AMD highlights strategic partnership with Meta

AMD describes the work as spanning CPUs, GPUs, networking, systems and software for Meta services and AI infrastructure.[2] The companies say they are aligning product roadmaps and engineering decisions around Meta workloads, and building integrated AI systems at gigawatt scale.[1] For technology teams, the announcement points to collaboration across more of the infrastructure stack than accelerators alone.[2][1] The companies say they are building integrated AI systems at gigawatt scale.[1]

Meta has installed initial AMD Helios rackscale systems and is testing them, according to AMD.[1] AMD characterized that testing as being on the path to broader deployment in the near future.[1] Meta has deployed millions of EPYC processors across multiple generations, and the companies are collaborating on future generations, including Verano.[1]

AMD says its Instinct GPUs are in production at Meta for training, ranking and recommendation engines.[1] Meta was an early definition customer for AMD’s MI450 architecture and plans to deploy a custom Instinct GPU optimized for its workloads.[1] On software, the companies expanded work on large-scale AI training, including optimization of recommendation models on MI350X GPUs and contributions to PyTorch, Triton and ROCm.[1] AMD also said it delivered Day-0 support for Muse Spark Open on Instinct GPUs and Muse Glimmer on Ryzen AI PCs and Radeon GPUs.[1]

AMD said it will invest $1 billion in Meta’s equity offering.[1] Meta announced plans to raise $95 billion for AI infrastructure investments, with $35 billion initially planned through underwritten public equity offerings and another $60 billion through an at-the-market program.[3] Meta said Nvidia, AMD and Broadcom were interested in buying up to $5 billion in shares.[3] Separately, a warrant issued under an amendment to the companies’ purchase agreement entitles Meta, subject to its terms, to buy 160 million AMD shares; the holder may exercise it before Feb. 23, 2031.[4]