AMD has agreed to a multibillion-dollar agreement with artificial intelligence developer Anthropic that includes an investment of up to $5 billion by the chipmaker in the AI firm. Under the terms of the deal, Anthropic will commit to purchasing tens of billions of dollars’ worth of AMD’s latest AI server chips, supporting the company's efforts to challenge Nvidia in the AI hardware market.
The collaboration will expand Anthropic’s computing capacity by 2 gigawatts, with the first gigawatt—comparable to the output of a single nuclear reactor—expected to become operational next year. This deal marks another example of a circular arrangement between leading AI companies and semiconductor manufacturers, exemplified previously by partnerships involving OpenAI and other chip suppliers.
Tom Brown, co-founder and chief compute officer at Anthropic, emphasized the importance of reliable compute resources in maintaining the performance and availability of the company’s Claude AI model amid rising customer demand. By adding AMD’s processors, Anthropic diversifies its AI infrastructure, which already incorporates chips from Nvidia, Amazon, and Google.
AMD’s investment is contingent on Anthropic meeting certain milestones. The agreement arrives as the AI sector faces a broad semiconductor shortage, driving up chip stock values, with AMD’s shares increasing by more than 14% this year. The race for computing capacity is intensifying: OpenAI recently secured 3.2 gigawatts of power for a forthcoming data center in Georgia, in a deal involving tens of billions of dollars and scheduled to come online between 2028 and 2032.
In 2024, Anthropic has made other significant deals to increase its computing capability, including agreements with Google, Broadcom, and Amazon Web Services. These contracts give the San Francisco-based company access to over 10 gigawatts of new capacity, as it looks to fulfill growing demand for its AI services. Additionally, Anthropic reached a substantial arrangement with Elon Musk’s SpaceX to lease data center space, reportedly paying $1.2 billion a month, which represents a notable partnership between two industry competitors.
The expansion of AI data centers faces opposition from local communities concerned about environmental and infrastructural impacts. In response, OpenAI has committed hundreds of millions of dollars in state and local taxes associated with its Georgia project, along with $80 million designated for community benefits, including support for schools and healthcare.
Anthropic’s CEO, Dario Amodei, who had previously criticized aggressive early spending on compute capacity by competitors, now seems to have shifted strategy amid the rapid build-out of AI infrastructure across the industry. The new deals with AMD and others highlight the urgency AI companies place on securing sufficient computational resources to sustain the accelerating demand for advanced machine learning models.