What Meta and AMD's Deal Means for AI Data Centre Hyperscale

Meta has selected Advanced Micro Devices (AMD) as a long‑term partner to power its next generation of AI infrastructure, aligning Meta’s rising compute needs with AMD’s strategy to be at the centre of global AI build‑outs.

Under the agreement, Meta plans to deploy up to 6GW of AMD Instinct GPU‑powered capacity to support its expanding AI workloads across its worldwide operations.

The collaboration spans hardware and software, with first GPU shipments slated for the second half of 2026 and systems built on the Helios rack‑scale architecture, which is open hardware designed to make data centre infrastructure more efficient, scalable and sustainable.

“We're excited to form a long‑term partnership with AMD to deploy efficient inference compute and deliver personal superintelligence,” said Mark Zuckerberg, Founder and CEO of Meta.

“This is an important step for Meta as we diversify our compute. I expect AMD to be an important partner for many years to come.”

Dr Lisa Su, Chair and CEO at AMD, adds: “We are proud to expand our strategic partnership with Meta as they push the boundaries of AI at unprecedented scale.

"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 optimised for Meta's workloads, accelerating one of the industry's largest AI deployments and placing AMD at the centre of the global AI buildout.”

Market momentum and ROI questions

Meta’s commitment is valued at approximately US$60bn, and Forrester analyst Alvin Nguyen estimates US tech companies will spend around US$660bn on AI assets.

Broader AI‑driven data centre investment could reach US$3tn across hyperscalers and AI leaders including OpenAI, Google and NVIDIA.

The surge has raised familiar questions about returns on massive AI spending and the durability of valuations. While concerns around an “AI bubble” persist, NVIDIA’s move becoming the first US$5tn company late last year underscored ongoing investor confidence.

For Meta, a US$60bn programme carries clear ROI risks, but company value and revenue growth continue to trend positively.

AI at hyperscale: Meta’s campus strategy

Meta says the AMD partnership is a cornerstone of its Meta Compute initiative and part of a broader plan to build “gigawatt‑plus” AI data centre campuses across multiple US states, including Louisiana, Texas, Wisconsin and Wyoming.

Key projects include:

• Prometheus Hyperscale in Ohio, expected online in 2026 as Meta’s first multi‑gigawatt AI data centre.
• Hyperion, a campus that could scale to five gigawatts over several years.

Mark Zuckerberg has suggested one site will span an area approaching the size of Manhattan (22.8 square miles), underscoring the spatial and power demands of next‑gen AI facilities.

Meta Compute initiative

Meta frames its AMD agreement as central to scaling infrastructure for “the era of personal superintelligence,” co‑designing high‑performance, energy‑efficient hardware tightly integrated with its software stack.

“We believe this portfolio approach will enable us to advance and innovate at an unmatched pace, rolling out powerful, efficient new hardware co‑designed with our software stack to handle massive growth,” the company says.

“We look forward to working with AMD to power our AI innovations and secure our ability to deliver world‑class AI experiences to billions of people globally.”