Meta strikes massive AMD deal to dominate global AI landscape

Meta is accelerating its push in the global AI race with a new multiyear agreement that places a fresh spotlight on its expanding data center ambitions. Just days after announcing its plan to deploy millions of processors from Nvidia, the company has now confirmed a major partnership with Advanced Micro Devices that further broadens its hardware strategy and deepens its hold in the AI infrastructure arena.

A sweeping commitment to AMD’s data center technology

The agreement centers on Meta’s plan to deploy up to 6 gigawatts of AMD’s GPUs within its AI data centers over several years. According to AMD, the deployment will include AI-optimized CPUs and the company’s latest MI450 GPUs housed in its Helios rack-scale servers. The first shipments are expected to roll out before the end of the year, marking a significant step forward in AMD’s efforts to become a stronger competitor in the high-demand AI chip market.

AMD leader Lisa Su highlighted that the company’s technology is designed to bring high performance and strong energy efficiency to Meta’s growing workloads. The partnership also positions AMD as a central player in what has become one of the largest AI infrastructure expansions underway across the tech industry.

Meta’s rising investments in AI scale

This new AMD deal arrives during a period of substantial capital spending for Meta. The company recently announced a plan to invest up to 135 billion dollars this year as it works to keep pace with other AI leaders including OpenAI and Anthropic. The spending is earmarked for infrastructure, hardware and the continued growth of its large footprint of data centers. Meta currently has plans for 30 total facilities, including 26 located in the United States, reinforcing the scale of its AI-driven operations.

Industry analyst Ben Bajarin from Creative Strategies noted that Meta’s ability to integrate a variety of chip technologies across its systems gives it a unique advantage. He added that the new AMD deployment introduces a level of customization not currently associated with the company’s arrangement with Nvidia. With the first wave of AMD GPUs designed specifically for Meta’s needs, the agreement demonstrates the flexibility large AI developers are seeking as they address an industrywide shortage in computing capacity.

AMD’s rising competitive profile

For AMD, the partnership represents a major moment as it seeks to gain more ground in a market overwhelmingly dominated by Nvidia. Although Nvidia currently holds close to 90 percent of the AI chip segment, demand for alternatives has increased as companies search for greater availability and more diversified supply chains. AMD’s Helios platform offers the first large-scale alternative to Nvidia’s Grace Blackwell systems, which have seen intense demand since their launch.

Meta has long used both AMD and Nvidia technologies, and it also continues to build its own in-house processors. A report last year indicated that the company had explored future collaboration with Google to potentially bring the search giant’s TPUs into Meta’s data centers by 2027, further underscoring its multi-vendor strategy.

A new chapter in the AI hardware race

The new agreement reinforces Meta’s strategy of building a diversified and powerful AI infrastructure that can support increasingly complex workloads. The partnership also strengthens AMD’s credibility as a second major option for large-scale AI deployments, following its previous collaboration with OpenAI.

The competition between top chipmakers continues to intensify. Nvidia is set to report earnings this week, and analysts expect strong revenue growth as demand for AI processors remains high. AMD recently delivered year-over-year sales growth in the double digits, reflecting its momentum as organizations seek more alternatives in the high-performance computing space.

Meta’s latest partnership with AMD signals that the race to secure enough computing power is far from over. As companies push to build faster and more efficient AI systems, multivendor strategies are quickly becoming the new normal. With millions of processors already committed across multiple chipmakers, Meta’s evolving approach offers a clear glimpse into how the world’s largest tech companies are preparing for the next phase of artificial intelligence.

Source: CNBC