AMD and Meta AI Partnership: 6 Gigawatts of Instinct GPUs Deployed
AMD (NASDAQ: AMD) and Meta have announced a massive expansion of their strategic partnership, centered on a multi-generational agreement to deploy up to 6 gigawatts of AMD Instinct™ GPUs. This collaboration is designed to rapidly scale Meta’s AI infrastructure and accelerate the development of next-generation AI models.
 |
| AMD and Meta AI Partnership: 6 Gigawatts of Instinct GPUs Deployed |
Custom Silicon for Gigawatt-Scale AI
The partnership moves beyond off-the-shelf components, focusing on deeply aligned roadmaps across silicon, systems, and software. Key highlights of the agreement include:
- Custom MI450 Architecture: The initial deployment will feature a custom AMD Instinct GPU based on the MI450 architecture, specifically optimized for Meta’s unique workloads.
- Next-Gen Processing: The infrastructure will be powered by 6th Gen AMD EPYC™ CPUs (codenamed "Venice"), ensuring a high-performance balance between CPU and GPU operations.
- Advanced Rack Architecture: The deployment utilizes the AMD Helios rack-scale design, which was first introduced at the 2025 Open Compute Project Global Summit.
Also Read: Lenovo and AMD collaborate to provide portable workstations for maximum productivity
Accelerating the Future of Open AI
Shipments for the first gigawatt of this deployment are expected to commence in the second half of 2026. By integrating customized AMD hardware, Meta aims to enhance the efficiency and speed at which it trains and deploys its most sophisticated AI models.
"Meta’s commitment to AMD Instinct GPUs validates our strategy of delivering high-performance, energy-efficient AI solutions at massive scale," said AMD executives. The collaboration marks one of the largest infrastructure agreements in the history of the semiconductor industry, signaling a shift toward specialized, large-scale AI data centers.
Deepening Ecosystem Collaboration
Beyond hardware, AMD and Meta are deepening their work on software ecosystems, including the ROCm™ open software stack. This ensures that Meta’s developers can maximize the performance of the custom Instinct GPUs, fostering a more open and accessible environment for AI innovation.
This multi-year partnership cements AMD’s position as a primary provider of the heavy-duty compute power required to sustain the global AI revolution.