Every figure on this page is a dated statement by xAI, SpaceX, a court record or a regulator. The profile follows the same method as every buyer profile in this independent reference on AI training data licensing.
Each number carries its source and check date, and how DataPacts sources and dates every figure is set out in the methodology.
What xAI Acquires for Grok: Licensed Text and Images, Web, Users and Cursor Data
SpaceXAI's public summary for Grok 4.7, version 1 of 2026-09-21, lists the routes by which training content reached the model, and only one of them, commercial licenses with rightsholders, is stated for text and images alone.
- Licensed content: commercial licenses with rightsholders, text and images only.
- Private datasets: third-party datasets, all modalities.
- Web crawling: the SpaceXAI Web Crawler.
- Public datasets: Common Corpus (Pleias), OpenAlex metadata and Crossref metadata.
- Users: user interaction data, and X data of users who did not opt out.
- Synthetic data: generated with xAI's own models, such as Grok 4.6.
- Developer data: anonymized Cursor workflow data.
SpaceXAI states size bands of more than 10 trillion tokens of text, more than 1 billion images, and more than 1 million hours each of audio and video. These are the company's claims. The Commission template asks only for Yes or No plus modality for licensed content, so no licensor is named.
For a seller, the summary acknowledges that licensed text and images and private datasets reach Grok, but it names no price, counterparty or application route. The wider map of how AI companies get training data places this mix beside the other labs.
xAI Company Identity: SpaceXAI LLC and Its SpaceX Parent
The provider of Grok is SpaceXAI LLC, formerly xAI LLC, of 1450 Page Mill Rd, Palo Alto, a subsidiary of Space Exploration Technologies Corp. since the merger took effect on 2026-02-02.
The table lists the identity attributes on record as of 2026-11-09.
| Attribute | Value | Source |
|---|---|---|
| Provider (legal name) | SpaceXAI LLC, formerly xAI LLC | Grok 4.7 summary |
| Address | 1450 Page Mill Rd, Palo Alto | Grok 4.7 summary |
| Brand and website | xAI, Grok; x.ai | x.ai |
| Parent | Space Exploration Technologies Corp., SEC CIK 0001181412 | SEC 424B4 |
| Merger effective | 2026-02-02 | SEC 424B4 |
| Own registry number | none published in the sources checked | SEC 424B4 |
| EU authorised representative | EDSR, Valukoja 8/2, Tallinn | Grok 4.7 summary |
| Names in the complaints of 2026-11-04 | SpaceX Super Intelligence Corp. (f/k/a SpaceXAI Corp. and X.AI Corp.); SpaceX Super Intelligence LLC (f/k/a X.AI LLC); SpaceX Super Intelligence Holdings LLC | N.D. Cal. complaint |
| Founding date | not recorded | n/a |
The complaints of 2026-11-04 use different names, as the table shows, so a licensor confirms which entity signs. The contracting party in xAI's own terms is not part of the record checked here, and no contracting entity is stated on this page.
xAI Program Terms and Payouts, Dated
xAI publishes no program, application page or terms for licensing company data; the table lists each term a data holder would check and what the record holds, as of 2026-11-09.
| Term | What the record shows | Checked | Source |
|---|---|---|---|
| Company data program or application page | none on record | 2026-11-09 | x.ai |
| License scope, term, exclusivity and deletion for company data | none published | 2026-11-09 | x.ai |
| Payout to data holders | none stated | 2026-11-09 | x.ai |
| Referral fee | none recorded | 2026-11-09 | x.ai |
| Licenses with rightsholders | Yes for text and image; no licensor or value named | 2026-11-09 | Grok 4.7 summary |
| Named data agreement | Cursor (Anysphere), 2026-04-19, compute and option agreement | 2026-11-09 | SEC 424B4 |
| Crawler | SpaceXAI Web Crawler, January 2024 to August 2026 | 2026-11-09 | Grok 4.7 summary |
| Opt-out honored | X users who opted out; crawl instructions against training use | 2026-11-09 | Grok 4.7 summary |
| GPAI Code of Practice copyright commitments | not a signatory | 2026-11-09 | Grok 4.7 summary |
The first four rows read "none on record". That leaves no scope, term, exclusivity, deletion or payment trigger to compare against the program terms compared for buyers that publish them, and the wording does not say xAI never acquires data.
The only disclosed data terms are the Cursor filing and the unnamed text and image licenses. Neither gives a company holder a route to apply.
xAI Data Deals on Record: Cursor, Rightsholder Licenses and Data Sales
One data agreement names a counterparty, the 2026-04-19 agreement between SpaceX, including xAI, and Anysphere, Inc., the maker of Cursor; xAI's licenses with rightsholders are acknowledged but unnamed.
The table lists the three items newest first, as of 2026-11-09, with values as stated by the parties.
| Date | Counterparty or item | What is acquired | Value | Source |
|---|---|---|---|---|
| 2026-09-21 | rightsholders (unnamed) | text and image under commercial licenses (Grok 4.7 summary) | not disclosed | media.x.ai |
| 2026-04-19 | Anysphere, Inc. (Cursor) | developer interaction data (coding prompts, iteration cycles) under a compute and option agreement | termination fee USD 1.5bn plus deferred services fee USD 8.5bn in defined cases (SpaceX 424B4, filed 2026-06-12) | sec.gov |
| Q1 2026 | xAI as seller (SpaceX AI segment) | data licensing arrangements revenue, up USD 12 million against Q1 2025 (SpaceX's statement) | stated increase, not a deal value | sec.gov |
The fee figures are contractual amounts in defined cases, not a price for data.
In the Cursor filing, SpaceX describes a compute and option agreement under which access to Cursor's developer interaction data is expected to enhance model training. SpaceXAI's Grok 4.7 summary states that training data included anonymized Cursor workflow data.
The terms, parties and status of the Cursor and SpaceX AI agreement have their own record.
For rightsholder licenses, the summary answers Yes for text and image and names no licensor, value or term.
xAI is also a seller. SpaceX's filing states an increase of USD 12 million in data licensing arrangements revenue in Q1 2026 against Q1 2025, in its AI segment. That is a revenue change, not a deal value, and makes xAI a possible competitor for some holders.
X Posts and Grok Users as a Source of xAI Training Data
SpaceXAI states for Grok 4.7 that user interaction data was used for training and that X data of X users who opted out was not.
Both statements are xAI's own, dated 2026-09-21. X's own terms are not part of the record checked here, so only the xAI side is stated.
The Irish Data Protection Commission acted twice over Grok training. It made an urgent High Court application against Twitter (now X) on 2024-08-08 under section 134 over training on EU and EEA users' data, and its outcome is not stated here.
It announced a section 110 inquiry into X Internet Unlimited Company on 2025-04-11 over training on EU and EEA users' public posts, and no decision is recorded. The wider regulatory frame is the Data Protection Commission oversight of AI training.
No X or xAI program for licensing content posted on X is on record. Control sits with the user setting.
How xAI Collects Web Content and What It Committed to in the EU
SpaceXAI names one crawler for Grok 4.7, the SpaceXAI Web Crawler, active from January 2024 to August 2026.
No user agent string is recorded, so none is named here. How bots are identified, allowed and blocked in general is covered on the page on AI crawlers.
SpaceXAI states three opt-out measures: it does not train on X data from X users who opted out, it respects instructions that crawled content should not be used for training, and it uses the USTR Notorious Markets list as an exclusion signal.
On the Code of Practice for general-purpose AI, SpaceXAI states on 2026-09-21 that it is not a signatory of the copyright commitments. The Commission page (updated 2026-10-07, checked 2026-11-09) lists xAI for the Safety and Security chapter only, and xAI is not among the 22 full signatories.
Article 53(1)(c) of Regulation (EU) 2024/1689 requires a copyright policy that honors rights reservations, and Article 53(1)(d) requires a public training content summary. Licensed datasets are disclosed only as Yes or No plus modality. The reading of each section sits with the Article 53 training data summary and copyright policy page.
xAI Reputation: Lawsuits, Regulators and Transparency, Dated
As of 2026-11-09, xAI is a defendant in three publisher copyright suits filed on 2026-11-04 and the plaintiff in a challenge to California's training data disclosure law, and the Irish regulator has twice acted over Grok training.
The three suits are in the Northern District of California: Cognella (3:26-cv-13189), Chicken Soup for the Soul (3:26-cv-13188) and Cambronne (5:26-cv-13187-SVK, filed as 3:26-cv-13187).
The complaints allege that the plaintiffs' books were used to build Grok, including acquisition from pirate libraries such as LibGen. These are complaint allegations, and no court has ruled as of 2026-11-09.
The record of authors and publishers v. xAI holds the claims and status.
Stanford's Foundation Model Transparency Index of December 2025 scores xAI at 14 of 100, level with Midjourney and against a mean of 41 across 13 companies.
X.AI LLC filed a constitutional challenge to California's AB 2013, arguing trade secret, takings and compelled speech grounds. The district court denied a preliminary injunction, and the appeal is pending in the Ninth Circuit, where amicus briefs were filed on 2026-07-22. No decision is recorded as of 2026-11-09.
The page on the rulings and status of X.AI v. Bonta carries the court record.
Benefits, Risks, Downsides and Alternatives for a Data Holder Considering xAI
For a data holder, xAI offers an acknowledged appetite for licensed text and images but no published route, terms or counterparties, alongside an active litigation record, so the practical alternatives are buyers that publish programs, and the option of not licensing at all.
Benefits (as the record supports them):
- A published EU summary that acknowledges licensed text and images and private third-party datasets.
- An identifiable provider and an SEC-registered parent.
- A named agreement on record in a filing, the Cursor agreement, with fee terms tied to data access.
Risks:
- Pending copyright suits that allege use of pirate-library books.
- No signature of the Code of Practice copyright commitments.
- A transparency score of 14 of 100 in the December 2025 index, against a mean of 41.
- A court fight against training data disclosure in California.
- Entity names differ between the Grok 4.7 summary and the complaints of 2026-11-04, so the signing party needs confirming.
Downsides:
- No application route, payout or referral fee for company data is on record.
- No licensor, price or term is published for any rightsholder license.
- xAI is also a seller of data licenses, a competitor for some holders.
Alternatives (alphabetical; each one attribute as the party's claim, checked 2026-11-09):
- Handshake AI Data Partnerships: Handshake states "Earn $100K to $4M+", while the estimator on the same page reads "Your data could be worth up to $100,000".
- Labelbox Prism: datasets "typically start around $50K", and enterprise partnerships "can reach $1M+" (2026-11-09).
- micro1 Company Data Partnerships: USD 100K to 500K for qualified enterprise data partnerships (tier 1, 2026-11-09).
- Microsoft Publisher Content Marketplace, for publishers: publishers "will be paid on delivered value", with no rates published.
- Scale Data Partnerships: "$10K-$1M+ illustrative value per data partnership" (2026-11-09).
- Not licensing at all, which keeps the data and every later option open.
These figures are each party's claims, not payments made. The wider set of alternatives to selling data includes keeping the data.
xAI Among Frontier Labs That Acquire Data
xAI is one of the frontier labs profiled as data buyers and, as of 2026-11-09, the only one whose EU training content summary has been located.
It sits in the comparison of frontier AI labs as data buyers, and the hub on who buys AI training data places it beside every other buyer. Where a lab publishes no program, choosing a buyer for your data covers the next step.
- Other labs: OpenAI, Anthropic, Google, Meta, Microsoft, Mistral AI and Apple.
- Sources and opt-outs: user data and platform terms, opting out of AI training and robots.txt for AI crawlers.
- Law: California rules including AB 2013.
- Data types: selling source code to AI companies, and book publishers licensing books for AI training.
- Record: the log of data buyer incidents.
More About xAI as a Data Buyer
xAI AI Tutor Roles: Where Contributor Work Is Covered
Searches for xAI tutor roles concern paid contributor work for individuals, not the licensing of data a company holds.
The pay, assessments and reputation of that work sit in the comparison of AI training jobs and platforms compared.
Where xAI's Cases and Deals Are Tracked
Each record above has its own tracker page with full terms, sources and status.
- Authors and publishers v. xAI: the three publisher suits of 2026-11-04.
- DPC v. X over Grok training: the High Court application and the inquiry.
- Cursor and SpaceX AI: the agreement of 2026-04-19.
- The tracker of lawsuits over training data holds every case.
- The tracker of AI data licensing deals holds every deal.
Frequently Asked Questions About xAI and Training Data
Does xAI pay for training data?
xAI states it holds commercial licenses with rightsholders for text and images, without naming licensors or prices, and SpaceX's filing records fees in its Cursor agreement; no payout program for data holders is on record.
Does Grok train on user data and X posts?
SpaceXAI states for Grok 4.7 that user interaction data was used and that X data of X users who opted out was not.
Can a company license its data to xAI?
No xAI program, application page or terms for company records appear in the sources checked as of 2026-11-09.
Which legal entity provides Grok?
SpaceXAI LLC, formerly xAI LLC, of 1450 Page Mill Rd, Palo Alto, a subsidiary of Space Exploration Technologies Corp. since February 2026.
Is xAI a signatory of the EU Code of Practice?
Not of the copyright commitments, by SpaceXAI's own statement of 2026-09-21; the Commission lists xAI for the Safety and Security chapter only.