AMD and Anthropic agree circular transaction

By the Editorial Team · Thursday, July 23, 2026

AMD has agreed to invest up to $5 billion in AI developer Anthropic, which will, in turn, purchase tens of billions of dollars in AMD’s AI server processors to expand its computing capacity by about 2 gigawatts. This deal strengthens AMD’s competitive position against rivals like Nvidia and supports Anthropic’s strategy to diversify and scale its AI infrastructure amid growing industry demand.

AMD has entered into a multibillion-dollar agreement with artificial intelligence developer Anthropic, which includes a potential investment of up to $5 billion by the U.S. chipmaker. This investment is contingent upon Anthropic achieving specified milestones. In return, Anthropic has agreed to purchase tens of billions of dollars’ worth of AMD’s latest AI server processors, enhancing AMD’s position against competitors like Nvidia.

The agreement will enable Anthropic to add approximately 2 gigawatts of computing capacity to its AI platform. The first gigawatt—comparable in power output to a single nuclear reactor—is expected to be operational by next year. This addition forms part of Anthropic’s broader effort to diversify its AI infrastructure, which already incorporates chips from Nvidia, Amazon Web Services, and Google.

Tom Brown, co-founder and chief compute officer of Anthropic, emphasized the importance of access to computing power to maintain the competitiveness of their AI model, Claude. The transaction mirrors similar arrangements in the industry, where leading AI firms secure supply chains and investments with chip manufacturers to guarantee access to critical hardware amid high global demand and semiconductor supply constraints.

This deal arrives during a period of intense competition among AI companies to secure large-scale computing resources. AMD has seen its shares rise more than 14% so far this year amid a broader semiconductor shortage affecting the industry. Other major players are making comparable investments: OpenAI recently secured 3.2 gigawatts of power for a data center in Georgia through a multibillion-dollar contract with the state’s electric utility, with the facility expected to be operational between 2028 and 2032.

Anthropic has also finalized significant agreements this year with Google, Broadcom, and Amazon Web Services, aiming to add over 10 gigawatts of computing capacity to meet growing customer demand. In a separate deal signed in May, Anthropic agreed to pay SpaceX $1.2 billion monthly to lease data center space, further intertwining the interests of the two companies.

The company’s leadership, including CEO Dario Amodei, has previously criticized competitors for aggressively locking in computing power beyond immediate needs. Nevertheless, Anthropic’s recent agreements signal a strategic shift toward securing extensive infrastructure ahead of anticipated demand surges.

As AI companies expand their data centers, they are encountering increased resistance from local communities concerned about environmental and infrastructure impacts. OpenAI, for example, has pledged hundreds of millions of dollars in state and local taxes related to its Georgia facility, along with an $80 million investment in community benefits such as local education and healthcare support.

This article was reported by synthesizing coverage from multiple independent news outlets to present a balanced, unbiased account. Read about our editorial standards. Spotted an error? Let us know.