Meta Signs $100 Billion AI Chip Deals With AMD and Nvidia in Unprecedented Infrastructure Push

Meta is doubling down on artificial intelligence at an almost incomprehensible scale. In the span of a single week, the company has locked in multi-year chip supply agreements with both Nvidia and AMD worth a combined sum that analysts estimate exceeds $100 billion — a figure that underscores just how central AI infrastructure has become to the tech giant's strategy.

The Nvidia Deal: CPUs, GPUs, and a First-of-Its-Kind Deployment

The first shoe dropped on February 18th, when Nvidia announced a sweeping multiyear agreement with Meta to supply millions of its Grace and Vera CPUs alongside Blackwell and Rubin GPUs for Meta's data centers. The deal is notable not just for its size but for what Nvidia describes as "the first large-scale Nvidia Grace-only deployment," promising significant performance-per-watt improvements across Meta's infrastructure.

The partnership also includes plans to integrate Nvidia's next-generation Vera CPUs into Meta's facilities starting in 2027, signaling that this is not a one-off purchase but a long-term architectural commitment. Neither company disclosed the financial terms, but industry watchers note that 2026 AI spending from Meta, Microsoft, Google, and Amazon is estimated to exceed the entire cost of the Apollo space program.

AMD Enters the Ring: Six Gigawatts of Processing Power

Just days later, AMD confirmed a parallel deal with Meta valued at roughly $100 billion over multiple years. The agreement calls for Meta to deploy six gigawatts worth of AMD processors across its AI data centers — a staggering volume that could see Meta owning up to 10 percent of AMD's stock.

The AMD arrangement follows a similar partnership between AMD and OpenAI announced last year, in which the ChatGPT-maker committed to deploying AMD Instinct MI450 GPUs at gigawatt scale. For AMD, the Meta deal represents a major validation of its strategy to challenge Nvidia's dominance in the AI accelerator market.

Why Meta Is Hedging Its Bets

Meta's decision to sign with both chip giants simultaneously is a calculated diversification play. The company has been developing its own in-house AI chips, but according to reports from the Financial Times, that effort has encountered "technical challenges and rollout delays." By securing supply from both Nvidia and AMD, Meta ensures it won't be bottlenecked by any single vendor as it scales up AI-powered products across Facebook, Instagram, WhatsApp, and its metaverse platforms.

The dual-sourcing strategy also gives Meta leverage in price negotiations and reduces its exposure to the kind of supply chain disruptions that have plagued the semiconductor industry in recent years.

The Bigger Picture: An AI Arms Race

Meta's spending spree is part of a broader industry trend that shows no signs of slowing. The company's AI capital expenditure for 2026 alone is projected to reach tens of billions of dollars, joining Microsoft, Google, and Amazon in what has become a full-blown infrastructure arms race.

The scale of investment raises questions about sustainability and returns. Critics point out that AI revenues have yet to match the enormous upfront costs, while proponents argue that the companies building the most capable infrastructure today will dominate the next decade of technology.

What's Next

With Nvidia's Vera CPUs arriving in 2027 and AMD's MI450 GPUs rolling out in the second half of 2026, Meta's data centers are set for a generational upgrade. The company's ability to translate that raw compute power into products that justify the investment will be the real test — one that investors, competitors, and the broader tech world will be watching closely.