AMD commits $1 billion to Meta as the companies deepen their AI partnership

AMD is adding a financial commitment to its role supplying Meta’s computing infrastructure. In an October 27, 2026 partnership announcement, AMD said it will invest $1 billion in Meta’s equity offering. The companies are also deepening their technical work, with initial Helios systems installed at Meta and undergoing testing before a broader deployment.

The expanded relationship covers central processors, graphics processors, networking, complete systems and software for Meta’s services and AI infrastructure. Rather than describing a single chip purchase, AMD outlined collaboration across the equipment and software Meta uses today and the products the companies are helping shape for future workloads.

AMD included that commitment in the same announcement as the infrastructure work, bringing a financial relationship alongside the companies’ existing engineering and deployment ties.

Helios is installed, but the wider rollout is ahead

The initial rack-scale systems—computing systems organized at the rack level—are installed and being tested by Meta. AMD expects broader deployment “in the near future,” without giving a specific date.

AMD positions Helios as infrastructure for Meta’s frontier AI models and services.

Together, we are building and deploying infrastructure at gigawatt scale and shaping future generations of silicon, software and systems to enable Meta’s most ambitious superintelligence models and services.

Lisa Su, AMD chair and CEO, in the partnership announcement

An existing production relationship

Helios testing sits alongside AMD hardware already doing production work at Meta. AMD says its Instinct graphics processors are used for training, ranking and recommendation engines.

The processor relationship is larger than the initial Helios installation. Meta has deployed millions of AMD EPYC central processors across multiple generations and is a lead customer for the sixth generation. EPYC processors also support the dedicated computing environments behind Meta’s Muse personal AI agent, according to AMD.

Meta’s workloads help shape the next hardware

The collaboration also reaches back into product design. AMD says Meta was an early customer helping define the Instinct MI450 architecture and will deploy a custom Instinct graphics processor optimized for its workloads.

Beyond that chip, AMD says the partnership is informing its MI500-series graphics processors and next-generation rack-scale systems. The companies are also collaborating on future EPYC processors, including a generation named Verano.

AMD describes the arrangement as aligning product roadmaps and engineering decisions around Meta’s current and future workloads. The scope therefore extends beyond supplying available components: Meta’s computing needs are helping guide what AMD develops next, from processors to the larger systems in which they will operate.