Alphabet CEO Sundar Pichai visited the White House on Thursday for a series of high-level meetings with senior Trump administration officials.

Reports indicate the discussions centered on a key concern: U.S. fears of a shortage in artificial intelligence computing capacity, which could hinder the government's ability to stockpile sufficient AI computational resources to maintain its national defense systems.

This apprehension stems from the limited release of the Claude Mythos advanced model by AI firm Anthropic, which is currently accessible only to a select group of corporations and government agencies. While Anthropic cited safety as the reason for the restricted access, the move has generated widespread unease within U.S. political and Washington circles.

Recent tests have shown that the Mythos model can identify and exploit security vulnerabilities in critical software infrastructure at speeds unattainable by traditional cybersecurity teams.

The Trump administration is concerned that Anthropic's own computational reserves—referring to processing power in industry terms—may be insufficient. Even for high-priority institutional users, the company might be forced to limit access to the Mythos model. Some officials worry that in a crisis, the U.S. government could lack the necessary AI tools to patch vulnerabilities in its software systems.

This situation presents an opportunity for companies like Google and OpenAI. While the government appears to be managing its relationship with Anthropic, it is also looking to reduce reliance on the Claude model by utilizing Google's Gemini, OpenAI's GPT series, and other advanced AI technologies.

Separately, the U.S. Department of Defense announced last week that it has secured several cooperative agreements to integrate multiple AI models, besides Claude, into classified work.

A significant challenge for Google is that its proprietary AI processing chips, Tensor Processing Units (TPUs), are not yet compatible with some classified operational environments. Anthropic has become a convenient government choice largely because its models predominantly operate on the Amazon Web Services platform. Amazon has invested billions to obtain the industry's highest-level security certification.

The U.S. government is now exploring ways to expedite the security clearance process for Google's TPUs to achieve equivalent certification.

In related news, Anthropic is reportedly close to forming a $15 billion joint venture with Wall Street giants such as Goldman Sachs and Blackstone, aimed at selling AI tool services to private equity-held companies.