AMD and Anthropic Forge Deep Partnership, 2 GW of AI Compute Power Set for Deployment

On July 22 local time, AMD and Anthropic officially announced a strategic partnership. The two parties will pursue long-term collaboration across multiple fronts, including large-scale compute deployment, joint AI technology R&D, and strategic investment.

Gigawatt‑Scale Compute Infrastructure Takes Shape, First GW to Start in 1H2027

As the AI industry advances at rapid speed, demand for compute capacity for AI training and inference continues to surge. To sustain the ongoing expansion of the Claude large language model and meet rapidly growing customer demand, Anthropic is scaling out its compute infrastructure. Anthropic will deploy up to 2 gigawatts of AMD Instinct MI450 series GPUs based on AMD Helios rackscale solutions, with the initial 1 GW deployment scheduled to kick off in the first half of 2027.

This deployment is not a greenfield trial; it builds on Anthropic's existing deployment of AMD Instinct MI355X GPUs. The solution is powered by AMD Instinct MI455X GPUs and features a full-stack configuration consisting of AMD EPYC Venice CPUs, AMD Pensando networking and the ROCm software stack. Purpose-built for next-generation advanced AI workloads, the AMD Helios rackscale platform delivers exceptional performance, power efficiency and scalability, making it well-suited for rigorous use cases such as ultra-large-scale foundation model training and complex inference. This gigawatt-scale compute deployment also marks a major leap forward in strengthening AMD's core position in the global AI infrastructure landscape.

Bidirectional Technology Synergy: Claude Empowers AMD Software Development

Under this partnership, AMD and Anthropic will also launch a multi-year engineering collaboration. Leveraging the technical strengths of Anthropic's Claude large language model, the two sides will optimize workloads running on AMD Instinct GPUs and accelerate development of AMD's core ROCm software. AMD will broadly adopt Claude across its internal engineering and product development teams.

Deep integration between hardware compute capacity and AI large language models enables precise matching and efficient utilization of compute resources, ensuring AI workloads across diverse use cases are mapped to optimal hardware configurations. Meanwhile, it will continuously enrich AMD's AI software ecosystem, laying a solid technical foundation for future compute deployments and hardware adaptation.

$5 Billion Strategic Investment: Deep Capital Ties Underpin Long-Term Cooperation

To reinforce the foundation of their long-term strategic partnership, AMD has committed to making a strategic equity investment of up to $5 billion in Anthropic. Deep capital linkage will drive mutual benefit and coordinated growth. This sizable strategic investment not only demonstrates AMD's strong recognition of Anthropic's cutting-edge large model capabilities, but also constitutes a key move by AMD to step up investments in the AI sector and refine its AI industrial layout.

Competition in the AI industry has entered a new phase defined by multi-dimensional rivalry spanning compute power, technology, ecosystems and capital. AMD boasts strengths in high-performance chips and compute infrastructure, while Anthropic holds core technologies for frontier large language models. This in-depth collaboration integrates the core strengths of both companies, accelerating the large-scale commercialization of AI technologies and positioning AMD Helios as a flagship platform for next-generation AI infrastructure.