.

Software and IT · Financial services and insurance · Professional services · Healthcare and life sciences · Retail and e-commerce · Real estate and property · Media, games and entertainment · Education and training · Telecommunications · Manufacturing and industrials · Construction and engineering · Logistics, transport and fleets · Energy, utilities and mining · Automotive and mobility · Aerospace and defence · Agriculture and food · Robotics and automation · Drones, survey and inspection · Security and facilities · Sports and fitness

### Which regions do we work in?

We work with companies, not individuals, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

### What is fearofgod.ai?

fearofgod.ai licenses private and niche data from companies, securely, so they generate additional, long-term revenue from data they already hold. fearofgod.ai works with companies, not individuals.

- **What it is:** Private and niche data for those who want to simulate the world.
- **Value to you:** License your data securely to generate additional, long-term revenue for your company.
- **Typical partnership:** $50K to $3M+
- **Paid to partners:** $17M
- **Partner companies:** 60+
- **Works with:** Companies, not individuals
- **Regions:** United States, European Union, United Kingdom, Canada, Australia, New Zealand, Japan
- **Data:** Any industry: operations and workflow records; documents and expertise; code and systems; physical-world video and imagery, and more
- **De-identification:** We de-identify the data before it goes to buyers.
- **Ownership:** Your company can keep ownership of its data. This is agreed case by case.
- **Defined retention periods:** Raw data is deleted according to a schedule agreed with you.
- **Buyers:** AI labs and teams building world models, robotics and simulation. Never named.

### Frequently asked questions

#### How much can our data earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. fearofgod.ai has paid $17M to partners and works with 60+ partner companies. What a partnership is worth depends on the data, its rights and the agreed scope.

#### Is this a one-off payment or long-term revenue?

It can be long-term: your company can be paid more than once, through renewals and repeat licences. You license data your company already holds, so it earns additional revenue instead of being given away for free. Partnerships typically range from $50K to $3M+.

#### How much work is it for our team?

Little. For data that already exists, there is nothing new to collect. Your team describes the data on the [partnership form](https://www.fearofgod.ai/partner), talks it through with us, and agrees and approves the scope and terms. We de-identify the data, and your company is paid.

#### How is our data kept private and de-identified?

We de-identify the data before it goes to buyers. Only the scope you approve is licensed, and anything you leave out stays out. Raw data is deleted according to a schedule agreed with you.

#### Will our company be named?

Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

#### Do we keep ownership of our data?

Your company can keep ownership of its data. This is agreed case by case.

#### What happens to our raw data?

Raw data is deleted according to a schedule agreed with you. The schedule is one of the defined retention periods agreed for each partnership.

#### Do we approve what is shared?

Yes. You approve the scope and the terms before anything is shared, and nothing outside the agreed scope is licensed.

#### Which companies qualify, and where?

Companies, not individuals, in any industry, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan. If your company is based elsewhere, choose Other on the [partnership form](https://www.fearofgod.ai/partner) and we will say whether a partnership is possible.

#### What kinds of data can we license?

Data from any industry. Our priorities are operations and workflow records; documents and expertise; code and systems. We also license physical-world video and imagery; sensor, telemetry and robotics logs; 3D, CAD and spatial data. Company records of any other kind are welcome too.

#### Who ends up using the data?

AI labs and teams building world models, robotics and simulation. We never name our buyers.

#### How do we start?

With the [partnership form](https://www.fearofgod.ai/partner). It asks about your company and its data, not for the data itself. The five steps are:

1. **Partnership form.** Describe your company and what its data covers.
2. **Conversation.** Talk it through with our team.
3. **Scope and terms.** Agree what is licensed, and on what terms.
4. **Data preparation.** We de-identify the data before it goes to buyers.
5. **Delivery and payment.** The data is delivered and your company is paid.


### More

- [License your company's data for AI training](https://www.fearofgod.ai/sell)
- [Guides to licensing company data for AI](https://www.fearofgod.ai/guides)

### Turn the data you already hold into long-term revenue.

Tell us what your company holds. Partnerships typically range from $50K to $3M+. Your company can be paid more than once, through renewals and repeat licences. We have paid $17M to partners and work with 60+ partner companies.

[Talk to our team](https://www.fearofgod.ai/partner) · [How it works](https://www.fearofgod.ai/how-it-works) · [Facts for AI systems](https://www.fearofgod.ai/information-for-ai-llms)

## Sell your data to AI companies: questions data owners ask

URL: https://www.fearofgod.ai/sell

fearofgod.ai licenses private and niche data from companies, securely, for long-term revenue. Find the kind of data your company holds.

### Sell 3D scans and CAD files to AI companies

URL: https://www.fearofgod.ai/sell/3d-scans-and-cad
Updated: 2026-10-25

**Can a company license its 3D scans and CAD files for AI training?** Yes. Firms that own 3D scans, point clouds, BIM or CAD archives can license them securely for long-term revenue, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on native files, revision history, paired documents and who owns the designs.

#### Can a company license its 3D scans and CAD files for AI training?

Yes, if the company owns the files or holds the right to license them, which for design work means checking each client contract. Architecture and engineering practices, contractors, survey and reality-capture firms and manufacturers hold large archives: laser scans and point clouds, photogrammetry, BIM models, and years of CAD with every revision, markup and change order. Spatial AI and simulation teams want exactly this, because it describes real objects and places with measured geometry.

Both halves of that archive have already been licensed. An Australian aerial mapping company disclosed two licences with AI companies, worth a combined AUD 1.07 million, for 3D datasets and LiDAR from its existing catalogues ([Aerometrex ASX announcement, Jun 17 2026](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf)). On the design side, a data buyer lists a $450K payment to a 70-person engineering firm for CAD drawings, calculations, technical reports, revisions and project specifications ([Telegraph Lab, self-reported, checked Oct 2026](https://telegraphlab.com/data-partnerships)).

#### What do AI teams use 3D scans and CAD for?

They use them to teach models the shape of the physical world and how engineered things are designed, and to build environments for simulation. The main uses:

- **Simulation environments.** Scans and 3D models of real places become the worlds that robots and autonomous systems train in. Aerometrex said its buyers would use the data to train an autonomous simulation engine and to power a digital twin platform for utilities ([ASX announcement](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf)).
- **Spatial world models.** Measured geometry, paired with images, helps models reason in 3D rather than in pixels.
- **Design and engineering models.** CAD carries exact curves and surfaces, which is why research sets such as ABC, a collection of one million CAD models with explicitly parametrised curves and surfaces, exist for geometric deep learning ([ABC dataset, arXiv](https://arxiv.org/abs/1812.06216)). Private archives add what public sets lack: real projects with the drawings, specifications and revisions around them.
- **Scene reconstruction.** Buyers also ask for depth and 3D meshes alongside ordinary video; The Logic reported one such request for footage of homes and offices with depth information and object meshes ([The Logic, May 25 2026](https://thelogic.co/news/ai-training-data-video-versos/)).

#### What makes 3D scans and CAD more valuable?

Files are worth more when they keep their native structure, their history and their context. In rough order of impact:

1. **Native formats.** Parametric CAD with feature history, and BIM with object classes, beat exported meshes or PDFs.
2. **Revision history.** Successive versions, markups, RFIs and change orders show how a design evolved and why, which a single final file cannot.
3. **Paired records.** Drawings linked to specifications, calculations, simulations, quality reports and the as-built scan of what was actually made or built.
4. **Accuracy metadata.** Scanner or camera model, registration reports, coordinate system and georeferencing, units and stated tolerances.
5. **Raw captures.** Source photos behind photogrammetry and unfiltered point clouds, not only the cleaned model.
6. **Breadth and repetition.** Many building or product types, and the same site scanned at several stages.

#### What rights and consent do 3D scans and CAD need?

Ownership is the main question for design files, and export control and confidentiality can rule some files out entirely. The gates that come up most:

- **Who owns the designs.** Under US copyright law, a commissioned work counts as a work made for hire only if it falls within nine listed categories and both parties signed a written agreement saying so ([US Copyright Office, Circular 30](https://www.copyright.gov/circs/circ30.pdf)). Most design drawings do not fit those categories, so ownership depends on what each contract assigned or licensed. Contracts may also restrict reuse of the client's project information.
- **Defence and controlled technology.** Under ITAR, "technical data" includes blueprints, drawings and plans required to design, build or maintain defence articles ([22 CFR 120.33](https://www.ecfr.gov/current/title-22/chapter-I/subchapter-M/part-120/subpart-C/section-120.33)). Under the EAR, releasing controlled technology to a foreign person, even inside the US, counts as an export ([15 CFR 734.13](https://www.ecfr.gov/current/title-15/subtitle-B/chapter-VII/subchapter-C/part-734/section-734.13)). Files touching either regime need an export-control review, and some may need to be excluded.
- **Client confidentiality.** NDAs, security clauses for sensitive buildings, and manufacturing clients' trade secrets.
- **People and property in scans.** Interior and street scans can capture people, documents on desks and plates. In the EU and UK identifiable people in imagery are personal data ([EDPB Guidelines 3/2019](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-32019-processing-personal-data-through-video_en)).

This page is general information, not legal advice.

#### What are 3D scans and CAD files worth?

Disclosed prices are scarce and come as whole-archive deals rather than per-file rates. Each figure is dated and labelled by market side; none is a fearofgod.ai price.

| Reference point | Market side | Figure and source |
|---|---|---|
| Two licences: 3D datasets and LiDAR from existing catalogues | Disclosed licence, Jun 2026 | [AUD 1.07 million combined (Aerometrex, ASX)](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf) |
| 70-person engineering firm: CAD drawings, calculations, reports, revisions, specifications | Buyer's self-reported past deal, checked Oct 2026 | [$450K paid to the business (Telegraph Lab)](https://telegraphlab.com/data-partnerships) |

The pattern in both disclosed deals is that the archive mattered, not single files: catalogues of scans in one case, and a whole firm's drawings and their revisions in the other. Partnerships typically range from $50K to $3M+.

#### How does licensing 3D scans and CAD work with fearofgod.ai?

You license scans and drawings your company already holds, securely, and earn additional, long-term revenue from them. The archive already sits in a document-management system, a CAD vault, a cloud drive or a scan server, so licensing it means re-licensing what exists. No new data collection is needed.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and anything you want left out, such as a client's projects or export-controlled files, stays out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

A useful first description covers file counts and formats, years covered, project types, scanners or software used, and how client contracts treat ownership. [Talk to our team](https://www.fearofgod.ai/partner) when you are ready. Survey firms with aerial capture as well can read [licensing drone footage](https://www.fearofgod.ai/sell/drone-and-aerial-footage).

#### Selling 3D scans and CAD files to AI companies

A firm that owns its 3D scans, point clouds and CAD files can sell them to AI companies by licensing them with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences.

#### Frequently asked questions

##### Can we sell our 3D scans and CAD files to AI companies?

Yes, if your company owns the files or holds the right to license them. You sell them by licensing them with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case. Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

##### We are an engineering firm with 20 years of CAD. Can we license it?

Possibly, project by project. The deciding question is whether your contracts left you owning the drawings or free to reuse them. Projects where the client took ownership, or that involve controlled technology, may need to be excluded.

##### Is CAD revision history worth more than final drawings?

It often adds value. Successive versions, markups and change orders show how a design was reasoned through, which a final drawing alone cannot. Keep version history and the records that explain each change.

##### Can we license point clouds of client buildings?

Only if your contract allows reuse and the scans can be handled for people, documents and sensitive areas. Security-sensitive sites may need to be excluded.

##### Do export controls affect licensing CAD to AI companies?

They can. ITAR technical data includes drawings for defence articles, and under the EAR releasing controlled technology to a foreign person counts as an export. Have files that touch either regime reviewed before scope is agreed. This is general information, not legal advice.

##### How much can our 3D and CAD archive earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on formats, revision history, paired records, accuracy metadata and rights. fearofgod.ai does not publish per-file prices.

##### Does fearofgod.ai buy 3D models from individual designers?

No. fearofgod.ai works with companies, not individuals. A firm that owns its archive can apply.

##### Which countries does fearofgod.ai work in?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

### Sell company records and workflow data to AI companies

URL: https://www.fearofgod.ai/sell/company-records-and-workflows
Updated: 2026-10-25

**Can a company license its Slack, email, documents and tickets for AI training?** Yes. A company can license the records its work already produces, such as chat, email, documents, tickets and code, securely and for long-term revenue, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on history, linked systems and handling of personal data.

#### Can a company license its records and workflow data for AI training?

Yes, if the company owns the records and can handle the personal, client and vendor information inside them. Chat threads, email, shared documents, support tickets, CRM history and code repositories record how a business actually gets work done, and teams building AI agents want that record. Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free.

The market is real and noisy. Business Insider reported owners receiving six-figure "exploding offers" that expire in 48 hours, and Gartner's estimate that spending on data for AI will grow from under $1 billion to almost $6.5 billion in 2027 ([Business Insider, Oct 13 2026](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)). An offer with a deadline that short is a reason to slow down, not speed up.

#### What do AI teams use company records for?

They use them to teach agents how real work unfolds across people, tools and time, rather than one prompt at a time. Business Insider quoted an executive at an AI training-data startup saying AI developers increasingly want "long-horizon" data that gives a full picture of how work gets done ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)). The useful part is the sequence: a customer request, the internal discussion, the document that was drafted, the ticket that closed it.

#### What makes company records more valuable?

Records are worth more when they cover a long period and connect across systems. In rough order of impact:

1. **Span.** Years of continuous history rather than a recent slice.
2. **Linked systems.** Tickets that reference commits, chat threads that reference documents, CRM records that connect to email.
3. **Outcomes.** Resolved tickets, merged code, signed deals and closed projects show how a piece of work ended.
4. **Domain depth.** Specialised industries and workflows that public data does not describe.
5. **Clean exports.** Complete archives with timestamps, thread structure and roles intact.

#### What rights and consent do company records need?

This is the data type with the most people in it, so the rights work is heavier than for footage or sensor logs. The gates that come up most:

- **Platform terms.** Slack's API terms say an app offered outside your organisation may not use API data to train a large language model, and may not bulk export message and file data unless an additional agreement allows it ([Slack API Terms, effective Oct 10 2025](https://slack.com/terms-of-service/api)). Read your own Slack agreement before exporting anything.
- **Employees.** Whether employees' consent extends to a sale of their messages is, in one law professor's words, "an ethically gray area" ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)). Lawmakers are watching: on Oct 8 more than 120 US lawmakers urged Google and Spirit Airlines to protect former employees' privacy in the airline's proposed data deal ([IBTimes UK, Oct 20 2026](https://www.ibtimes.co.uk/companies-sell-workplace-data-ai-firms-1826648)).
- **Customers and clients.** Client confidentiality clauses, privileged communications and regulated data (health, financial, children's) can rule whole folders out.
- **Vendors' IP.** Records held in a supplier's system can embed the supplier's own work. The head of an airline-operations vendor told Business Insider he had less than 48 hours' notice that intellectual property his company had built over 15 years could be given away in its client's data sale ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)).
- **Code.** Credentials and secrets in repositories, third-party and open-source code under its own licences, and client-owned work.

This page is general information, not legal advice.

#### What are company records worth?

Public figures vary by orders of magnitude, because they mix bankrupt-estate auctions, small wind-down sales and buyers' advertised ranges. Each is dated and labelled by market side; none is a fearofgod.ai price.

| Reference point | Market side | Figure and source |
|---|---|---|
| Bankrupt airline's documents, emails and data | Court-process bids, Sep 2026 | [$10 million top bid, then a $12.5 million proposal; another bidder offered $7.5 million; no sale court-approved as of Sep 14 (Fortune)](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/) |
| Same archive, for scale | Reported size, Oct 2026 | [about 100 million emails and 500 million Teams messages behind a proposed $10 million deal (IBTimes UK)](https://www.ibtimes.co.uk/companies-sell-workplace-data-ai-firms-1826648) |
| Dead startup's codebase and workspace data | Reported wind-down sale, Oct 2026 | [roughly $25,000 for code, about $15,000 pending for workspace data (MIT Technology Review)](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/) |
| A data buyer's programme for midsize companies | Published buyer offer, reported Oct 2026 | [$100,000 to $2 million for anonymised data (Business Insider)](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10) |

Fortune noted there is no reliable market average for a failed company's data ([Fortune](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)), and the same caution applies to live companies. Partnerships typically range from $50K to $3M+.

#### How does licensing company records work with fearofgod.ai?

You license records your company already keeps, securely, and earn additional, long-term revenue from them. They already exist in the software your teams use every day, so there is nothing new to collect.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and anything you want left out, such as a client, a channel, a mailbox or a period, stays out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

fearofgod.ai leads with data for simulating the physical world, but welcomes company records of any kind. If your business also holds footage, scans or sensor logs, see [licensing industrial sensor data](https://www.fearofgod.ai/sell/industrial-sensor-and-telemetry-data) or [licensing egocentric video](https://www.fearofgod.ai/sell/egocentric-video). [Talk to our team](https://www.fearofgod.ai/partner) when you are ready.

#### Selling company records to AI companies

A company can sell the records its work already produces to AI companies by licensing them with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

#### Frequently asked questions

##### Can we sell our company records to AI companies?

Yes, if your company owns the records and has the rights to license them. You sell them by licensing them with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case. Your company can be paid more than once, through renewals and repeat licences.

##### Can we sell our Slack data to AI companies?

A company can license its own workspace data, but not through just any app: Slack's API terms bar third-party apps offered outside your organisation from using API data to train large language models, and from bulk exporting messages unless an additional agreement allows it. Read your own Slack agreement before exporting anything. This is general information, not legal advice.

##### Do our employees have to consent?

The law varies by country, and whether past consent covers a sale is contested. Common safeguards are de-identification and leaving sensitive channels, mailboxes or people out of scope. Take legal advice for your jurisdictions.

##### We got an offer that expires in 48 hours. Should we take it?

Treat a short deadline as a warning sign. Business Insider has reported such "exploding offers"; take the time to check what is being licensed, for how long, and what happens to the data afterwards.

##### How is fearofgod.ai different from other companies that license company records?

fearofgod.ai focuses on private and niche data for those who want to simulate the world, and licenses company records alongside physical-world data. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner. Partnerships typically range from $50K to $3M+.

##### How much can our company records earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on span, linked systems, domain and rights. fearofgod.ai does not publish per-message or per-seat prices.

##### Which countries does fearofgod.ai work in?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan, in any industry.

### Sell dashcam footage and fleet video to AI companies

URL: https://www.fearofgod.ai/sell/dashcam-and-fleet-video
Updated: 2026-10-25

**Can a fleet operator license its dashcam footage for AI training?** Yes. A fleet operator that owns its dashcam footage can license it, ideally with the matching telematics, for long-term revenue, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on hours, synced vehicle data, event labels and privacy handling.

#### Can a fleet operator license its dashcam footage for AI training?

Yes, if the company owns the footage and can deal with the drivers, other road users and licence plates that appear in it. Trucking, delivery, field-service, transit and rental fleets record road-facing and often driver-facing video every working day, mostly to settle claims and coach drivers. Years of that footage, plus the vehicle data recorded with it, describe real driving across real roads in a way few other sources do.

Most of this video already lives in a fleet camera or telematics platform. Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free.

#### What do AI teams use dashcam and fleet video for?

They use it to teach models how traffic behaves and how scenes unfold from a moving vehicle, especially in the rare moments that matter most. The main uses:

- **World models for driving.** Long sequences of road video show how other vehicles, cyclists and pedestrians move, which is the raw material for predicting what happens next.
- **Rare and safety-critical events.** Near-misses, hard braking, collisions and unusual road conditions are hard to collect on purpose, and fleets record them as a matter of course.
- **Simulation.** Real drives help reconstruct roads and situations for testing autonomous systems in simulation.
- **Perception.** Varied weather, lighting, regions and vehicle types make detection and tracking models more robust.

#### What makes dashcam footage more valuable?

Video is worth much more when it is synced with what the vehicle was doing. In rough order of impact:

1. **Synced telematics.** GPS, speed, acceleration, braking, steering and, where available, CAN bus signals, time-aligned to the video.
2. **Event labels.** The triggers and review outcomes your safety system already records: harsh braking, collision, near-miss, coaching notes.
3. **Continuous drives.** Some systems keep only short clips around events. Continuous footage, where you have it, shows the build-up and the normal driving around incidents.
4. **Multiple views.** Road-facing, side and rear cameras together, with camera position and calibration.
5. **Breadth.** Many routes, regions, seasons, road types, times of day and vehicle classes, across several years.
6. **Original quality.** Full-resolution files rather than platform-compressed previews.

#### What rights and consent does dashcam footage need?

Your company needs to own the footage and be able to use it for a new purpose lawfully, for both your drivers and everyone else the cameras saw. The gates that come up most:

- **Drivers.** Driver-facing footage and driving records are about identifiable employees or contractors. The EDPB's connected-vehicle guidance says data on driving style, distance, wear and tear, location and camera footage can concern the driver and other people, and permits their identification ([EDPB Guidelines 01/2020](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-012020-processing-personal-data-context_en)). Check employment contracts, driver notices and any union agreement that covers in-cab cameras.
- **Other road users and plates.** In its video guidance, the EDPB treats licence plates and information that could identify passers-by as information relating to people, and says a dashcam installed to collect evidence of accidents should not constantly record traffic and people near the road ([EDPB Guidelines 3/2019](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-32019-processing-personal-data-through-video_en)). The UK regulator covers surveillance in vehicles and number-plate recognition in its guidance ([ICO](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/cctv-and-video-surveillance/guidance-on-video-surveillance-including-cctv/)).
- **Driver-facing analytics.** Illinois' biometric privacy law counts a scan of face geometry as a biometric identifier and bars private companies from selling or profiting from biometric identifiers or information ([740 ILCS 14](https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57)). Keep fatigue or distraction scores built on face analysis out of any licence.
- **Vendor contracts.** Read your camera or telematics agreement for who owns recorded video and vehicle data, and whether bulk export is allowed.
- **Leased, rented or customer vehicles.** If the vehicles or the routes belong to a customer, their contract may say who controls the footage.

This page is general information, not legal advice.

#### What is dashcam and fleet video worth?

Public figures for driving video come from data marketplaces, an imagery API and collection vendors; none is a licence for a fleet archive, and none is a fearofgod.ai price.

| Reference point | Market side | Figure and source |
|---|---|---|
| 360-degree 4K driving footage, 1,750+ hours, North America | Marketplace listing, checked Oct 2026 | [$150,000 for the full library (Opendatabay)](https://www.opendatabay.com/data/ai-ml/4382f6ff-5440-4883-ae76-124270115f15) |
| Same seller, Greater Los Angeles, 310+ hours | Marketplace listing, checked Oct 2026 | [$30,800 (Opendatabay)](https://www.opendatabay.com/data/ai-ml/f6a3eb9f-0a2d-4084-b514-ed122e294f43) |
| Same seller, Colorado and Wyoming, 215+ hours | Marketplace listing, checked Oct 2026 | [$21,700 (Opendatabay)](https://www.opendatabay.com/data/ai-ml/ee011773-ae18-4395-a972-d02a854ed534) |
| Event videos from a dashcam mapping network | Published API price, checked Oct 2026 | [$6 per AI event video; street-level images at $0.005 each (Bee Maps)](https://beemaps.com/pricing) |
| Managed dashcam and ambient video collection | Vendor collection cost, Mar 2026 | [$15 to $40 per captured hour; crowd platforms $8 to $15 (DataX Power)](https://www.dataxpower.com/blog/video-data-collection-cost) |
| Video licensed for AI training, general | Reported licence rate, May 2026 | [about $1 a minute (The Logic)](https://thelogic.co/news/ai-training-data-video-versos/) |

The listed driving libraries work out to roughly $86 to $101 per hour of footage, and the seller says the footage is used in film and television driving scenes ([Opendatabay](https://www.opendatabay.com/data/ai-ml/4382f6ff-5440-4883-ae76-124270115f15)). Fleet archives are a different product: less cinematic, but often with years of synced vehicle data and real incidents. Partnerships typically range from $50K to $3M+.

#### How does licensing fleet video work with fearofgod.ai?

You license footage your fleet already holds, securely, and earn additional, long-term revenue from it. The footage and telematics already sit in your fleet platform, so there is nothing new to record.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and anything you want left out, such as driver-facing video, a customer's routes or a region, stays out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

A useful first description covers the number of vehicles, years of retained footage, camera views, which telematics fields are recorded, and the platform it lives in. [Talk to our team](https://www.fearofgod.ai/partner) when you are ready. Fleets that also hold sensor histories from depots or equipment can read [licensing industrial sensor data](https://www.fearofgod.ai/sell/industrial-sensor-and-telemetry-data).

#### Selling dashcam footage to AI companies

A fleet operator that owns its dashcam footage can sell it to AI companies by licensing it with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences.

#### Frequently asked questions

##### Can we sell our dashcam footage to AI companies?

Yes, if your company owns the footage and has the rights to license it. You sell it by licensing it with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case. Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

##### We have eight years of dashcam footage from our trucks. Can we license it?

Very likely, if the company owns the footage and can handle drivers, other road users and plates lawfully. Footage that comes with the matching speed, GPS and braking data, and with your safety events, is the most useful.

##### Is road-facing video enough, or do buyers want telematics too?

Road-facing video alone has value, but synced vehicle data makes it far more useful, because it tells a model what the vehicle was doing at each moment. Keep the two together and time-aligned if you can.

##### Do our drivers have to agree?

It depends on your jurisdiction, your driver notices and any union agreement. In the EU and UK, driving data and driver-facing video are personal data, so the new use needs a lawful basis. This is general information, not legal advice. Driver-facing footage can also be left out of scope entirely.

##### Our footage lives in our telematics vendor's platform. Is that a problem?

Usually not, but read the vendor agreement first for who owns recorded video and whether bulk export is allowed. Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free.

##### How much can our fleet footage earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on hours, synced vehicle data, event labels, regions and rights. fearofgod.ai does not publish per-hour prices.

##### Does fearofgod.ai buy dashcam clips from individual drivers?

No. fearofgod.ai works with companies, not individuals. The footage has to belong to the company that runs the fleet.

##### Which countries does fearofgod.ai work in?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

### Sell drone footage and aerial video to AI companies

URL: https://www.fearofgod.ai/sell/drone-and-aerial-footage
Updated: 2026-10-25

**Can a company license its drone footage for AI training?** Yes. Companies that own drone or aerial footage, from survey and inspection flights to media shoots, can license it securely for long-term revenue, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on telemetry, coverage, repeat surveys and ownership.

#### Can a company license its drone footage for AI training?

Yes, if the company holds the rights to the footage and can deal with any people, vehicles or private property that appear in it. Survey and mapping firms, inspection contractors, agricultural operators, construction companies and production studios all build up large drone archives, and much of that footage can sit unused on hard drives once the deliverable is sent.

Aerial data owners are already licensing archives for this purpose. Australian aerial mapping company Aerometrex announced two off-the-shelf data licences with AI companies worth a combined AUD 1.07 million, covering 3D datasets from its existing catalogues for an autonomous simulation engine and LiDAR datasets for a utility digital twin platform ([Aerometrex ASX announcement, Jun 17 2026](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf)). That is a disclosed licence, not a price list, but it shows that an existing catalogue, not new flights, was what the buyers wanted.

#### What do AI teams use drone footage for?

They use it to teach models what the world looks like from above and how it changes, and to build 3D environments for simulation. The main uses:

- **Simulation and digital twins.** Overlapping aerial frames and LiDAR can be turned into 3D models of real places, which simulation teams use as training environments. The Aerometrex buyers named exactly these uses ([ASX announcement](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf)).
- **World models.** Long, continuous flights show how a scene shifts with motion, light and weather, which helps models learn to predict what comes next.
- **Autonomous flight and navigation.** Footage paired with position and attitude data shows a model what a camera sees for a given flight path.
- **Perception and inspection.** Footage of roofs, towers, lines, pipelines, crops and sites, linked to the defects or findings an inspector logged, is supervised data for detection models.

#### What makes drone footage more valuable?

Footage is worth more when a buyer knows exactly where the camera was, where it pointed, and what was found. In rough order of impact:

1. **Flight telemetry.** GPS position, altitude, speed and gimbal angles for each frame, from flight logs or sidecar files. Without it, footage is just video.
2. **Camera details.** Sensor model, lens, calibration, and for survey work the ground control points and processing reports.
3. **Repeat coverage.** The same site flown monthly or yearly shows change over time, which single flights cannot.
4. **Findings attached.** Inspection reports, defect logs or crop assessments tied to the flights that produced them.
5. **Other sensors.** LiDAR, thermal and multispectral captures alongside RGB video.
6. **Raw over edited.** Full-length originals, including the "boring" transit legs, are often more useful for training than polished cuts.
7. **Breadth.** Many regions, terrains, seasons, weather conditions and times of day.

#### What rights and consent does drone footage need?

You need to own the footage, or have the right to license it, and you need to manage the people and private property that appear in it. The gates that come up most:

- **Client-commissioned flights.** If a client paid for the flight, check the contract before anything else. Under US copyright law, a commissioned work counts as a work made for hire only if it falls within nine listed categories (one is "a part of a motion picture or other audiovisual work") and both parties signed a written agreement saying so ([US Copyright Office, Circular 30](https://www.copyright.gov/circs/circ30.pdf)). Otherwise ownership turns on what the contract assigned or licensed, and the contract may also restrict reuse of client data.
- **Freelance pilots.** Footage shot by subcontracted pilots may belong to them unless rights were assigned in writing.
- **People and vehicles below.** In the EU and UK, identifiable passers-by and licence plates are personal data. The EDPB's video guidance treats licence plates and information that could identify passers-by as the kind of information that brings a camera into data protection law ([EDPB Guidelines 3/2019](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-32019-processing-personal-data-through-video_en)). The UK regulator has a section on drones in its surveillance guidance ([ICO](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/cctv-and-video-surveillance/guidance-on-video-surveillance-including-cctv/)).
- **Stock agency agreements.** If clips were placed with a stock agency under an exclusive agreement, those clips may not be free to license elsewhere.
- **Sensitive sites.** Footage of critical infrastructure or restricted sites may be covered by client security terms.

This page is general information, not legal advice.

#### How does an AI licence differ from selling stock drone footage?

A stock licence sells individual clips for use in a production, while an AI-training licence covers a whole collection for teaching models, so the economics and what matters differ.

| | Stock footage licence | AI-training licence |
|---|---|---|
| What is licensed | Individual clips, chosen by a buyer | A collection or archive, often in bulk |
| What matters most | The best-looking shots | Volume, coverage, telemetry, consistency |
| Raw and unedited footage | Less often what buyers pick | Can be useful |
| Typical price unit | Per clip or download | Per minute, hour, clip or whole collection |
| Contributor share at one agency | Agency-set royalty | [20% average royalty on data-licence revenue at Shutterstock](https://submit.shutterstock.com/help/en/articles/10594694-shutterstock-data-licensing-and-the-contributor-fund) |

#### What is drone footage worth?

No one publishes a price list for licensed drone archives, but several reported rates and published price lists for video licensed to AI developers give a range. Each is dated and labelled by market side; none is a fearofgod.ai price.

| Reference point | Market side | Figure and source |
|---|---|---|
| Two AI data licences, existing 3D and LiDAR catalogues | Disclosed licence, Jun 2026 | [AUD 1.07 million combined (Aerometrex, ASX)](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf) |
| Unused creator footage | Reported licence rate, Jan 2025 | [$1 to $4 per minute (Bloomberg Law)](https://news.bloomberglaw.com/ip-law/youtubers-are-selling-their-unused-video-footage-to-ai-companies) |
| Adobe buying video for a generative model | Reported purchase rate, Apr 2024 | [average $2.62 per minute (PYMNTS, citing Bloomberg)](https://www.pymnts.com/news/artificial-intelligence/2024/adobe-paying-videos-images-train-ai-models/) |
| Footage licensed through a video broker | Reported licence rate, Mar 2025 | [$0.75 to $3 per minute (New York Post via AOL)](https://www.aol.com/influencers-making-big-money-selling-181507475.html) |
| Video licensed for AI training, general | Reported licence rate, May 2026 | [about $1 a minute; specialised scenes up to $30 per 30-second clip (The Logic)](https://thelogic.co/news/ai-training-data-video-versos/) |
| Film archive's AI-training licence | Published price list, checked Oct 2026 | [$3 per clip (proxies) or $10 per clip (4K masters), $10,000 minimum (Stockfilm)](https://stockfilm.com/ai-training) |

Per-minute rates apply to video in general. Survey-grade aerial data with telemetry, LiDAR or 3D outputs is a different product, and the Aerometrex licences were disclosed as catalogue deals, not per-minute rates. Partnerships typically range from $50K to $3M+.

#### How does licensing drone footage work with fearofgod.ai?

You license flights your company has already flown, securely, and earn additional, long-term revenue from work you have already done. No new data collection is needed.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and flights you want left out, such as a particular client's sites, stay out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

A useful first description covers total hours or terabytes, years flown, regions, sensors (RGB, thermal, LiDAR), whether telemetry was kept, and which flights were for clients. [Talk to our team](https://www.fearofgod.ai/partner) when you are ready. If your archive includes processed point clouds and models, see [licensing 3D scans and CAD](https://www.fearofgod.ai/sell/3d-scans-and-cad).

#### Selling drone footage to AI companies

A company that owns its drone and aerial footage can sell it to AI companies by licensing it with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences.

#### Frequently asked questions

##### Can we sell our drone footage to AI companies?

Yes, if your company owns the footage and has the rights to license it. You sell it by licensing it with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case. Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

##### Can I sell drone footage that I shot for a client?

Only if the contract lets you. Check who owns the footage and whether you may reuse it. In the US, a commissioned work is a work made for hire only in listed categories and with a signed written agreement, so ownership usually depends on what the contract assigned or licensed. This is general information, not legal advice.

##### Is drone footage worth more to AI companies than to stock sites?

It can be, because AI buyers value whole collections, raw flights and telemetry that stock buyers ignore. Stock sales reward a few great shots; AI licensing rewards volume, coverage and metadata.

##### Do we need to blur people and licence plates?

In the EU and UK, identifiable people and licence plates are personal data, so they need to be handled before footage is shared. This is general information, not legal advice. fearofgod.ai de-identifies the data before it goes to buyers.

##### What metadata should we keep from our flights?

Keep flight logs and per-frame GPS, altitude and gimbal angles, the camera and lens details, any ground control and processing reports, and the inspection or survey findings tied to each flight.

##### How much can our drone archive earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on hours, coverage, telemetry, other sensors and rights. fearofgod.ai does not publish a per-minute price.

##### Does fearofgod.ai work with individual drone pilots?

No. fearofgod.ai works with companies, not individuals.

##### Which countries does fearofgod.ai work in?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

### Sell or license egocentric video for AI training

URL: https://www.fearofgod.ai/sell/egocentric-video
Updated: 2026-10-25

**Can a company license its egocentric video for AI training?** Yes. A company that owns first-person footage of real work, such as helmet-cam or body-cam video, can license it securely for long-term revenue, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on hours, task coverage, synced sensors and consent.

#### Can a company license its egocentric video for AI training?

Yes, provided the company owns the footage and the people in it can either be covered by an appropriate legal basis or removed and de-identified. Egocentric video is footage recorded from the wearer's point of view: a helmet camera on a building site, smart glasses on a maintenance round, a chest camera on a delivery route, a head-mounted rig used to film a procedure for new hires. Often it was captured for safety, quality or training, then archived.

Data-collection startups are paying gig workers to film household chores for robot training. The Globe and Mail reported that one contributor recorded 17 hours of household tasks and was paid US$50, about US$2.94 an hour ([The Globe and Mail, Oct 9 2026](https://www.theglobeandmail.com/business/article-startups-are-hiring-low-paid-gig-workers-to-train-robots-on-manual/)). Footage of trained people doing skilled work on real equipment, in working facilities, is a different thing and is much harder to stage in a kitchen.

#### What do AI teams use egocentric video for?

They use it to learn how hands, tools and objects interact during a task, seen from roughly the angle a robot's own head or wrist camera would have. The main uses:

- **Robot manipulation.** Video of a person doing a task teaches a model what the steps are and what success looks like before any robot attempts it. One collection vendor matches rig to training goal: plain action cameras for pretraining visual encoders, instrumented glasses for retargeting human actions to robots, stereo-depth helmets for metric 3D ([Dexset, Jul 14 2026](https://dexset.ai/blogs/egocentric-data-collection-cost-comparison/)).
- **World models.** Long first-person sequences show how a scene changes as someone acts on it, which is the core of predicting what happens next.
- **Simulation and digital twins.** Footage of real workspaces helps rebuild them for training in simulation. The Logic reported a buyer asking for phone or action-camera footage of homes and offices together with depth information and 3D meshes of the objects in view ([The Logic, May 25 2026](https://thelogic.co/news/ai-training-data-video-versos/)).
- **Benchmarks.** The best-known research set, Ego4D, holds 3,670 hours from 931 camera wearers, collected with consenting participants and de-identification where relevant ([Ego4D, arXiv](https://arxiv.org/abs/2110.07058)).

#### What makes egocentric video more valuable?

Footage is worth more when a buyer can see what task is being done, with which tools, from start to finish, and can line the video up with other signals. In rough order of impact:

1. **Hands and work in frame.** Clips where the wearer's hands, the tool and the workpiece stay visible are the core of manipulation data.
2. **Whole tasks.** Complete jobs, from setup to clean-up, beat short highlights. Footage of mistakes and how they were fixed is useful too.
3. **Labels you already have.** Work orders, checklists, SOP step names or inspection results tied to a clip turn raw video into supervised data.
4. **Synced sensors.** Depth or stereo, an IMU, gaze and audio, with calibration files. One marketplace listing describes head-mounted stereo with a 6-DOF IMU at 300 Hz aligned to each exposure and calibration per session ([Opendatabay listing](https://www.opendatabay.com/data/robotics/09ef5189-a2fd-4412-8a27-737c25e4346d)).
5. **Metadata.** Date, site type, equipment model, task name and the wearer's role (not their name).
6. **Span and variety.** Many sites, crews, seasons and equipment generations, recorded over years, at original resolution and frame rate rather than compressed exports.

#### What rights and consent does egocentric video need?

Your company needs to own the footage, and the new use has to be lawful for the people who appear in it, including the person wearing the camera. In practice that means reading what workers, clients and site visitors were told when the footage was recorded. The gates that come up most:

- **Identifiable people.** Under EU and UK data protection law, footage of identifiable people is personal data. The EDPB's video guidance adds that footage only becomes biometric (special category) data when it is put through specific technical processing to identify someone ([EDPB Guidelines 3/2019](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-32019-processing-personal-data-through-video_en)). The UK regulator covers body-worn video and action cameras in its own guidance ([ICO](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/cctv-and-video-surveillance/guidance-on-video-surveillance-including-cctv/)).
- **Biometrics.** Illinois' Biometric Information Privacy Act counts a scan of face geometry as a biometric identifier (photographs are excluded) and bars private companies from selling or otherwise profiting from biometric identifiers or information ([740 ILCS 14](https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57)). Keep face templates or face analytics derived from footage out of any licence.
- **Employees.** Whether employees' consent extends to a sale of their data is, in an Emory law professor's words quoted by Business Insider, "an ethically gray area" ([Business Insider, Oct 13 2026](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)).
- **Client sites.** Footage shot on a client's premises may fall under the confidentiality terms of that job. Screens, documents and products in view can belong to someone else.

This page is general information, not legal advice. Have counsel review the notices and contracts that cover your footage.

#### What is egocentric video worth?

There is no public price list for licensed footage of professional work, so the figures available mostly describe what it costs to record new footage or what raw hours are listed for. They are reference points, each dated and labelled by market side, not offers.

| Reference point | Market side | Figure and source |
|---|---|---|
| Gig contributor filming household tasks | Pay to a collector, Oct 2026 | [US$50 for 17 hours, about US$2.94 an hour (The Globe and Mail)](https://www.theglobeandmail.com/business/article-startups-are-hiring-low-paid-gig-workers-to-train-robots-on-manual/) |
| Head-mounted capture, by rig | Vendor collection cost, Jul 2026 | [$15 to $20 (action camera) up to $25 to $40 (instrumented glasses) per finished hour (Dexset)](https://dexset.ai/blogs/egocentric-data-collection-cost-comparison/) |
| First-person data in China | Reported market quote, Sep 2026 | [about CNY 100 per hour (Xinhua Finance)](https://www.cnfin.com/hg-lb/detail/20260924/4474923_1.html) |
| 49.3 hours of raw industrial stereo, non-exclusive | Marketplace listing, checked Oct 2026 | [8 USD per hour, listed at $395 (Opendatabay)](https://www.opendatabay.com/data/robotics/09ef5189-a2fd-4412-8a27-737c25e4346d) |
| 25,000 hours of factory assembly-line footage | Marketplace listing, checked Oct 2026 | [$111,800 (Opendatabay)](https://www.opendatabay.com/data/robotics/5f447749-cc5a-4c7e-975b-eb0fddaadced) |
| Video licensed for AI training, general | Reported licence rate, May 2026 | [about $1 a minute, specialised scenes up to $30 per 30-second clip (The Logic)](https://thelogic.co/news/ai-training-data-video-versos/) |

Even the labelled 25,000-hour factory listing above works out to under $5 an hour ([Opendatabay](https://www.opendatabay.com/data/robotics/5f447749-cc5a-4c7e-975b-eb0fddaadced)). Hours alone are not what is scarce; depth is: skilled work with synced sensors, work orders and outcomes attached, and the rights cleared. fearofgod.ai does not publish per-hour prices. Partnerships typically range from $50K to $3M+.

#### How does licensing egocentric video work with fearofgod.ai?

You license footage your company already holds, securely, and earn additional, long-term revenue from it. If it already sits on a camera vendor's platform or a cloud drive, licensing it means re-licensing what you already hold. No new data collection is needed.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and anything you want left out, such as a site, a client or a period, stays out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

A rough first description is enough: hours, cameras, kinds of work, and the notices given to staff. [Talk to our team](https://www.fearofgod.ai/partner) when you are ready. Companies with robot logs as well as human video can read [licensing robot and teleoperation data](https://www.fearofgod.ai/sell/robot-and-teleoperation-data).

#### Selling egocentric video to AI companies

A company that owns first-person footage of its work can sell it to AI companies by licensing it with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences.

#### Frequently asked questions

##### Can we sell our egocentric video to AI companies?

Yes, if your company owns the footage and has the rights to license it. You sell it by licensing it with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case. Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

##### We have years of helmet-cam footage from our job sites. Can we license it?

Usually, yes, if the company owns the footage and the people in it can be handled lawfully, which in practice often means de-identification and a review of what crews were told at the time. Footage of whole tasks, with hands and tools in view, is the most useful part.

##### Do the workers who appear in the footage have to consent?

It depends on where the footage was recorded and what workers were told. In the EU and UK, identifiable people in video are personal data, so the new use needs a lawful basis. The usual routes are fresh consent or removing and de-identifying the people concerned. Ask counsel to review your notices before you share anything. This is general information, not legal advice.

##### Why would AI teams license our footage if they can pay people to film at home?

Home capture mostly shows household chores filmed by untrained people. Footage of trained staff using real tools in real facilities is harder to stage, and it often comes with work orders, checklists or inspection results that explain what each clip shows.

##### What should we keep with the footage?

Keep originals rather than compressed exports, plus any synced sensor files, calibration data, timestamps, site and equipment details, and the job records tied to each recording. The wearer's role is useful; their name is not needed.

##### How much can our egocentric video earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on hours, task coverage, labels, synced sensors and rights, so fearofgod.ai does not publish a per-hour price. The public figures on this page are market reference points, not fearofgod.ai prices.

##### Does fearofgod.ai buy footage from individual camera wearers?

No. fearofgod.ai works with companies, not individuals. fearofgod.ai licenses data that a company owns.

##### What happens to the footage after it is delivered?

fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

##### Which companies can apply?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan, in any industry.

### Sell industrial sensor data and telemetry to AI companies

URL: https://www.fearofgod.ai/sell/industrial-sensor-and-telemetry-data
Updated: 2026-10-25

**Can a company license its industrial sensor and telemetry data for AI training?** Yes. Manufacturers, utilities and operators that own years of sensor, historian and machine telemetry can license it securely for long-term revenue, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on history, labelled events, tag context and rights.

#### Can a company license its industrial sensor and telemetry data for AI training?

Yes, if the company owns the data, the contracts with equipment makers and customers allow it, and sensitive operational detail can be kept out. Plants, utilities, mines, water operators and equipment makers with connected fleets store years of time series: process historian tags, PLC and SCADA logs, alarms and events, vibration and condition monitoring, energy metering, and the maintenance work orders that explain what happened. It is kept for operations and compliance, and rarely thought of as something another party would pay for.

To a team building models of the physical world, that history is a record of how real machines and processes behave, including the rare failures that cannot be staged. Utilities' physical networks are already a target for this kind of modelling: one buyer of licensed aerial LiDAR runs a digital twin platform that utilities use to predict network failures ([Aerometrex ASX announcement, Jun 17 2026](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf)).

#### What do AI teams use industrial sensor data for?

They use it to learn how physical systems respond over time, so models can predict, simulate and eventually control them. The main uses:

- **World models and simulation.** Long records of inputs, setpoints and outcomes let a model learn the dynamics of a process, which simulation teams use to calibrate digital twins against reality.
- **Failure prediction.** Sensor histories joined to the failures and repairs that followed are supervised data for predictive maintenance and anomaly detection.
- **Robotics in industrial settings.** Robots working on a line or in a plant need context about the process around them, not just their own camera feeds.
- **Agents for operations software.** Alarm logs, operator actions and the work orders raised in response show how people actually run a site.

#### What makes industrial sensor data more valuable?

A tag history is worth far more when a buyer can tell what each signal measures and what was happening at the time. In rough order of impact:

1. **Labelled events.** Failures, trips, alarms, maintenance work orders and root-cause notes, time-aligned with the sensor data.
2. **Tag context.** Descriptions, units, ranges, sampling rates, the asset hierarchy, and diagrams that show which sensor sits where.
3. **Long, raw history.** Years at the original sampling rate, before compression or averaging, including periods of abnormal operation.
4. **Control and outcome data.** Setpoints, recipes, operator actions, batch records and quality results joined to the process data.
5. **Repetition.** Many similar assets or sites, so a model can separate what is normal for one machine from what is normal for the type.
6. **Other modalities.** Machine-vision images, thermal imaging or acoustic data recorded alongside the time series.

#### What rights and consent does industrial data need?

Your company needs to own the data or have the right to license it, and some operational detail may be off limits for security or export reasons. The gates that come up most:

- **Equipment and platform contracts.** Data from connected machines often passes through the maker's or a software vendor's platform. Check who the service agreement says owns the data and whether you may export it in bulk.
- **Customers' machines.** If you make equipment and collect telemetry from customer sites, that data may belong to, or be controlled by, your customers under their contracts.
- **Critical infrastructure.** Grid entities with high- or medium-impact BES Cyber Systems must run information protection programmes for BES Cyber System Information ([NERC CIP-011-3](https://www.nerc.com/globalassets/standards/reliability-standards/cip/cip-011-3.pdf)). Keep network diagrams, configurations and security data of that kind out of scope.
- **Export-controlled know-how.** Under US export rules, releasing controlled technology to a foreign person counts as an export ([15 CFR 734.13](https://www.ecfr.gov/current/title-15/subtitle-B/chapter-VII/subchapter-C/part-734/section-734.13)). Process data that embeds controlled production know-how needs review.
- **Trade secrets.** Recipes, yields and process windows may be the most valuable thing you own. Decide what stays private before discussing scope.
- **People in the data.** Operator IDs, badge events and location traces are personal data. The EDPB's connected-vehicle guidance notes that even technical data such as wear and tear can relate to an identifiable person ([EDPB Guidelines 01/2020](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-012020-processing-personal-data-context_en)).

This page is general information, not legal advice.

#### What is industrial sensor data worth?

We could not find a public per-unit price for licensed industrial sensor or historian data as of October 2026. The nearest public reference points are adjacent, so treat them as context, not comparables. None is a fearofgod.ai price.

| Reference point | Market side | Figure and source |
|---|---|---|
| Aerial LiDAR and 3D datasets licensed to AI companies, one for a utility digital twin | Disclosed licence, Jun 2026 | [AUD 1.07 million across two licences (Aerometrex, ASX)](https://announcements.asx.com.au/asxpdf/20260617/pdf/070qcgp7xlkknh.pdf) |
| De-identified structured records, priced per record delivered (healthcare, not sensor data; a related-party deal) | Disclosed licence, Oct 2026 | [$100 per delivered record, invoiced monthly (Tempus AI 8-K)](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm) |
| Spending on data for AI, all types | Analyst forecast, reported Oct 2026 | [almost $6.5 billion in 2027, from under $1 billion (Gartner, via Business Insider)](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10) |

The Tempus deal matters here for its structure rather than its sector: structured records can be priced per unit delivered, with updates paid separately ([Tempus AI 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)). Partnerships typically range from $50K to $3M+.

#### How does licensing industrial data work with fearofgod.ai?

You license history your plants and assets have already recorded, securely, and earn additional, long-term revenue from it. Historians, SCADA archives, maintenance systems and IoT platforms already hold the data. Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free. No new data collection is needed.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and anything you want left out, such as recipes, security-relevant data or a site, stays out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

A useful first description covers the systems involved, years of history, the number of tags or assets, sampling rates, and whether work orders or quality records can be joined to the data. [Talk to our team](https://www.fearofgod.ai/partner) when you are ready. Plants that also film work on the floor can read [licensing egocentric video](https://www.fearofgod.ai/sell/egocentric-video).

#### Selling industrial sensor data to AI companies

A company that owns its industrial sensor and telemetry history can sell it to AI companies by licensing it with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences.

#### Frequently asked questions

##### Can we sell our industrial sensor data to AI companies?

Yes, if your company owns the data and has the rights to license it. You sell it by licensing it with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case. Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

##### We have 15 years of historian data from our plants. Is it worth anything to AI companies?

It can be, especially if it can be joined to maintenance records and failures, and if the tags are documented. Undocumented tags with no events attached are much harder for a buyer to use.

##### Do we need the equipment maker's permission?

Sometimes. If the data sits on the maker's platform, or the service agreement gives them rights over it, read that contract before you scope anything. Data in your own historian is usually simpler.

##### Can a utility license operational data without creating a security risk?

Only with care. Security-sensitive information, such as network configurations and cyber system data covered by NERC CIP-011, should stay out of scope. Process and condition data with that detail removed is a different question.

##### Are failures and alarms more valuable than normal operation?

Both matter. Normal operation shows a model what "healthy" looks like; failures and alarms, with the work orders that followed, show what goes wrong. The combination is what makes a history useful.

##### How much can our sensor data earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on history, labels, tag context, repetition across assets and rights. fearofgod.ai does not publish per-tag or per-year prices.

##### Which systems hold useful industrial data?

Process historians, SCADA and PLC logs, maintenance and work-order systems, condition-monitoring tools, energy meters and IoT platforms. The most useful histories join several of them, so sensor readings can be matched with the events and repairs that explain them.

##### Which countries does fearofgod.ai work in?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

### Sell robot and teleoperation data to AI companies

URL: https://www.fearofgod.ai/sell/robot-and-teleoperation-data
Updated: 2026-10-25

**Can a robotics company license its teleoperation and robot data for AI training?** Yes. A robotics company that owns teleoperation demonstrations or robot logs can license them securely, even while winding down, and fearofgod.ai partnerships typically range from $50K to $3M+. Value depends on synced actions, embodiment detail, labels and rights.

#### Can a robotics company license its teleoperation and robot data for AI training?

Yes, if the company owns the recordings and the customer contracts and people involved allow it. Robotics companies accumulate exactly the data that robot-learning teams find hardest to produce: teleoperated demonstrations, autonomous runs with their successes and failures, test-bench recordings and field-deployment logs, often stored as ROS bags or similar log files. A team that pivoted, shut down a product line or is winding down can still hold years of it.

That data is expensive to recreate. One collection vendor puts teleoperation at $28 to $60 per raw hour once rigs, operators and quality checks are counted ([Dexset, Jul 2 2026](https://dexset.ai/blogs/robot-training-data-costs-pricing-complete-2026/)), and that figure covers capture only, not the years of engineering behind a working robot.

#### What do AI teams use robot and teleoperation data for?

They use it to train policies that map what a robot senses to what it should do, and to test whether those policies transfer across robots and settings. The main uses:

- **Generalist robot policies.** The Open X-Embodiment project pooled data from 22 different robots across 21 institutions, covering 527 skills, and found that a model trained on the mix improved the capabilities of multiple robots by drawing on experience from other platforms ([Open X-Embodiment, arXiv](https://arxiv.org/abs/2310.08864)). Data from robots outside such collections can widen that mix.
- **Imitation learning.** Teleoperated demonstrations show a model executable trajectories for a task, not just video of a person doing it.
- **World models and simulation.** Logs of how objects and robots actually moved help models predict outcomes and help simulation teams calibrate physics against reality.
- **Evaluation.** Recorded runs with known outcomes are used to test new policies before they touch hardware.

#### What makes robot and teleoperation data more valuable?

Logs are worth most when every observation is paired with the action that followed and a buyer can reconstruct the robot that produced them. In rough order of impact:

1. **Synced observations and actions.** Camera frames, joint states, end-effector poses, gripper state and the commands sent, on a shared clock.
2. **Embodiment detail.** Robot model, kinematic description (URDF or equivalent), camera placement and calibration, controller settings.
3. **Outcomes.** Whether each episode succeeded, and how failures happened and were recovered. Failure data is useful, not an embarrassment.
4. **Task and language labels.** Task names, step boundaries, operator notes or instructions per episode.
5. **Variety.** Many objects, scenes, lighting conditions, operators and sites; long-horizon tasks as well as short picks.
6. **Clean structure.** Consistent formats and a manifest of what each log contains. The Open X-Embodiment work released its data in standardised formats for this reason ([arXiv](https://arxiv.org/abs/2310.08864)).

#### What rights and consent does robot data need?

Your company needs to own the logs and be free to license them, so the contracts around the data matter as much as the data. The gates that come up most:

- **Customer deployments.** Logs recorded on a customer's premises may be the customer's confidential information under the deployment contract. Read it before anything else.
- **People in view.** Camera streams from warehouses, homes or hospitals can show workers, residents or patients. In the EU and UK identifiable people in video are personal data ([EDPB Guidelines 3/2019](https://www.edpb.europa.eu/our-work-tools/our-documents/guidelines/guidelines-32019-processing-personal-data-through-video_en)).
- **Operators.** Teleoperator IDs, voices on audio channels and any VR headset tracking tied to named staff are personal data too.
- **Funding and research agreements.** Grants, university partnerships and loans can carry terms about who owns or may sell data and IP.
- **Companies winding down.** A dissolved Delaware corporation continues for three years to wind up its affairs, including to "dispose of and convey their property" ([8 Del. C. § 278](https://delcode.delaware.gov/title8/c001/sc10/index.html)). Data sales from shut-down startups are already happening: MIT Technology Review reported a founder netting roughly $25,000 for a dead startup's codebase, with more pending for its workspace data ([MIT Technology Review, Oct 13 2026](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)).

This page is general information, not legal advice.

#### What is teleoperation data worth?

Public figures for robot data are almost all collection costs and market quotes; disclosed licences of an existing robot archive are rare. Each figure below is dated and labelled by market side, and none is a fearofgod.ai price.

| Reference point | Market side | Figure and source |
|---|---|---|
| Teleoperation capture, rigs plus operators plus QA | Vendor collection cost, Jul 2026 | [$28 to $60 per raw hour (Dexset)](https://dexset.ai/blogs/robot-training-data-costs-pricing-complete-2026/) |
| Fully loaded teleoperation data, pick-and-place with wrist and external cameras | Vendor's own estimate (sells collection services), Mar 2026 | [$118 per hour, down from about $340 per hour in early 2024 (Robotics Center)](https://roboticscenter.ai/state-of-robotics-2026) |
| Real-robot teleoperation data in China | Reported market quote, Sep 2026 | [about CNY 500 to 1,000 per hour (Xinhua Finance)](https://www.cnfin.com/hg-lb/detail/20260924/4474923_1.html) |

The same Xinhua Finance report puts first-person human video at about CNY 100 per hour ([Xinhua Finance](https://www.cnfin.com/hg-lb/detail/20260924/4474923_1.html)), so real-robot data is quoted at several times the price of human video. Collection costs are falling, which suggests the lasting value sits in logs from robots, tasks and sites that are hard to reproduce. Partnerships typically range from $50K to $3M+.

#### How does licensing robot data work with fearofgod.ai?

You license logs your robots have already recorded, securely, and earn additional revenue from them. The logs already exist, whether on a fleet server, in cloud storage or on the drives of a team that has moved on. No new data collection is needed.

The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. You approve the scope before anything is shared, and anything you want left out, such as a customer's deployments or a proprietary controller, stays out. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.

A useful first description covers which robots, how many hours or episodes, what each log contains, the file formats, and whether customer contracts touch the data. [Talk to our team](https://www.fearofgod.ai/partner) when you are ready. If you also hold human first-person video of the same tasks, see [licensing egocentric video](https://www.fearofgod.ai/sell/egocentric-video).

#### Selling robot and teleoperation data to AI companies

A robotics company that owns teleoperation demonstrations or robot logs can sell them to AI companies by licensing them with fearofgod.ai, securely. Partnerships typically range from $50K to $3M+. A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

#### Frequently asked questions

##### Can we sell our robot and teleoperation data to AI companies?

Yes, if your company owns the data and has the rights to license it. You sell it by licensing it with fearofgod.ai. Your company can keep ownership of its data. This is agreed case by case.

##### Our robotics startup is shutting down. Can we still license our robot data?

Often, yes. A dissolved Delaware corporation continues for three years to wind up its affairs and dispose of its property, and shut-down startups are already selling data to AI buyers. Check investor, lender and customer agreements first, and talk to whoever is running the wind-down. This is general information, not legal advice.

##### Are failed runs worth including?

Yes. Episodes that failed, with the reason recorded, show a model what goes wrong and how recoveries look. Keep success and failure labels with every episode.

##### Our data is in ROS bags with custom message types. Is that a problem?

No, as long as the message definitions and the robot description are kept with the bags. Without them, a buyer cannot tell what each field means.

##### Do customer contracts stop us from licensing deployment logs?

They might. A deployment contract may treat data recorded on site as the customer's confidential information. Logs from your own labs and test benches are often simpler to license than field data.

##### How much can our teleoperation data earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Value depends on hours, embodiment, sensors, labels, variety and rights. fearofgod.ai does not publish per-hour or per-episode prices.

##### Does fearofgod.ai work with individual teleoperators?

No. fearofgod.ai works with companies, not individuals. The data has to belong to the company that recorded it.

##### Which countries does fearofgod.ai work in?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

## Guides

URL: https://www.fearofgod.ai/guides

Plain answers for companies deciding whether, how and on what terms to license their data.

### AI data licensing agreement: clause by clause

URL: https://www.fearofgod.ai/guides/data-licensing-agreement-ai-training
Updated: 2026-10-25

**What should a data licensing agreement for AI training contain?** A data licensing agreement for AI training sets out what data is licensed, what the buyer may do with it and for how long, how and when the seller is paid, and what the seller promises about rights and privacy. The clauses that move the most value are the grant, acceptance, deletion, warranties and the liability cap.

> **TL;DR**
>
> - micro1's public Data License Master Terms (updated Oct 7 2026) take a "worldwide, royalty-free, fully paid-up, nonexclusive, irrevocable, perpetual license" that the buyer can sublicense through multiple tiers ([micro1](https://www.micro1.ai/legal/data-license-master-terms)).
> - Under those terms the seller can invoice only after the buyer accepts the data, and the buyer has 30 business days after delivery to reject it ([micro1](https://www.micro1.ai/legal/data-license-master-terms)).
> - Deletion falls due 120 days after final delivery, but nothing requires the buyer to delete or retrain models, weights or derivative works ([micro1](https://www.micro1.ai/legal/data-license-master-terms)).
> - Each side's liability is capped at three times the fees, except for fraud, gross negligence and wilful misconduct, and the buyer can claw back every payment if the seller's description of the data proves materially inaccurate ([micro1](https://www.micro1.ai/legal/data-license-master-terms)).
> - Per-unit pricing appears in public filings: Tempus licenses de-identified neurology records to Pathos at $100 per delivered record, invoiced monthly, plus $5.0 million for record updates. It is a related-party deal, since the two companies share executives and owners ([Tempus 8-K, Oct 23 2026](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
> - Exclusivity can be a window rather than forever: Tempus agreed not to license the same neurology records to other third parties until 18 months after the later of delivery and Oct 22 2027 ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
> - Terms are finite. Reddit's AI data licenses run two to three years ([Reddit prospectus, SEC](https://www.sec.gov/Archives/edgar/data/1713445/000162828024012380/reddit-final424b4.htm)), and Tempus's oncology license with Bristol Myers Squibb ends Dec 31 2031, with 12 months' notice to terminate early ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
> - Some clauses are required by law. In California, a contract to sell or license de-identified patient information must prohibit re-identification ([Cal. Civ. Code § 1798.148(c)](https://cppa.ca.gov/regulations/pdf/ccpa_statute.pdf)).

#### What is a data licensing agreement for AI training?

A data licensing agreement for AI training is the contract under which a company lets a model developer use a defined copy of its data, for defined purposes, for a defined time, in exchange for payment. Under a license, unlike a sale, the company keeps ownership. The agreement usually has two layers: master terms that cover every delivery, and an order form or statement of work for each dataset. Tempus's 2026 deals follow that pattern. The Bristol Myers Squibb oncology license is a statement of work under a master services agreement dated 2018, and the Pathos neurology license is an order form under a master agreement restated in 2024 ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).

Most of these contracts are confidential, so this guide uses the two most detailed public sources: micro1's published [Data License Master Terms](https://www.micro1.ai/legal/data-license-master-terms), a buyer's standard form, and the [Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm) of Oct 23 2026, which describes negotiated terms. A standard form shows where a negotiation starts, not where it has to end.

#### What should the data description schedule contain?

The schedule should describe the data precisely enough that both sides can tell whether a delivery matches it, because acceptance, pricing and every warranty point back to it. A good schedule lists:

- **Sources and contents:** each system or archive, and what it holds.
- **Volume and span:** hours, files, records or scans, and the dates covered.
- **Formats and fields:** resolution, frame rate, sensor channels, metadata.
- **Exclusions:** what is deliberately left out, such as HR files, named clients or audio.
- **De-identification:** what was removed or transformed, and how.
- **Delivery method and schedule:** in one batch, in tranches, or updated over time.

*Datasheets for Datasets* ([Gebru et al.](https://arxiv.org/abs/1803.09010)) gives a template: document each dataset's motivation, composition, collection process and recommended uses. That documentation also speeds up the buyer's own compliance work.

#### What does the grant clause decide?

The grant clause decides what the buyer may do with your data, and it is where most of the value changes hands. Read every adjective in it. micro1's grant is "worldwide, royalty-free, fully paid-up, nonexclusive, irrevocable, perpetual" with sublicensing "through multiple tiers" ([micro1, s. 1.2](https://www.micro1.ai/legal/data-license-master-terms)). In plain terms:

| Word in the grant | What it means for the seller |
|---|---|
| Worldwide | No territorial limit on the buyer's use |
| Royalty-free, fully paid-up | One price, with no further payment however the data or the models are used |
| Nonexclusive | You can license the same data to others |
| Irrevocable | You cannot take the rights back, even after a dispute |
| Perpetual | The rights never expire |
| Sublicensable through multiple tiers | The buyer can pass rights to others, who can pass them on again |

A negotiated grant can be much narrower. Tempus lets Pathos use the neurology data "for its internal purposes" and to help one named partner, and a side letter confirms Tempus's consent to the sublicense to that partner ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)). Points to fix in the grant:

- **Purpose.** Training, fine-tuning, evaluation, or all three.
- **Users.** Internal use only, or named affiliates.
- **Sublicensing.** None, or only with your consent.
- **Term.** A fixed term, not a perpetual one.
- **Exclusivity.** Whether it applies, and its scope (see [exclusive vs non-exclusive data license](https://www.fearofgod.ai/guides/exclusive-vs-non-exclusive-data-license)).

#### How are fees and payments usually structured?

AI data licenses are usually paid as a fixed fee, a per-unit fee, or a mix, and the public examples show all three:

- **Fixed payments over a term.** Bristol Myers Squibb pays Tempus $137.5 million in aggregate fixed payments to the end of 2031, invoiced quarterly, for a data license bundled with software access and research services ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
- **Per unit delivered.** Pathos pays $100 per delivered record, invoiced monthly, for up to 500,000 records over five years ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
- **Fees for updates.** The same order form adds $5.0 million for the availability of record updates, invoiced quarterly ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
- **Multi-year contract value.** Reddit disclosed data licensing arrangements with an aggregate contract value of $203.0 million and terms of two to three years ([Reddit prospectus, SEC](https://www.sec.gov/Archives/edgar/data/1713445/000162828024012380/reddit-final424b4.htm)).

A per-unit price moves volume risk onto the seller; a fixed price moves it onto the buyer. Whichever you choose, the contract should say when each payment falls due and what triggers it.

#### How do delivery and acceptance clauses affect when you are paid?

Acceptance clauses decide when you are paid, because many buyers pay only for data they have accepted. Under micro1's terms the buyer can reject data within 30 business days of delivery, and the seller "may invoice Licensee only after Licensee has given written notice that the Data has been accepted, or such Data has been deemed accepted" ([micro1, ss. 1.3 and 3](https://www.micro1.ai/legal/data-license-master-terms)).

Acceptance does not end the seller's exposure. The same terms give the buyer 60 days after acceptance to claim a latent defect or a material mismatch with the Data Manifest (micro1's name for the data schedule), and that time limit does not apply to claims about rights, privacy, consents or provenance ([micro1, s. 1.3](https://www.micro1.ai/legal/data-license-master-terms)).

Protect yourself with objective acceptance criteria tied to the data schedule: volume, format, fields, and de-identification checks that can be measured. Add deemed acceptance if the buyer stays silent, and a short cure period for fixing a rejected delivery.

#### How do privacy and de-identification clauses work?

Privacy clauses say who is responsible for personal data in the delivery, how it was de-identified, and what the buyer must never do with it. The most important is a ban on re-identification. For de-identified patient information, California requires one: the contract must state that the data includes de-identified patient information, prohibit re-identification by the buyer, and bar onward disclosure except to parties bound by the same terms ([Cal. Civ. Code § 1798.148(c)](https://cppa.ca.gov/regulations/pdf/ccpa_statute.pdf)). California's general definition of "deidentified" also requires the holder to bind every recipient by contract ([§ 1798.140(m)](https://cppa.ca.gov/regulations/pdf/ccpa_statute.pdf)).

If any personal data leaves the EU or EEA, the agreement needs a transfer mechanism. The European Commission's [standard contractual clauses](https://commission.europa.eu/law/law-topic/data-protection/international-dimension-data-protection/standard-contractual-clauses-scc_en), modernised on June 4 2021, are the usual one. The [legal guide](https://www.fearofgod.ai/guides/is-it-legal-to-sell-company-data-to-ai) covers when data counts as de-identified in each region. fearofgod.ai de-identifies the data before it goes to buyers.

#### What warranties and indemnities will you give?

You will warrant that you had the right to collect and license the data, and you will indemnify the buyer if that proves false. micro1's terms ask the seller to warrant that it obtained "all consents, opt-ins or other valid legal bases" required by privacy law, obtained the releases needed for anyone whose likeness appears, and did not scrape or harvest the data from online sources without the owner's authorization ([micro1, s. 5](https://www.micro1.ai/legal/data-license-master-terms)). The seller also warrants that it has not already granted anyone rights to the same data "on the same or substantially similar terms" ([micro1, s. 5.1](https://www.micro1.ai/legal/data-license-master-terms)).

Two clauses give those promises their weight. micro1 caps each party's total liability at three times the fees under the order form, except for fraud, gross negligence and wilful misconduct ([micro1, s. 7.2](https://www.micro1.ai/legal/data-license-master-terms)). And if the Data Manifest proves "inaccurate or misleading in any material respect", the buyer can terminate at once, nothing more is payable, and the seller refunds everything already paid, even after acceptance ([micro1, s. 4.3](https://www.micro1.ai/legal/data-license-master-terms)). Ask for:

- **Knowledge qualifiers** ("to the seller's knowledge") on facts you cannot check, such as what every former employee consented to.
- **A materiality bar and a cure period** before a description error can unwind the whole deal.
- **A clear split of responsibility:** your duty covers the data as delivered, and the buyer's covers how it uses the data and what its models produce.

Buyer forms also contain protections for the seller, and they are worth keeping. micro1's terms bar the buyer from knowingly trying to re-identify anyone in de-identified data, from using the data to compete with the seller's business, and from reselling the same data as a standalone product ([micro1, s. 1.4](https://www.micro1.ai/legal/data-license-master-terms)).

#### What happens to the data and the models when the license ends?

Usually the files are deleted and the models are kept, so read the deletion clause for what it leaves out. micro1 requires deletion 120 days after final delivery or on termination, whichever comes first, with written confirmation at the seller's request, and adds that this does "not require Licensee to delete or retrain any model, weights, or derivative works created from Data" ([micro1, s. 4.4](https://www.micro1.ai/legal/data-license-master-terms)). A license for a fixed term does not stop a model trained during that term from being used after it.

Points to negotiate:

- **Certification.** A written confirmation once deletion is done, as of right rather than on request.
- **Backups.** Whether copies in backups are covered, and by when.
- **Surviving rights.** Whether models trained on your data may still be used, sold or sublicensed after the term, and whether that is reflected in the price.

Raw data is deleted according to a schedule agreed with the partner.

#### Which confidentiality, term and governing-law clauses matter?

These clauses decide who knows about the deal, how long it lasts, and where disputes go.

- **Confidentiality of the seller.** micro1's terms say the licensee "shall not disclose Licensor's identity or the source of the Data without Licensor's prior written consent" ([micro1, s. 2.3](https://www.micro1.ai/legal/data-license-master-terms)). Ask for the same wherever your name could harm your relationships with staff or clients.
- **Term and termination.** Tempus's oncology license runs to December 31 2031, and the buyer can end it early on 12 months' notice and payment of a substantial termination fee ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
- **Governing law and disputes.** micro1's terms choose the laws of California and the United States, with arbitration in California ([micro1, ss. 9.3 and 9.4](https://www.micro1.ai/legal/data-license-master-terms)). A seller elsewhere should know in advance where a dispute would be heard.

#### Which clauses should a seller negotiate first?

Negotiate the grant, acceptance, deletion and liability clauses first, because they decide how much of the data's value you keep. This checklist covers the full agreement:

| Clause | What to ask for |
|---|---|
| Data schedule | Precise sources, volume, span, formats, exclusions and de-identification |
| Grant | Named purposes, no open-ended sublicensing, a fixed term |
| Exclusivity | None, or limited by field, records and time, and priced as sales you give up |
| Fees | Payment triggers and due dates; per-unit or fixed, chosen on purpose |
| Acceptance | Objective criteria, deemed acceptance, a cure period |
| Privacy | A ban on re-identification and onward disclosure; transfer clauses where needed |
| Warranties | Knowledge qualifiers on facts you cannot verify |
| Liability | A cap that covers the indemnity, and limits on clawback for description errors |
| Deletion | A deadline, certification, backups, and an answer on models |
| Confidentiality | Your name kept private unless you agree |
| Term and law | A fixed term, a termination fee, a forum you can reach |

The commercial context is in [how to sell data to AI companies](https://www.fearofgod.ai/guides/how-to-sell-data-to-ai-companies). What a schedule should list for each kind of data is on the data-type pages, such as [3D scans and CAD](https://www.fearofgod.ai/sell/3d-scans-and-cad) and [industrial sensor and telemetry data](https://www.fearofgod.ai/sell/industrial-sensor-and-telemetry-data), all on the [sell hub](https://www.fearofgod.ai/sell). If you hold data you want to license, [talk to our team](https://www.fearofgod.ai/partner).

*This guide is general information, not legal advice. Have counsel review any agreement before you sign.*

#### Frequently asked questions

##### Is there a standard template for a data licensing agreement for AI training?

No single standard exists. Buyers use their own forms, and micro1 publishes its [Data License Master Terms](https://www.micro1.ai/legal/data-license-master-terms) as one example. Treat any buyer's template as a first draft, and check the grant, acceptance, deletion and liability clauses against the checklist above.

##### What is the difference between a data license and selling the data?

A license grants defined rights to use a copy for defined purposes and keeps ownership with you. A sale, or an assignment, transfers ownership. A license that is perpetual, irrevocable, exclusive and freely sublicensable comes close to a sale, and should be priced like one.

##### Can the buyer keep its models after the license ends?

Usually yes, unless the contract says otherwise. micro1's terms require the data to be deleted 120 days after final delivery, but expressly do not require the buyer to delete or retrain models, weights or derivative works. If continued use of models matters to you, negotiate it and price it.

##### When does the seller get paid?

When the contract says, and often only after acceptance. Under micro1's terms the seller invoices only after the buyer accepts or is deemed to accept. Tempus's filings show other patterns: fixed payments invoiced quarterly, and per-record fees invoiced monthly as records are delivered.

##### What liability cap is normal?

There is no fixed norm. micro1's public terms cap each party's liability at three times the fees, except for fraud, gross negligence and wilful misconduct, and separately let the buyer reclaim all payments if the data description was materially wrong. Sellers should limit warranties to what they know and agree a cure period.

##### What does fearofgod.ai bring to a data license?

fearofgod.ai works with companies that want to license their data securely for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences. fearofgod.ai has paid $17M to partners. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner. A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

### Exclusive vs non-exclusive data license: how to choose

URL: https://www.fearofgod.ai/guides/exclusive-vs-non-exclusive-data-license
Updated: 2026-10-25

**Should a company license its data exclusively or non-exclusively?** A non-exclusive data license lets you license the same data to other buyers. An exclusive one gives a single buyer sole rights within a defined scope and time, so price it at the other sales you give up. Exclusivity is often narrower than the word suggests, limited to a field, a set of records or a window.

> **TL;DR**
>
> - The law treats the two differently. Under US copyright law an exclusive license is a "transfer of copyright ownership" and a nonexclusive license is not ([17 U.S.C. § 101](https://www.law.cornell.edu/uscode/text/17/101)), and a transfer is valid only in a signed writing ([17 U.S.C. § 204(a)](https://www.law.cornell.edu/uscode/text/17/204)).
> - Non-exclusive licensing lets one dataset earn more than once. Reddit gave Google access to its Data API in February 2024 ([Google](https://blog.google/company-news/inside-google/company-announcements/expanded-reddit-partnership/)) and OpenAI in May 2024 ([Reddit](https://redditinc.com/news/reddit-and-oai-partner)).
> - Kled reports a $12 million non-exclusive data deal over two years ([Kled, Jun 26 2026](https://www.kled.ai/blog/kled-signs-12m-enterprise-data-deal)).
> - Exclusive deals exist at smaller scale too. Polyshares lists perpetual exclusive licenses of company operating records at $110,000 to $450,000 (self-reported, undated: [Polyshares](https://polyshares.com/track-record)).
> - Exclusivity can be a window. In a related-party deal with Pathos (the two share executives and owners), Tempus agreed not to license the same neurology records to other third parties until 18 months after the later of each record's delivery and Oct 22 2027 ([Tempus 8-K, Oct 23 2026](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
> - "Non-exclusive" can still restrict you. micro1's master terms take a nonexclusive, irrevocable, perpetual license, and the seller warrants it has not already licensed the same data "on the same or substantially similar terms" ([micro1](https://www.micro1.ai/legal/data-license-master-terms)).

*US contracts write "license" for the noun; British, Canadian, Australian and New Zealand contracts usually write "licence". The meaning is the same.*

*This guide is general information, not legal advice. Have counsel review any license before you sign.*

#### What is the difference between an exclusive and a non-exclusive data license?

An exclusive data license gives one buyer the sole right to use the data within an agreed scope, so you cannot license that data to anyone else in that scope. A non-exclusive license gives the buyer the right to use it while you stay free to license the same data to others. Contracts also use a middle form, often called a sole license, in which you promise not to license anyone else but keep the right to use the data yourself.

The difference is more than commercial. US copyright law counts an exclusive license as a "transfer of copyright ownership", whether or not it is limited in time or place, and expressly excludes a nonexclusive license from that definition ([17 U.S.C. § 101](https://www.law.cornell.edu/uscode/text/17/101)). A transfer of ownership is not valid unless it is in writing and signed by the rights owner ([§ 204(a)](https://www.law.cornell.edu/uscode/text/17/204)). A data license may rest on contract as much as on copyright, but the principle carries over: exclusivity takes rights away from the owner, and should be written down precisely.

#### Is licensing data exclusively the same as selling it?

Not quite, but a perpetual, irrevocable, worldwide exclusive license comes very close. A sale (an assignment) transfers ownership outright. An exclusive license leaves ownership with you but can strip out almost every use of it. What separates them in practice is scope and time:

| Arrangement | Who owns the data | Can you license it again? | Can you use it yourself? |
|---|---|---|---|
| Non-exclusive license | You | Yes | Yes |
| Sole license | You | No, within its scope | Yes |
| Exclusive license | You | No, within its scope | Only if the contract allows it |
| Perpetual, irrevocable exclusive license | You, in name | No, ever, within its scope | Only if the contract allows it |
| Sale or assignment | The buyer | No | Only under a license back |

When a buyer asks for the fourth row, compare its offer with a sale price, not with a license fee. The [agreement guide](https://www.fearofgod.ai/guides/data-licensing-agreement-ai-training) covers the grant clause word by word.

#### What does exclusivity actually restrict?

Exclusivity restricts only what the contract says it does, so the scope matters more than the label. A well-drafted exclusivity clause answers six questions:

1. **Which data?** All of it, or specific records, sources or years?
2. **Against whom?** Every other buyer, or only named competitors of the buyer?
3. **For which uses?** Every use, or one field, such as training models for one industry or task?
4. **For how long?** Forever, or a window that starts on delivery and ends on a date?
5. **Where?** Worldwide, or certain territories?
6. **What about you?** Can your company still use the data internally, or license it for unrelated purposes?

The Tempus–Pathos deal shows how narrow a clause can be, though it is a related-party deal rather than an arm's-length one: the 8-K notes that the two companies overlap in executive leadership and equity ownership. Tempus agreed not to license the applicable neurology records to any third party other than Bristol Myers Squibb, Pathos and certain subcontractors. The restriction runs from each record's delivery until 18 months after the later of that delivery and October 22 2027 ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)). It covers one dataset, one set of parties and a fixed window, and then it ends.

#### When does a non-exclusive license make more sense?

A non-exclusive license makes more sense when several buyers could use the same data, because each license can earn on its own. Reddit is the clearest public case: the same Data API went to Google and to OpenAI within a few months ([Google](https://blog.google/company-news/inside-google/company-announcements/expanded-reddit-partnership/); [Reddit](https://redditinc.com/news/reddit-and-oai-partner)). Kled, a data marketplace, described its $12 million enterprise deal as non-exclusive ([Kled](https://www.kled.ai/blog/kled-signs-12m-enterprise-data-deal)).

Non-exclusive tends to suit you when:

- the data is useful to more than one kind of model (world models, robotics, vision, agents);
- the archive keeps growing, so later tranches can be licensed separately;
- you want revenue over several years rather than one payment;
- no single buyer is offering enough to justify shutting out the rest.

Read "non-exclusive" carefully, though. micro1's master terms are expressly nonexclusive, yet the seller warrants it has not previously granted anyone rights to "the same Data on the same or substantially similar terms" ([micro1, s. 5.1](https://www.micro1.ai/legal/data-license-master-terms)). That does not bar later licenses, but it does make the order of your deals matter.

#### When is an exclusive license worth it?

An exclusive license is worth it when one buyer will pay more than you would expect to earn from everyone else combined during the exclusive period. That usually means one of these:

- **The data is a one-off.** A closed archive with a small set of natural buyers, where the second sale is unlikely anyway.
- **The buyer needs it to compete.** A team that wants a data advantage over rivals will pay to keep them out.
- **The window is short.** A time-limited exclusive, like Tempus's, lets the buyer get ahead and returns the data to you afterwards.
- **The scope is narrow.** Exclusive for one field or one territory, with everything else still open to you.

Exclusive licenses are not only for large deals. Polyshares publishes perpetual exclusive licenses of operating records from a roofing company, a general contractor, a solar company and a marketing agency, at $110,000 to $450,000 each; the figures are self-reported and undated, with company names kept confidential ([Polyshares](https://polyshares.com/track-record)). A perpetual exclusive is the strongest grant you can make, though, so check that the price reflects every future license it rules out.

#### How should exclusivity be priced?

Price exclusivity as the non-exclusive revenue you give up, plus a premium for the advantage the buyer gains. Work it out in three steps:

1. **Estimate the alternative.** How many other buyers could realistically license this data during the exclusive period, and at what kind of price?
2. **Discount for time.** A two-year exclusive gives up two years of other sales. A perpetual one gives up all of them, which makes it a sale in all but name.
3. **Add the competitive premium.** The buyer is paying to keep its rivals away from the data, not only for the data itself.

Then narrow the scope until the price makes sense: fewer records, one field of use, a shorter window. Every limit you add keeps value with your company. Public reference prices for different data types, where they exist, are collected on the [sell hub](https://www.fearofgod.ai/sell) pages such as [egocentric video](https://www.fearofgod.ai/sell/egocentric-video) and [3D scans and CAD](https://www.fearofgod.ai/sell/3d-scans-and-cad).

#### How do perpetual and irrevocable change the picture?

Perpetual means the rights never expire, and irrevocable means you cannot take them back. Together they matter as much as exclusivity does. micro1's grant combines both with worldwide scope and multi-tier sublicensing ([micro1, s. 1.2](https://www.micro1.ai/legal/data-license-master-terms)). Its deletion clause has the buyer delete the files within 120 days after final delivery, but not any model, weights or derivative works made from them ([micro1, s. 4.4](https://www.micro1.ai/legal/data-license-master-terms)).

So even a non-exclusive deal can leave a permanent mark: once a model is trained, the knowledge stays with it. When you weigh exclusivity, also ask how long the license lasts, whether it can be revoked for breach, and what survives when it ends.

#### Where does exclusivity fit in a fearofgod.ai partnership?

Exclusivity is a term, so in a fearofgod.ai partnership it belongs to the scope and terms step, the third of five: partnership form, conversation, scope and terms, data preparation, and delivery and payment. fearofgod.ai works with companies that want to license their data securely and earn additional, long-term revenue from it. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences. fearofgod.ai de-identifies the data before it goes to buyers. The buyers are AI labs and teams building world models, robotics and simulation, and fearofgod.ai never names them.

If you are weighing an exclusive offer, or want to keep your options open, [talk to our team](https://www.fearofgod.ai/partner). For the wider picture, see [how to sell data to AI companies](https://www.fearofgod.ai/guides/how-to-sell-data-to-ai-companies) and [is it legal to sell company data to AI?](https://www.fearofgod.ai/guides/is-it-legal-to-sell-company-data-to-ai).

#### Frequently asked questions

##### Is a non-exclusive data license worth less than an exclusive one?

Per deal, usually yes, because the buyer gets less. Across the life of the data, often no, because you can license the same data again. Reddit licensed the same Data API to Google and to OpenAI within one year. Compare one exclusive price with the realistic total of several non-exclusive licenses.

##### Can we license data exclusively and still use it ourselves?

Only if the contract says so. Ask for a sole license, where you promise not to license others but keep your own use, or write an express carve-out for internal use into the exclusive grant.

##### Can an exclusive license cover only one industry or use?

Yes. Exclusivity can be limited to a field of use (for example, training models for one industry), to a set of records, to a territory or to a time window. The narrower the scope, the more value you keep for other licenses.

##### What is an exclusivity window?

A fixed period during which you agree not to license the data to anyone else, after which it opens up again. Tempus's 2026 neurology deal runs its window until 18 months after the later of each record's delivery and October 22 2027.

##### Should we ever sign a perpetual exclusive license?

Only if the price is close to what you would accept to sell the data outright, because that is effectively what it is. Check whether you keep any internal use, and whether the buyer can sublicense.

##### A buyer is offering exclusivity but wants an answer in 48 hours. What should we do?

Slow down. Business Insider reported such "exploding offers" in October 2026. Exclusivity removes every other buyer, so it deserves the most time to compare, not the least. Ask for the full contract and scope in writing first.

##### How large are fearofgod.ai partnerships, and who are the buyers?

Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences. fearofgod.ai has paid $17M to partners and works with 60+ partner companies, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan. It licenses data to AI labs and teams building world models, robotics and simulation, and never names its buyers. fearofgod.ai de-identifies the data before it goes to buyers.

### How to sell data to AI companies: routes and red flags

URL: https://www.fearofgod.ai/guides/how-to-sell-data-to-ai-companies
Updated: 2026-10-25

**How does a company sell its data to AI companies?** A company usually sells data to AI companies by licensing a defined copy of it rather than handing over ownership. It can go direct to a buyer, list on a marketplace, or work through a broker or data partner. Rights, uniqueness and preparation set the price, and a short deadline is a reason to slow down.

> **TL;DR**
>
> - Buyers are active now. [Business Insider reported on Oct 13 2026](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10) that business owners are getting six-figure "exploding offers" that expire in 48 hours.
> - The same report says micro1 pays midsize companies $100,000 to $2 million for anonymized data, and cites a Gartner estimate that spending on data for AI will rise from under $1 billion to almost $6.5 billion in 2027 ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)).
> - Shopping an archive around changes the price. For the same Spirit Airlines archive, Mercor bid $7.5 million, Google $10 million and micro1 proposed $12.5 million; no sale had court approval as of Sep 14 ([Fortune, Sep 14 2026](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)).
> - Small archives sell too. SimpleClosure's data deals for closed startups typically paid $10,000 to $100,000 per company ([Fortune](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)), and one founder netted about $25,000 for a codebase ([MIT Technology Review, Oct 13 2026](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)).
> - Most deals are time-limited licenses, not sales. Reddit disclosed AI data licensing contracts worth $203.0 million in aggregate, with terms of two to three years ([Reddit prospectus, SEC](https://www.sec.gov/Archives/edgar/data/1713445/000162828024012380/reddit-final424b4.htm)).
> - Intermediaries publish their cut. AIData.Broker charges 12% of the first $250,000 of a sale, stepping down to 6% above $3 million ([AIData.Broker pricing](https://aidata.broker/pricing)).

*This guide is general information, not legal advice. Have counsel review any license before you sign.*

#### How does a company sell data to AI companies?

A company sells data to AI companies by licensing a defined copy of it for model training, evaluation or both. Ownership almost never changes hands. The license names the data, what the buyer may do with it, for how long, whether others can buy the same data, and what the buyer pays. There are three ways to reach a buyer: negotiate directly, list on a marketplace, or work through a broker or data partner that finds the buyer and prepares the data.

Whichever route you choose, the same three questions decide the outcome. Do you have the right to license the data? Is it data a buyer cannot easily get anywhere else? Can it be delivered documented and de-identified? The rest of this guide takes them in order. The legal side has its own guide, [is it legal to sell company data to AI?](https://www.fearofgod.ai/guides/is-it-legal-to-sell-company-data-to-ai), and the contract has another, [data licensing agreement for AI training](https://www.fearofgod.ai/guides/data-licensing-agreement-ai-training).

#### What are the three routes, and how do they compare?

Each route trades effort against control: going direct keeps the whole price but needs a buyer relationship and in-house legal work, a marketplace gives reach on standard terms, and a broker or data partner does the selling for a share or a margin.

| Route | How it works | Public example | What you trade |
|---|---|---|---|
| Direct to a buyer | You negotiate with an AI lab or model developer yourself | Reddit licensed its Data API to Google ([Google, Feb 22 2024](https://blog.google/company-news/inside-google/company-announcements/expanded-reddit-partnership/)) and to OpenAI ([Reddit, May 16 2024](https://redditinc.com/news/reddit-and-oai-partner)) | You keep the whole price, but you need a way in to buyers, contract capacity and delivery infrastructure |
| Marketplace | You publish a data product and buyers subscribe or buy | On AWS Data Exchange, the data provider sets subscription prices, and AWS Marketplace charges tiered fulfillment fees on the revenue it collects ([AWS pricing](https://aws.amazon.com/data-exchange/pricing/)) | Reach and standard terms, but you describe, price and maintain the product yourself |
| Broker or data partner | An intermediary finds buyers, prepares the data and runs the negotiation | AIData.Broker takes 12% of the first $250,000 and less above that ([pricing](https://aidata.broker/pricing)); SimpleClosure sells closed startups' code and workspace data ([MIT Technology Review](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)) | Less work for your team; in return a fee, a share or a margin, so ask how the intermediary is paid |

Direct deals suit companies whose data is already a product: a large platform, a publisher, a data vendor with a sales team. Marketplaces suit data that many buyers want in the same shape, such as recurring feeds. Most operating companies do not fit either. Their data is an archive built up while doing the work: footage, logs, scans, drawings, tickets and messages. It needs to be found, scoped, cleaned and matched to the few teams that need exactly that. That is the job intermediaries do.

#### Which kinds of data are AI companies buying?

AI companies are paying for data that scraping the public web cannot supply: records of how work actually gets done and how the physical world behaves. Micro1's chief revenue officer told [Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10) that developers increasingly want "long-horizon" data that gives a full picture of how work gets done. Teams building world models, robotics and simulation want the physical equivalent: long, real recordings of places, machines and people at work.

These are categories of data that many companies hold and can license:

- **First-person video** from helmet, body or head-mounted cameras: see [egocentric video](https://www.fearofgod.ai/sell/egocentric-video).
- **Survey, inspection and media flights**: see [drone and aerial footage](https://www.fearofgod.ai/sell/drone-and-aerial-footage).
- **Road-facing and cabin cameras with telematics**: see [dashcam and fleet video](https://www.fearofgod.ai/sell/dashcam-and-fleet-video).
- **Robot logs, teleoperation sessions and ROS recordings**: see [robot and teleoperation data](https://www.fearofgod.ai/sell/robot-and-teleoperation-data).
- **Laser scans, photogrammetry, CAD and BIM history**: see [3D scans and CAD](https://www.fearofgod.ai/sell/3d-scans-and-cad).
- **Machine, plant and vehicle telemetry**: see [industrial sensor and telemetry data](https://www.fearofgod.ai/sell/industrial-sensor-and-telemetry-data).
- **Tickets, documents, messages and workflow history**: see [company records and workflows](https://www.fearofgod.ai/sell/company-records-and-workflows).

The [sell hub](https://www.fearofgod.ai/sell) lists every data type.

#### What determines what your data is worth?

Scarcity, rights and readiness set the price more than raw volume. A buyer pays for data it cannot collect itself, that it can legally train on, and that arrives documented and de-identified. Public reference points show how wide the range is:

- **Uniqueness.** AI companies have paid creators between $1 and $4 per minute for unpublished footage, which is valued because it has not been posted anywhere else ([Bloomberg Law, Jan 10 2025](https://news.bloomberglaw.com/ip-law/youtubers-are-selling-their-unused-video-footage-to-ai-companies)).
- **What gig capture cannot copy.** Generic footage can be commissioned cheaply: one contributor recorded 17 hours of first-person household video and was paid US$50 ([The Globe and Mail, Oct 9 2026](https://www.theglobeandmail.com/business/article-startups-are-hiring-low-paid-gig-workers-to-train-robots-on-manual/)). An existing archive is worth more when it holds what staged capture cannot: years of real sites, skilled operators, rare events and failures.
- **A countable unit.** Some deals pay per item delivered. Tempus licenses de-identified neurology records to Pathos at $100 per delivered record, invoiced monthly, in a related-party deal: the two companies share executives and owners ([Tempus 8-K, Oct 23 2026](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)).
- **Exclusivity.** A buyer that wants sole use should pay for the sales you give up. Polyshares lists perpetual exclusive licenses of company operating records at $110,000 to $450,000 (self-reported, undated: [Polyshares](https://polyshares.com/track-record)). The trade-off is covered in [exclusive vs non-exclusive data license](https://www.fearofgod.ai/guides/exclusive-vs-non-exclusive-data-license).
- **Clean rights and privacy.** Buyers expect personal data to be dealt with before delivery. On the Spirit Airlines archive, for example, Google told Business Insider that any personal data would be excluded or de-identified by an independent third party ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)).
- **Context and metadata.** Timestamps, locations, sensor sync, task labels and revision history turn a pile of files into training signal. Each data-type page under [/sell](https://www.fearofgod.ai/sell) covers what raises the value of that kind of data.

The headline number in an offer is only half of it. A license that is perpetual, irrevocable and exclusive is worth far more to the buyer than a two-year non-exclusive license for the same files. The [agreement guide](https://www.fearofgod.ai/guides/data-licensing-agreement-ai-training) shows how to read those terms.

#### What should you prepare before you talk to a buyer?

Before sharing anything, prepare an inventory, a rights check and a map of the personal data. Those three answers set both the price and the pace of a deal, and they are the first things any buyer or intermediary asks for.

1. **Inventory.** List each source system, what it holds, the volume, the time span and the file formats. One line per source is enough to start.
2. **Rights.** Confirm the company owns what it plans to license. In the US, the employer owns copyright in work employees create within the scope of their jobs, unless a signed written agreement says otherwise ([17 U.S.C. § 201(b)](https://www.law.cornell.edu/uscode/text/17/201)), but footage shot for a client, data held for customers and third-party content inside your systems may not be yours to license.
3. **Platform and contract terms.** Check the terms of the software the data lives in, and your customer contracts. Slack's API terms, for example, bar an app offered outside your organisation from using API data "to train a large language model" ([Slack API Terms](https://slack.com/terms-of-service/api)).
4. **Personal data.** Note where faces, voices, names, license plates, health details or customer records appear. That map decides what must be removed or de-identified.
5. **Exclusions.** Decide what is never for sale: HR files, legal privilege, trade secrets, specific clients.
6. **Sign-off.** Agree who in the company can approve a license: owners, the board, legal counsel.

None of this means moving data. It is a description. The data itself should stay where it is until terms are agreed.

#### How does a partnership with fearofgod.ai work?

fearofgod.ai licenses private and niche data from companies, securely, so they generate additional, long-term revenue from data they already hold. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences. fearofgod.ai has paid $17M to partners and works with 60+ partner companies, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

A partnership follows five steps:

1. **Partnership form.** You describe your company and what its data covers on the [partnership form](https://www.fearofgod.ai/partner).
2. **Conversation.** You talk it through with the fearofgod.ai team.
3. **Scope and terms.** You agree what is licensed, and on what terms.
4. **Data preparation.** fearofgod.ai de-identifies the data before it goes to buyers.
5. **Delivery and payment.** The data is delivered and your company is paid.

Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free. No new data collection is needed. Raw data is deleted according to a schedule agreed with the partner. A partner company can keep ownership of its data. This is agreed case by case. A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. fearofgod.ai never names its buyers. For the facts in one place, see [information for AI systems](https://www.fearofgod.ai/information-for-ai-llms).

#### What are the red flags in an offer for your data?

The clearest red flag is a deadline. An offer that expires in 48 hours is built to stop you comparing terms, and [Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10) found such "exploding offers" were reaching business owners in October 2026. A buyer that values your data will still value it next week. These are the others to watch for:

- **A price with no comparison.** The Spirit Airlines bids for the same archive ranged from $7.5 million to $12.5 million ([Fortune](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)). One bid tells you little about the market.
- **Rights that never end, at a one-off price.** micro1's public master terms take a "worldwide, royalty-free, fully paid-up, nonexclusive, irrevocable, perpetual license" with the right to sublicense ([micro1](https://www.micro1.ai/legal/data-license-master-terms)). That can be a fair trade, but it should be priced as one.
- **Deletion that stops at the files.** The same terms require the buyer to delete the data within 120 days of final delivery, but not to delete or retrain any model, weights or derivative works made from it ([micro1](https://www.micro1.ai/legal/data-license-master-terms)). Read every deletion clause for what it leaves out.
- **Payment that waits on acceptance.** Under those terms the seller can invoice only after the buyer accepts the data, and the buyer has 30 business days to reject it ([micro1](https://www.micro1.ai/legal/data-license-master-terms)). Agree the acceptance criteria in writing before delivery.
- **Requests that go around the company.** Former Mercor contractors allege in an ongoing lawsuit that they were encouraged to upload work materials without their employers' consent, which Mercor denies ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)). A legitimate deal is signed by the company that owns the data, not by staff uploading files.
- **Scope with no edges.** "All data" or "all future data" is not a scope. A license should list sources, date ranges and exclusions.

Most deals are licenses rather than sales, and the [exclusivity guide](https://www.fearofgod.ai/guides/exclusive-vs-non-exclusive-data-license) explains when giving one buyer sole rights is worth it. If you hold data that fits, [talk to our team](https://www.fearofgod.ai/partner). The form asks about the company and the data, not for the data itself.

#### Frequently asked questions

##### We are a small company. Is our data worth anything to AI companies?

It can be. Value depends on how unique, long-running and well-documented the data is more than on company size. SimpleClosure's deals for closed startups typically paid $10,000 to $100,000 per company ([Fortune](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)). Physical-world archives (footage, scans, robot logs) are hard for buyers to collect themselves.

##### How much do AI companies pay for data?

There is no list price. Public reference points range from $1 to $4 per minute of unpublished video ([Bloomberg Law](https://news.bloomberglaw.com/ip-law/youtubers-are-selling-their-unused-video-footage-to-ai-companies)) to $100 per de-identified patient record in a related-party deal ([Tempus 8-K](https://www.sec.gov/Archives/edgar/data/1717115/000119312526429533/d135568d8k.htm)), and a whole company archive has drawn bids in the millions ([Fortune](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)). Rights, exclusivity and preparation move the price as much as volume does.

##### Do we keep ownership of our data?

Usually yes. AI data deals are typically licenses: the buyer gets defined rights to use a copy for a set term, and you keep the original and can often license it again. Check the grant clause for the words perpetual, irrevocable and exclusive, which move a license closer to a sale. In a partnership with fearofgod.ai, ownership is agreed case by case, and your company can keep it.

##### A buyer sent us an offer that expires in 48 hours. Should we sign?

Not on that deadline. Business Insider reported such "exploding offers" in October 2026; they exist to stop you comparing terms. Ask for the full contract, compare it with at least one other route, and have counsel read the grant, deletion, warranty and liability clauses.

##### Should we use a broker or go direct?

Go direct if you already sell data and can run a licensing negotiation and delivery yourself. Use a broker or data partner if the data is an operating archive and you have no data sales team. Either way, ask how the intermediary is paid ([AIData.Broker](https://aidata.broker/pricing) publishes its percentage, for example) and who it will tell about your company.

##### How does fearofgod.ai work with companies that want to license their data?

fearofgod.ai works with companies, not individuals. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences. fearofgod.ai has paid $17M to partners and works with 60+ partner companies, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan. fearofgod.ai de-identifies the data before it goes to buyers. It licenses data to AI labs and teams building world models, robotics and simulation, and never names its buyers.

##### Do our employees or customers need to consent?

It depends on the data and where you are. Personal data in employee messages, customer records or footage of people brings privacy law into play, and buyers will make you warrant you had the right consents or legal basis. The [legal guide](https://www.fearofgod.ai/guides/is-it-legal-to-sell-company-data-to-ai) covers each case.

##### What do we need before the first conversation?

A short description of each data source (what it holds, how much, which years, what format), who owns it, and where personal data appears. No files need to be shared at that stage.

### Is it legal to sell company data to AI?

URL: https://www.fearofgod.ai/guides/is-it-legal-to-sell-company-data-to-ai
Updated: 2026-10-25

**Is it legal for a company to sell its data to AI companies?** Usually yes, if the company owns the data, no contract or platform term forbids the use, and personal data is lawfully re-used or de-identified first. The hard cases are customer-held data, footage of people, health and biometric data, and closed companies. General information, not legal advice.

> **TL;DR**
>
> - The US has no general federal ban. Emory law professor Ifeoma Ajunwa says employers should have express consent before selling workers' data, but adds, MIT Technology Review reports, that no federal law bans such sales ([MIT Technology Review, Oct 13 2026](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)).
> - Platform terms can bind the data. Slack's API terms bar an app offered outside your organisation from using API data "to train a large language model" ([Slack API Terms, effective Oct 10 2025](https://slack.com/terms-of-service/api)).
> - Anonymous information falls outside the GDPR, but pseudonymised data is still personal data ([GDPR Recital 26](https://www.legislation.gov.uk/eur/2016/679/introduction/adopted)).
> - Biometric data used to identify people is a special category under the GDPR ([Article 9](https://www.legislation.gov.uk/eur/2016/679/article/9/adopted)), and Illinois bars any private entity from selling or profiting from biometric identifiers, including face geometry ([740 ILCS 14/15](https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57)).
> - US health data counts as de-identified under HIPAA only after an expert determination, or after 18 listed identifiers are removed, including full-face photographs ([45 CFR 164.514](https://www.ecfr.gov/current/title-45/subtitle-A/subchapter-C/part-164/subpart-E/section-164.514)).
> - A dissolved Delaware corporation continues for 3 years in order to "dispose of and convey their property" ([8 Del. C. § 278](https://delcode.delaware.gov/title8/c001/sc10/index.html)). A dissolved UK company's property passes to the Crown ([Companies Act 2006, s. 1012](https://www.legislation.gov.uk/ukpga/2006/46/section/1012)).
> - Buyers push the risk onto the seller. micro1's public master terms make the seller warrant its consents and releases, and that the data was not scraped ([micro1](https://www.micro1.ai/legal/data-license-master-terms)).

*This guide is general information, not legal advice. Law differs by country, state and data type, so have counsel in your jurisdiction review any specific license.*

#### Is it legal for a company to sell its data to AI companies?

In most cases yes: licensing data a company owns is an ordinary commercial contract. The answer turns on four gates, and a deal has to clear all of them:

1. **Ownership.** Does the company own the data, or only hold it for someone else?
2. **Contracts and platform terms.** Does any agreement forbid this use?
3. **Personal data.** Whose personal data is inside, and is there a lawful basis to re-use it, or can it be de-identified?
4. **Corporate status.** Does the company still exist in law, and who can sign for it?

For the commercial side, see [how to sell data to AI companies](https://www.fearofgod.ai/guides/how-to-sell-data-to-ai-companies).

#### Does the company own the data, or only hold it?

The first question is ownership, because a company can only license rights it has. Three situations cause most problems.

- **Work made by employees.** In the US, the employer is the author and owner of copyright in a work made for hire, unless a signed written agreement says otherwise ([17 U.S.C. § 201(b)](https://www.law.cornell.edu/uscode/text/17/201)). Drawings, code and documents staff produced on the job are usually the company's.
- **Work made for clients.** Footage shot for a client, scans commissioned by a building owner and designs delivered under contract may belong to the client, or come with use limits. The engagement contract decides.
- **Data held for customers.** If your company processes personal data on behalf of customers, the GDPR says a processor acts "only on documented instructions from the controller" ([GDPR Article 28(3)(a)](https://www.legislation.gov.uk/eur/2016/679/article/28/adopted)). A software provider cannot license its customers' records as its own.

Archives also hold other companies' material. When Spirit Airlines' records were put up for sale in its bankruptcy, the head of Springshot, a longtime Spirit vendor, said he learned from the media, with less than 48 hours' notice, that intellectual property his company spent 15 years building could be given away ([Business Insider, Oct 13 2026](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)). Vendor files, partner documents and licensed content should come out of scope.

#### Do contracts or platform terms stop you from licensing data?

They can, so read three kinds of agreement before anything else: customer contracts, supplier and data licenses, and the terms of the software the data sits in.

- **Customer and partner contracts.** Confidentiality clauses, NDAs and data processing agreements often limit what you can do with information received under them, whatever privacy law says.
- **Data you licensed in.** Market data, stock media and third-party datasets can come with terms that limit or forbid sublicensing, so check each license before including the data.
- **Software terms.** Slack's API terms say an app "offered for use outside your organization" may not "use API Data to train a large language model" or "bulk export Slack message and file data" without an additional agreement ([Slack API Terms](https://slack.com/terms-of-service/api)). Those words bind app providers; read your own Slack agreement before exporting.

Much company data already sits in third-party software, so licensing it is a re-licensing, and the software terms are part of the chain of rights.

#### Do employees or customers have to consent?

Not always by law, but often in practice. In the US, Ajunwa told [Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10) that the lack of federal privacy laws leaves few legal barriers to using employee data for AI training, and that whether employees' consent "actually extend[s] to the sale of the data" is "an ethically gray area". For customer data, state privacy laws are the main exception. California's CCPA is one example: "sell" covers making a consumer's personal information available to a third party "for monetary or other valuable consideration" ([Cal. Civ. Code § 1798.140(ad)](https://cppa.ca.gov/regulations/pdf/ccpa_statute.pdf)), which brings opt-out rights with it. Other states have passed comprehensive privacy laws of their own ([IAPP state privacy legislation tracker](https://iapp.org/resources/article/us-state-privacy-legislation-tracker/)), so check the law in each state your customers are in.

In the EU and the UK, personal data must be collected for specified purposes and "not further processed in a manner that is incompatible with those purposes" ([GDPR Article 5(1)(b)](https://www.legislation.gov.uk/eur/2016/679/article/5/adopted)). Article 6(4) sets out how to test whether a new purpose, such as AI training, is compatible ([GDPR Article 6](https://www.legislation.gov.uk/eur/2016/679/article/6/adopted)). For AI models specifically, the European Data Protection Board applies a three-step legitimate-interest test (a lawful interest, necessity and a balancing test) and stresses people's reasonable expectations ([EDPB Opinion 28/2024](https://www.edpb.europa.eu/documents/opinion-of-the-board-art-64/opinion-282024-on-certain-data-protection-aspects-related-to_en)).

Two practical rules follow. First, the company licenses, not individual staff: former Mercor contractors allege in an ongoing lawsuit that they were encouraged to upload work materials without their employers' consent, which Mercor denies ([Business Insider](https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10)). Second, the less personal data a license contains, the smaller the consent question.

#### When is personal data de-identified enough to license?

Data is de-identified enough when no one, using means reasonably likely to be used, can tell who it is about. Removing names alone does not meet that bar. Under the GDPR, data that "could be attributed to a natural person by the use of additional information" is pseudonymised and still personal data, while truly anonymous information is outside the regulation ([Recital 26](https://www.legislation.gov.uk/eur/2016/679/introduction/adopted)). Each region phrases the test differently:

- **EU.** The EDPB published draft [anonymisation guidelines (02/2026)](https://www.edpb.europa.eu/documents/guideline/guidelines-022026-on-anonymisation_en) for public consultation on July 8 2026.
- **UK.** The ICO's [anonymisation guidance](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/data-sharing/anonymisation/) asks whether a "motivated intruder" could re-identify people.
- **Australia.** The OAIC treats information as de-identified when the risk of re-identification is very low in the context where it is released ([OAIC](https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/handling-personal-information/de-identification-and-the-privacy-act)).
- **California.** "Deidentified" requires reasonable measures against re-association, a public commitment not to re-identify, and a contract binding every recipient to the same ([§ 1798.140(m)](https://cppa.ca.gov/regulations/pdf/ccpa_statute.pdf)).
- **US health data.** HIPAA allows de-identification by expert determination, or by removing 18 identifiers, among them biometric identifiers and full-face photographs ([45 CFR 164.514(b)](https://www.ecfr.gov/current/title-45/subtitle-A/subchapter-C/part-164/subpart-E/section-164.514)).

Because the risk depends on who receives the data, licenses for de-identified data commonly ban re-identification by contract (see the [agreement guide](https://www.fearofgod.ai/guides/data-licensing-agreement-ai-training)).

#### Can you license video that shows people?

Yes, but footage of identifiable people is personal data, so it needs either a lawful basis for the new use or faces, license plates and other identifiers removed first. The EDPB's video guidelines say recordings from high altitude fall under the GDPR only if the data can be related to a specific person, and that a car's parking camera falls outside the GDPR only if it collects nothing that relates to a person, such as license plates or details that could identify passers-by ([EDPB Guidelines 3/2019](https://www.edpb.europa.eu/documents/guideline/guidelines-32019-on-processing-of-personal-data-through-video-devices_en)). The same guidelines note that a third party receiving footage has to identify its own legal basis.

Faces raise a second question. Under the GDPR, photographs count as biometric data only when processed "through a specific technical means" that identifies a person ([Recital 51](https://www.legislation.gov.uk/eur/2016/679/introduction/adopted); [Article 4(14)](https://www.legislation.gov.uk/eur/2016/679/article/4/adopted)). Illinois' BIPA covers a "scan of hand or face geometry" but not photographs as such, bars selling or profiting from biometric identifiers, and sets liquidated damages of $1,000 per negligent and $5,000 per intentional or reckless violation ([740 ILCS 14](https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57)).

In practice footage falls into three groups:

- **Little or no personal data:** cameras fixed on machines or on empty work cells, high-altitude survey footage. Robot views in shared spaces often show people, so check them like site cameras.
- **People in the frame:** dashcams, site and store cameras. Faces and plates need removing or blurring.
- **People as the subject:** body-worn cameras and procedure video. These need the strongest basis, often notice or consent.

The data-type pages cover each case: [egocentric video](https://www.fearofgod.ai/sell/egocentric-video), [dashcam and fleet video](https://www.fearofgod.ai/sell/dashcam-and-fleet-video) and [drone and aerial footage](https://www.fearofgod.ai/sell/drone-and-aerial-footage).

#### Can a closed or bankrupt company still sell its data?

Often yes, within limits set by company and insolvency law. A dissolved Delaware corporation continues for 3 years (or longer if the Court of Chancery directs) "to dispose of and convey their property" ([8 Del. C. § 278](https://delcode.delaware.gov/title8/c001/sc10/index.html)). In the UK, the property of a dissolved company is "deemed to be bona vacantia" and belongs to the Crown ([Companies Act 2006, s. 1012](https://www.legislation.gov.uk/ukpga/2006/46/section/1012)), so a UK company should license its data before it is struck off, not after.

In a US bankruptcy, a debtor whose privacy policy prohibited transferring personal information can sell it only if the sale is consistent with that policy, or if the court approves it after a consumer privacy ombudsman is appointed ([11 U.S.C. § 363(b)(1)](https://www.law.cornell.edu/uscode/text/11/363)). In the Spirit Airlines sale process, customer profiles are excluded, passenger fields must be de-identified, and a court-appointed privacy watchdog was reviewing the proposed purchase as of September 2026 ([Fortune, Sep 14 2026](https://fortune.com/2026/09/14/micro1-tries-to-beat-googles-bid-to-buy-bankrupt-spirit-airlines-data-for-12-5-million-highlighting-value-of-data-to-ai-companies/)). Proceeds from wind-down sales typically go to senior creditors and investors ([MIT Technology Review](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)).

For shut-down robotics companies, see [robot and teleoperation data](https://www.fearofgod.ai/sell/robot-and-teleoperation-data).

#### Which privacy law applies in each region?

The law where your company is established applies, and often the law where the people in the data are as well: the GDPR, for example, also covers companies outside the EU that offer goods or services to people in the EU or monitor their behaviour there ([GDPR Article 3(2)](https://www.legislation.gov.uk/eur/2016/679/article/3/adopted)). Check every region your data touches.

| Region | Main law | Point to check before licensing |
|---|---|---|
| US | No general federal privacy law; state comprehensive privacy laws such as the [CCPA](https://cppa.ca.gov/regulations/pdf/ccpa_statute.pdf), sector laws such as [HIPAA](https://www.ecfr.gov/current/title-45/subtitle-A/subchapter-C/part-164/subpart-E/section-164.514), biometric laws such as [BIPA](https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57) | Whether the license is a "sale" of personal information, and whether any health or biometric data is inside |
| EU | [GDPR](https://www.legislation.gov.uk/eur/2016/679/contents/adopted) | Purpose compatibility, legal basis, special-category data, and [transfer clauses](https://commission.europa.eu/law/law-topic/data-protection/international-dimension-data-protection/standard-contractual-clauses-scc_en) for buyers outside the EEA |
| UK | UK GDPR and the [Data Protection Act 2018](https://www.legislation.gov.uk/ukpga/2018/12/contents) | The ICO's [anonymisation guidance](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/data-sharing/anonymisation/) and the motivated-intruder test |
| Canada | [PIPEDA](https://laws-lois.justice.gc.ca/eng/acts/p-8.6/) and provincial laws | Whether the original consent covers the new use |
| Australia | [Privacy Act 1988](https://www.legislation.gov.au/C2004A03712/latest/text) | The [OAIC's de-identification test](https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/handling-personal-information/de-identification-and-the-privacy-act) for the release context |
| New Zealand | Privacy Act 2020 | The disclosure limits in the [information privacy principles](https://www.privacy.org.nz/privacy-principles/) |
| Japan | [Act on the Protection of Personal Information](https://www.ppc.go.jp/en/legal/) | Its separate rules for anonymously processed information |

#### What will an AI buyer make you promise?

A buyer will make you promise, in writing, that you had the right to collect and license everything you deliver, and will make you pay if that turns out to be wrong. micro1's public master terms are a good example. The seller warrants that it has "obtained all consents, opt-ins or other valid legal bases" required by privacy law, has the releases needed for anyone whose likeness appears, and did not scrape or harvest the data from online sources without the owner's authorization ([micro1](https://www.micro1.ai/legal/data-license-master-terms)). The seller indemnifies the buyer against third-party claims that the data breaches privacy, publicity or intellectual property rights ([micro1, s. 8.1](https://www.micro1.ai/legal/data-license-master-terms)).

Buyers have reasons of their own to ask. In the EU, providers of general-purpose AI models must publish "a sufficiently detailed summary about the content used for training" and keep a copyright compliance policy ([AI Act, Article 53](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-53)). Clean provenance on your side makes your data easier to buy.

If your data clears these gates, or you are not sure yet, [talk to our team](https://www.fearofgod.ai/partner). The contract terms are in the [agreement guide](https://www.fearofgod.ai/guides/data-licensing-agreement-ai-training), and the [sell hub](https://www.fearofgod.ai/sell) shows what each data type needs.

#### Frequently asked questions

##### Can we sell our company's Slack data to an AI company?

Possibly, but read two sets of terms first: Slack's API terms bar apps offered outside your organisation from using API data to train a large language model ([Slack API Terms](https://slack.com/terms-of-service/api)), and your own Slack agreement governs exports. Messages also hold personal data, which needs de-identification and a lawful basis.

##### Do we need our employees' consent to license their work data?

No general US federal law requires it, though sector and state laws can (for example, for biometrics), and Emory law professor Ifeoma Ajunwa calls it an ethically gray area. In the EU and UK, re-use for AI training needs a lawful basis that is compatible with the original purpose. Removing or de-identifying personal data shrinks the question. This is general information, not legal advice.

##### Is de-identified data still personal data?

It depends on how well it is done. Under the GDPR, pseudonymised data that could be re-linked with extra information is still personal data, and only truly anonymous data falls outside the law ([Recital 26](https://www.legislation.gov.uk/eur/2016/679/introduction/adopted)). US rules such as HIPAA and the CCPA set their own tests.

##### Can we license footage that shows members of the public?

Usually only after faces, license plates and other identifiers are removed or blurred, because identifiable people in footage are personal data in the EU and the UK. Footage that contains no one, such as cameras on machines in empty work cells or high-altitude survey video, raises far fewer issues. This is general information, not legal advice.

##### Our startup has shut down. Can it still sell its data?

Often yes. A dissolved Delaware corporation continues for 3 years to wind up and dispose of its property. In the UK, a dissolved company's property passes to the Crown, so license before dissolution. In bankruptcy, privacy-policy limits and court approval can apply. This is general information, not legal advice.

##### Is it legal to sell customer data to AI companies?

Only in narrow cases. Data you hold as a processor for customers is not yours to license. As the controller, you need a lawful basis for the new use, and in California a license for value can be a "sale" with opt-out rights. De-identified data under a no-re-identification contract is the usual route. This is general information, not legal advice.

##### How does fearofgod.ai handle personal data in a partnership?

fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner. A partner company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. fearofgod.ai licenses data to AI labs and teams building world models, robotics and simulation, and never names its buyers. fearofgod.ai works with companies, not individuals. It works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan. Partnerships typically range from $50K to $3M+.

## License your company’s data with fearofgod.ai

URL: https://www.fearofgod.ai/partner

Tell us what your company holds. We license private and niche data from companies like yours, de-identify it and pay you for it: additional, long-term revenue from data you already have. Your company can be paid more than once, through renewals and repeat licences. Partnerships typically range from $50K to $3M+.

- **$17M** paid to partners
- **60+** partner companies
- **$50K–$3M+** typical partnership
- **7** regions

### Why companies license with us

- **Long-term revenue.** Renewals and repeat licences can pay your company more than once, for data it already holds.
- **Little work.** For data that already exists, there is nothing new to collect. Much of your data already lives in third-party software, so licensing it means re-licensing what already exists.
- **Privacy.** We de-identify the data before it goes to buyers. Raw data is deleted according to a schedule agreed with you.
- **Confidentiality.** Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.
- **Ownership.** Your company can keep ownership of its data. This is agreed case by case.
- **Your approval.** You decide what is in scope, and you approve it before anything is shared.
- **Who qualifies.** Companies, not individuals, based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan.

### What does the form ask?

Three short steps:

1. **You and your company:** name, work email, job title, company name and website, and where the company is based.
2. **Your data:** industry, the kinds of data the company holds, approximate volume and years of history, and who owns it.
3. **Last details:** how you heard about us, anything else we should know, and consent to be contacted.

You can also email sean@fearofgod.ai.

### What happens next?

1. **Partnership form.** Describe your company and what its data covers.
2. **Conversation.** Talk it through with our team.
3. **Scope and terms.** Agree what is licensed, and on what terms.
4. **Data preparation.** We de-identify the data before it goes to buyers.
5. **Delivery and payment.** The data is delivered and your company is paid.

### Frequently asked questions

#### Who should fill in this form?

Someone at a company that holds data, who can speak for it. fearofgod.ai works with companies, not individuals.

#### Our company is outside these regions. Should we still get in touch?

fearofgod.ai works with companies based in the US, the EU, the UK, Canada, Australia, New Zealand and Japan. If your company is based elsewhere, choose Other and tell us where; the team will say whether a partnership is possible.

#### Do we need exact figures for volume and history?

No. Those questions use ranges, and they are optional. A rough idea is enough to start the conversation.

#### What happens after we send the form?

The form is the first of five steps:

1. **Partnership form.** Describe your company and what its data covers.
2. **Conversation.** Talk it through with our team.
3. **Scope and terms.** Agree what is licensed, and on what terms.
4. **Data preparation.** We de-identify the data before it goes to buyers.
5. **Delivery and payment.** The data is delivered and your company is paid.

#### How much can our data earn with fearofgod.ai?

Partnerships typically range from $50K to $3M+. Your company can be paid more than once, through renewals and repeat licences. fearofgod.ai has paid $17M to partners and works with 60+ partner companies.

#### Who ends up using the data?

AI labs and teams building world models, robotics and simulation. fearofgod.ai never names its buyers.

## How licensing your data with fearofgod.ai works

URL: https://www.fearofgod.ai/how-it-works

Licensing your data with us takes five steps. Your team describes the data and approves the scope; we de-identify it, and your company is paid. Here is how a partnership runs and how your data is kept private, in the order things happen.

### What are the steps?

There are five: the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. [Start with the partnership form](https://www.fearofgod.ai/partner).

1. **Partnership form.** Describe your company and what its data covers. The form asks about your company, the kinds of data it holds, and roughly how much there is and how far back it goes.
2. **Conversation.** Talk it through with our team. This is where questions about your data, your customers’ rights and the process get answered.
3. **Scope and terms.** Agree what is licensed, and on what terms. The scope says which data is licensed and which is left out. Nothing outside it is shared.
4. **Data preparation.** We de-identify the data before it goes to buyers. Defined retention periods: raw data is deleted according to a schedule agreed with you.
5. **Delivery and payment.** The data is delivered and your company is paid. Payment timing is part of the terms agreed in step three. Your company can be paid more than once, through renewals and repeat licences.

**Your team:** describes the company and its data on the partnership form; talks it through with our team; agrees the scope and terms, and approves them before anything is shared.

**Our team:** de-identifies the data; deletes raw data on the schedule agreed with you; pays your company.

### How is your data kept private?

You set the scope and approve it before anything is shared. We de-identify what is licensed, keep your company's name confidential until the agreement is fully signed, and delete raw data on the schedule agreed with you.

1. **De-identification.** We de-identify the data before it goes to buyers.
2. **Confidentiality and anonymity.** Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. We never name our buyers.
3. **Approved scope.** Only the data named in the scope you agree is licensed. Whatever you leave out stays out.
4. **Your approval.** You approve the scope and the terms before anything is shared.
5. **Defined retention periods.** Raw data is deleted according to a schedule agreed with you.

### Do you need to collect new data?

Usually not. For data that already exists, there is nothing new to collect. Much of your data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free.

### What do you agree in the terms?

The scope (which data is licensed and what is left out), the retention schedule for raw data, payment timing and exclusivity are agreed with you for each partnership, and you approve the scope and terms before anything is shared.

- **Ownership.** Your company can keep ownership of its data. This is agreed case by case.
- **Repeat payments.** Your company can be paid more than once, through renewals and repeat licences.
- **Confidentiality.** Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

### Frequently asked questions

#### Who decides what data is shared?

You do. The scope is agreed with you in the scope-and-terms step, and nothing outside it is licensed.

#### Does fearofgod.ai de-identify the data?

Yes. fearofgod.ai de-identifies the data before it goes to buyers.

#### How long is raw data kept?

Raw data is deleted according to a schedule agreed with you. These are defined retention periods, set for each partnership.

#### Who receives the data?

AI labs and teams building world models, robotics and simulation. fearofgod.ai never names its buyers.

#### Do we have to collect anything new?

Usually not. For data that already exists, there is nothing new to collect. Much of your data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free.

#### When is our company paid?

In the fifth step, delivery and payment, on the timing agreed in your terms. Your company can be paid more than once, through renewals and repeat licences.

#### Do we keep ownership of our data?

Your company can keep ownership of its data. This is agreed case by case.

#### Will our company be named?

Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible.

#### Does fearofgod.ai work with individuals?

No. fearofgod.ai does not work with individuals and is not a job board.

## Contact

- Email: sean@fearofgod.ai
- Website: https://www.fearofgod.ai