Meta has signed a multi-year, multi-generation agreement with AMD for the deployment of up to 6GW-worth of AMD Instinct GPUs.

In a statement, AMD said the partnership will see the two companies align their roadmaps across silicon, systems, and software to deliver AI platforms “purpose-built” for Meta’s workloads.

As part of the deal, AMD has issued Meta a performance-based warrant for up to 160 million shares of AMD common stock, tied to the achievement of specific milestones associated with Instinct GPU shipments.

The first deployment will comprise custom Instinct GPUs based on AMD’s MI450 architecture, with shipments supporting the first gigawatt deployment scheduled to begin in the second half of 2026. The initial deployment will be powered by the custom AMD GPUs and the company’s sixth-generation AMD Epyc CPUs, codenamed ‘Venice’, built on the AMD Helios rack-scale architecture.

First unveiled in June 2025, AMD’s double-wide Helios AI rack features the company’s Epyc ‘Venice’ CPUs, MI455X GPUs, and Pensando Vulcano NICs, delivering 2.9 exaflops of FP4 performance per rack. In the statement announcing the partnership with Meta, AMD said Helios was developed in collaboration with Meta through the Open Compute Project.

Meta will also be a “lead customer” for AMD’s ‘Venice’ CPUs and its ‘Verano’ Epyc CPUs, the latter of which are targeted for 2027.

“We are proud to expand our strategic partnership with Meta as they push the boundaries of AI at unprecedented scale,” said Dr. Lisa Su, chair and CEO, AMD. “This multi-year, multi-generation collaboration across Instinct GPUs, EPYC CPUs, and rack-scale AI systems aligns our roadmaps to deliver high-performance, energy-efficient infrastructure optimized for Meta’s workloads, accelerating one of the industry’s largest AI deployments and placing AMD at the center of the global AI buildout.”

Mark Zuckerberg, founder and CEO of Meta, added: “We’re excited to form a long-term partnership with AMD to deploy efficient inference compute and deliver personal superintelligence. This is an important step for Meta as we diversify our compute. I expect AMD to be an important partner for many years to come.”

The agreement closely mimics a deal signed by AMD with OpenAI in October 2025, under the terms of which AMD announced it would also supply the generative AI giant with 6GW-worth of GPUs. That effort is also expected to kick off in the second half of 2026, with OpenAI saying it would build a 1GW data center using AMD MI450 chips.

No information about where Meta or OpenAI will be deploying the initial 1GW has been provided; however, during AMD’s earnings call earlier this month, Su said that with regard to OpenAI, the “ramp starting in the second half of the year going into 2027” is “on track.”

Last week, Meta announced it had signed a multi-year and multi-generational agreement with Nvidia that will see it deploy the chipmaker’s CPUs alongside “millions” of Blackwell and Rubin GPUs and the Spectrum-X Ethernet networking platform.