OpenAI agrees to deploy AI models on Pentagon network

Sam Altman, OpenAI CEO, said on X that the company agreed to deploy its AI models on the US Defense Department’s classified network after Anthropic PBC’s Pentagon relationship collapsed over surveillance and autonomous-weapon concerns.

Pete Hegseth, defense secretary, said on X that Anthropic was declared a supply-chain risk and given six months to hand over AI services to another provider.

Altman said the agreement reflects OpenAI’s principles banning domestic mass surveillance and requiring human responsibility for the use of force, including autonomous weapons, and that safeguards were built for deployment.

Anthropic said that it stipulates its products not be used for surveillance of Americans or for strikes without human involvement.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

A political feud helped push Anthropic out

A contract fight turned political after the White House stepped in. President Trump then publicly ordered all federal agencies to stop using Anthropic’s technology 1.
The Pentagon tagged Anthropic as a “supply-chain risk,” a harsh label usually aimed at foreign-adversary-linked companies like Russia’s Kaspersky Lab (a cybersecurity firm) or China’s Huawei (a telecom equipment giant), rather than an American company 2. The tag could block Anthropic from working with U.S. military vendors or contractors 2.
The Pentagon sought rights to use Anthropic’s AI for “all lawful purposes,” and Anthropic worried that language would weaken its safeguards on surveillance or the use of force 3. Separately, a source told Fortune that the U.S. government was talking with OpenAI about terms that would let OpenAI keep control of its “safety stack” (its internal safety controls and processes) and include OpenAI’s “red lines,” though no contract had been signed 3.

Government contracts are shaping AI ethics

With no clear AI legislation, federal procurement contracts are setting practical rules for how AI can be used 4.
The administration’s blacklisting of a major U.S. AI firm warns Silicon Valley that rejecting government terms can trigger commercial fallout. That pressure can hit federal revenue plus partnerships with other government contractors 5.
If OpenAI fills the gap left by Anthropic, it could deepen its role as a “trusted” platform for sensitive government work 4. That shift could concentrate spending among a few dominant AI vendors, raising the barrier to entry for smaller competitors 4.