Meta Secures $100B AMD Deal for Massive AI Data Center Growth

Meta is scaling its artificial intelligence infrastructure through a major new semiconductor partnership with AMD.

The agreement outlines large-scale GPU and CPU deployments across future data centers and introduces an equity-linked structure tied to shipment and market performance milestones.

The $100 Billion Framework and Equity Structure

According to TechCrunch, Meta could purchase up to $100 billion worth of AMD chips, enough to support roughly six gigawatts of data center power demand.

AMD has issued Meta a performance-based warrant for up to 160 million shares, about 10 percent of the company, priced at $0.01 per share. Final vesting depends on AMD’s stock reaching $600, compared with its recent close of $196.60, as reported by The Wall Street Journal.

Meta will deploy MI540 GPUs alongside next-generation CPUs. AMD Chief Executive Lisa Su said CPU demand is rising with AI inference scaling. The partnership follows AMD’s prior equity-linked supply agreement with OpenAI and comes shortly after Meta expanded its relationship with Nvidia. 

Deployment Milestones and Market Impact

Reporting from CNBC states that Meta’s multiyear agreement involves deploying up to six gigawatts of AMD GPUs for AI data centers, including early shipments of MI450 GPUs in Helios rack-scale systems later this year.

The warrant vests in stages, beginning after the first one gigawatt of Instinct GPU shipments and extending through six gigawatts, tied to stock price and commercial milestones.

Lisa Su described the structure as aligned with long-term investment planning. Following the announcement, AMD shares rose 7 percent, while Meta traded slightly lower. Analyst Ben Bajarin of Creative Strategies told CNBC the scale implies a multiyear rollout potentially worth tens of billions of dollars.

Why This Matters for the AI Compute Landscape

Meta’s AMD agreement reflects broader competitive pressure in the AI chip market, where Nvidia controls roughly 90 percent of the share and remains the largest publicly traded company globally.

By structuring equity-linked supply commitments, AMD positions itself as a scaled alternative supplier while securing long term demand visibility. 

For Meta, the arrangement reduces reliance on a single vendor as it builds out dozens of global data centers to support large-scale AI systems.

The Road Ahead

Meta has pledged long term US data center and AI investments totaling at least $600 billion, including projected capital expenditures of $135 billion in 2026.

As hyperscalers expand compute capacity to support increasingly complex AI systems, multivendor semiconductor strategies and customized deployments are becoming central to competitive positioning. 

The massive wave of investment shows that AI infrastructure is only going to keep growing. Moving forward, how fast we can get the hardware, the availability of power, and the long-term chip roadmaps could be what truly defines the next era of enterprise AI.

Source: AMD and Meta Announce Expanded Strategic Partnership