AMD and AI startup Anthropic announced a strategic partnership on July 22, 2026, under which AMD will invest up to $5 billion in Anthropic and supply up to 2 gigawatts (GW) of its next-generation Instinct MI450 AI chips [1, 2, 3, 4, 5].
Anthropic plans to deploy the AMD Instinct MI450 chips starting in the first half of 2027, with the first gigawatt of GPU capacity targeted for deployment in that period [1, 2, 3, 4, 5]. The chips will be installed in AMD's Helios rack-scale solutions designed for AI workloads [3, 4, 5, 6].
Anthropic will use these chips in its own data centers as well as through leased capacity from cloud providers, aiming to support its high-demand Claude AI model [1, 7, 8]. Tom Brown, Anthropic's cofounder and chief compute officer, said the partnership helps secure needed capacity and optimize it "for training and serving Claude" [4].
AMD CEO Lisa Su said, "We are thrilled to deepen our partnership with Anthropic and deploy AMD Helios at gigawatt scale," and highlighted combining Anthropic’s AI leadership with AMD’s high-performance computing power [3, 5].
The deal signals AMD's intent to challenge Nvidia's dominance in the AI hardware market by offering a large-scale supply of new AI chips [1, 7, 5, 8, 6]. Industry estimates place the cost of 1 GW of AI computing power at around $50 billion, underscoring the scale of the investment and infrastructure required [1].
In addition to this partnership, Anthropic has signed multi-gigawatt AI infrastructure agreements with companies including SpaceX, Google, Amazon, and Broadcom, and is in talks with Meta [3, 4, 8].
AMD and Anthropic will also engage in a multi-year engineering collaboration, with plans to use Anthropic’s Claude model in AMD’s software and product development [4]. If Anthropic meets certain chip deployment milestones, AMD may complete the full $5 billion investment, including financial guarantees for future data center leases [7].
