tech
The US blacklisted Anthropic as a security threat. Its spy agencies are using Claude anyway.
A chip shortage forced the NSA to keep using Anthropic’s AI despite a Pentagon blacklist. The White House approved $9B for classified data centres.
TL;DR
- The Pentagon has blacklisted Anthropic as a national security supply chain threat.
- The NSA is continuing to use Anthropic's AI due to a lack of alternatives caused by a chip shortage.
- The White House has authorized an emergency $9 billion funding request to build specialized federal data centers for AI.
- This funding aims to secure high-end semiconductors for spy agencies to run generative AI on classified networks.
- The computing demands of modern AI have outpaced the capabilities of existing classified government networks.
- Anthropic's AI, particularly Claude, is considered one of the most capable reasoning models available.
- The chip shortage impacting the government mirrors issues in the consumer electronics market.
- The situation highlights a structural tension in US AI policy, where frontier AI capability is concentrated in a few private companies.
- The government's inability to control its AI supply chain is a direct consequence of the hardware gap.
- The $9 billion initiative seeks to create independent infrastructure to run any chosen AI models, reducing reliance on single providers.