> ## Content Index
> Fetch the complete content index at: https://www.metatalks.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Disclosed contracts reveal the depth of AI labs’ Pentagon ties
- URL: https://www.metatalks.ai/pentagon-contracts-ai-labs-forecast-own-risks/
- Published: 2026-09-09T10:59:00.000Z
- Updated: 2026-09-09T10:58:59.000Z
- Author: Al
- Tags: News, Military AI, AI safety, #newswire

**Anthropic, Google, OpenAI and xAI agreed to a much closer military partnership than their public statements suggested. Contracts obtained by** [**The Intercept**](https://theintercept.com/2026/09/08/military-ai-weapons-contracts-openai-anthropic-google/?ref=metatalks.ai) **detail their obligations to advise the Pentagon on AI strategy, train military personnel and forecast the risks of their own technology.**

The outlet secured more than 400 pages through a Freedom of Information Act lawsuit with Legal Advocates for Safe Science and Technology. The heavily redacted records cover agreements signed in July 2025, worth up to $200 million each, and subsequent amendments.

Google’s duties include briefing Pentagon officials on adversaries’ AI tactics. All four firms provide experts for military tabletop exercises. Military briefings and feedback would also inform the companies’ development and release of new technology.

An OpenAI amendment dated February 27 provides for company engineers to work inside military organizations, including combatant commands. Its entire oversight section is redacted. No equivalent section appears in the released Google and xAI documents.

The labs must also forecast dangers from their own upcoming models. The contracts credit them with advanced risk-forecasting expertise. Heidy Khlaaf, chief scientist at the AI Now Institute and a former OpenAI safety engineer, told The Intercept that relying on vendors to assess their own products creates a conflict of interest in decisions with life-or-death consequences.

Anthropic refused a follow-up agreement for classified systems without prohibitions on domestic spying and autonomous weapons. In March, Defense Secretary Pete Hegseth designated the company a “supply chain risk” and banned its services from government use; a federal judge overturned the ban in August. Google, OpenAI and xAI signed follow-up deals for classified networks.

Only OpenAI responded to The Intercept’s questions. Spokesperson Nate Evans said its agreement prohibits using the company’s technology for mass domestic surveillance, directing autonomous weapons systems or making high-stakes automated decisions.

The released documents include [OpenAI’s classified-network agreement](https://drive.google.com/file/d/149UWKnsgu46BG%5F8xF3iQt5giJkt9tgUv/view?usp=sharing&ref=metatalks.ai). The government has yet to provide the corresponding contracts for the other three labs.