AMD announced a $1 billion investment in Meta's equity offering and revealed that the first Helios rack-scale systems are already installed and under testing at the company.
What happened?
In an official statement on October 27, 2026, AMD highlighted the deepening of its strategic partnership with Meta. AMD EPYC processors power the dedicated computing environments behind Meta's Muse personal AI agent, and the company has deployed millions of EPYC CPUs across multiple generations. Meta is a lead customer for 6th Gen EPYC, with future collaboration including the "Verano" codename.
Additionally, AMD Instinct GPUs are in production for training, ranking and recommendations in Meta's infrastructure. There is Day-0 support for models such as Muse Spark Open on Instinct and Muse Glimmer on Ryzen AI PCs and Radeon. The partnership includes up to 6 GW of AI infrastructure and ROCm software expansion via PyTorch.
Why does this matter?
AMD's $1 billion investment follows NVIDIA's announcement of $3 billion in Meta, showing how major chip suppliers are financially supporting the AI buildout of large customers. Helios systems, with up to 72 GPUs per rack, represent the rack-scale that Meta is testing toward broader deployment.
Mark Zuckerberg highlighted AMD as a key partner in frontier model computing and the path to personal superintelligence. Lisa Su, AMD CEO, emphasized the evolution from years of co-design to integrated systems at gigawatt scale.
What changes in practice?
For the AI hardware market, the news confirms AMD's strength in server CPUs and the progress of its rack-scale solutions with one of the largest AI players. Meta continues diversifying suppliers beyond NVIDIA, while AMD gains validation and aligned capital.
The first Helios already running marks a concrete milestone in the multi-year partnership announced earlier. Details on deployment volume and scale timeline remain limited, but collaboration on models, software and silicon points to strong continuity.
By GeekikiBot
