AMD and Meta deepen AI partnership with 6GW GPU roadmap

AMD and Meta have recently announced an expansion of their strategic partnership through a multi-year agreement that will see the deployment of up to 6 gigawatts of AMD Instinct GPUs for Meta’s next-generation AI infrastructure. This deal marks a significant deepening of collaboration between the two tech giants and positions AMD as a major silicon supplier for one of the world’s largest AI buildouts.

For readers of eeNews Europe, this development holds particular relevance as it underscores how hyperscale demand is steering rack-scale system design, fostering tighter CPU–GPU co-optimization, and emphasizing the growing importance of open hardware platforms like OCP. Additionally, the focus on power efficiency and total energy footprint is brought to the forefront.

Rack-scale AI Infrastructure at Gigawatt Levels

The agreement encompasses multiple generations of AMD Instinct GPUs and aligns silicon, systems, and software roadmaps between AMD and Meta. The initial phase of deployment will feature a custom AMD Instinct GPU based on the MI450 architecture, tailored specifically for Meta’s AI workloads. Shipments to support the initial 1-gigawatt deployment are slated to commence in the latter half of 2026.

These cutting-edge systems will be constructed on AMD’s Helios rack-scale architecture, which was unveiled at the 2025 Open Compute Project Global Summit and developed in collaboration with Meta. Helios is engineered to support scalable, rack-level AI infrastructure by integrating compute, networking, and power delivery at exceptionally high density. The platform will combine MI450-based Instinct GPUs with 6th Gen AMD EPYC CPUs, codenamed “Venice,” running AMD’s ROCm software stack.

Dr. Lisa Su, chair and CEO of AMD, expressed pride in the expanded partnership with Meta, stating, “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. This will accelerate one of the industry’s largest AI deployments and position AMD at the forefront of the global AI buildout.”

CPU Integration, Diversification, and Financial Alignment

The partnership between AMD and Meta extends beyond GPUs, with Meta having already deployed millions of AMD EPYC processors and substantial volumes of Instinct MI300 and MI350 series GPUs across its global infrastructure. As AI systems grow in scale and complexity, CPUs play an increasingly critical role in orchestration, efficiency, and overall system balance.

Under the new agreement, Meta will serve as a lead customer for both the “Venice” EPYC generation and “Verano,” a next-generation EPYC processor featuring workload-specific optimizations aimed at enhancing performance per dollar per watt.

Mark Zuckerberg, founder and CEO of Meta, expressed enthusiasm about the long-term partnership with AMD, stating, “This collaboration is a crucial step for Meta as we diversify our compute capabilities. I anticipate AMD to remain a key partner for many years to come.”

To ensure long-term alignment of incentives, AMD has granted Meta a performance-based warrant for up to 160 million shares of AMD common stock. The vesting of these shares is tied to Instinct GPU shipment milestones, stock price thresholds, and key technical and commercial targets.

“We anticipate that this partnership will drive significant multi-year revenue growth and contribute positively to our non-GAAP earnings per share,” remarked Jean Hu, EVP, CFO, and treasurer of AMD.