AMD has agreed to invest up to $5 billion in the AI company Anthropic as part of a major infrastructure deal. The agreement covers tens of billions of dollars' worth of AI systems, according to reports from Reuters.
Anthropic to deploy AMD chips at massive scale
Under the deal, Anthropic will deploy up to two gigawatts of computing capacity using AMD's Instinct MI450-series accelerators. The first gigawatt — which is roughly equal to the power output of a large nuclear reactor — is scheduled to begin deployment in the first half of 2027, as reported by The Financial Times.
The deal is structured in two separate parts. AMD is making an equity investment in Anthropic, while Anthropic is placing a large hardware order with AMD. The companies have not said that Anthropic must use AMD's investment to pay for the systems, according to Artificial Intelligence News.
AMD ties investment to chip deployment milestones
AMD's investment is tied to Anthropic meeting certain deployment milestones. However, the companies have not disclosed the specific conditions attached to those milestones, according to Reuters.
This deal marks a significant push by AMD into the AI infrastructure space, directly competing with Nvidia's dominant position in the AI chip market. Anthropic, which develops the Claude AI assistant, is one of the leading AI companies alongside OpenAI and Google DeepMind.
Our Take: A strategic bet on AI hardware and software
This deal is a clear signal that AMD is serious about challenging Nvidia's near-monopoly in AI chips. By tying its investment to actual deployment milestones, AMD is ensuring that its chips are not just purchased but actually put to work at scale. For Anthropic, the deal provides both capital and guaranteed access to cutting-edge hardware — two things that are critical in the race to build more powerful AI systems. The fact that the first gigawatt won't come online until 2027 shows that both companies are thinking long-term. In our view, this is exactly the kind of infrastructure deal that will define the next phase of the AI industry — where the winners are not just those with the best models, but those with the most efficient and scalable computing capacity.
