Amazon doubles down on AI with $5 billion Anthropic deal

Amazon has confirmed a new $5 billion investment in Anthropic this Monday to deepen its ties with the startup behind the Claude artificial intelligence models, as demand for AI infrastructure continues to rise.

According to a recent report, the latest funding builds on the $8 billion Amazon had already committed to Anthropic. Under the expanded agreement, the company could invest up to an additional $20 billion, depending on the startup meeting specific commercial milestones, bringing the total potential investment to $25 billion.

As part of the deal, Anthropic has agreed to spend more than $100 billion over the next decade on Amazon Web Services infrastructure to support the training and deployment of its AI models. The company has also secured up to 5 gigawatts of computing capacity for Claude, with plans to bring roughly 1 gigawatt online using Trainium2 and Trainium3 chips by the end of the year.

“We need to build the infrastructure to keep pace with rapidly growing demand,” Anthropic chief executive Dario Amodei said in a release. He added that the partnership would help advance research efforts while expanding access to Claude for customers.

Amazon continues to lean heavily on its cloud business to anchor its position in the AI race, even as its in-house models, including Nova, have yet to gain significant traction.

The company is preparing for capital expenditures of about $200 billion this year, with a large share expected to go towards expanding AI infrastructure, according to Reuters.

At the same time, Amazon is increasing its exposure to leading AI firms. Its deeper commitment to Anthropic follows an earlier announcement that it could invest up to $50 billion in OpenAI, the developer of ChatGPT. Both Anthropic and OpenAI are competing to strengthen their market positions ahead of potential public listings that could take place as early as this year.

Anthropic growth, policy tensions and new model launch

Anthropic reported earlier this month that its annualised revenue had surged to more than $30 billion, tripling quarter-on-quarter and briefly surpassing OpenAI on that metric.

The company’s engagement with US policymakers has also shifted in tone. Dario Amodei met officials at the White House last week, signaling a more cooperative stance following tensions earlier this year. The dispute escalated in February after Anthropic resisted allowing its technology to be used for mass surveillance or fully autonomous weapons, drawing criticism from Pentagon chief Pete Hegseth.

“We discussed opportunities for collaboration, as well as shared approaches and protocols to address the challenges associated with scaling this technology,” a White House spokesperson told AFP.

The latest discussions contrast with earlier action by President Donald Trump, who had directed federal agencies to stop using Anthropic’s technology after the company declined to grant the Pentagon unrestricted access to its Claude models.

Anthropic has since challenged both the directive and its designation as a potential “supply chain risk” in court.

Separately, the company introduced a new AI model called Mythos earlier this month but chose not to release it publicly, citing concerns over possible cybersecurity risks.