Skip to content
TECH CEO Daily
AIDeal

AMD Anthropic deal: up to 2 gigawatts of Instinct GPUs and up to $5 billion AMD investment

Anthropic will deploy AMD Helios racks, starting with a first gigawatt in the first half of 2027, while AMD commits to becoming a shareholder.

By · Editor

Added to archive · Covers events of July 22, 2026 · 3 min read · Fact-checked

The 60-second brief

  • 1Anthropic will deploy up to 2 gigawatts of AMD MI450 Series GPUs; the first gigawatt starts deploying in H1 2027.
  • 2AMD committed to a strategic equity investment of up to $5 billion in Anthropic, which would make it supplier and shareholder.
  • 3AMD's August 5 quarterly filing disclosed up to $5.0 billion in investment commitments, subject to contingencies and expected through fiscal 2028, without naming the recipient; The Motley Fool linked them to Anthropic.

The news

In the AMD Anthropic deal announced on July 22, 2026, Anthropic will deploy up to 2 gigawatts of Advanced Micro Devices (AMD) Instinct MI450 Series GPUs, and AMD committed to invest up to $5 billion in Anthropic, tying a chip order to a planned stake.

According to AMD's announcement, Anthropic will use AMD Helios rack-scale systems built on Instinct MI455X graphics processing units (GPUs), part of the MI450 Series, together with 6th Gen AMD EPYC "Venice" CPUs, AMD Pensando networking and AMD's ROCm software. Deployment of the first gigawatt is scheduled to begin in the first half of 2027. The release stated the deployment in gigawatts and did not say how many GPUs that represents.

AMD said it has committed to make a strategic equity investment of up to $5 billion in Anthropic. The release did not disclose the timing, valuation or conditions of the investment, the value of the GPU purchases, or where the hardware will be hosted.

The new deployment builds on Anthropic's existing use of AMD Instinct MI355X GPUs, AMD said. Tom's Hardware reported that this was the first official confirmation that Claude workloads already run on AMD chips.

The companies also set up a multiyear engineering collaboration. They will use Claude to optimize workloads for Instinct GPUs and speed up development of ROCm, AMD's software stack for its GPUs, and AMD will adopt Claude broadly across its engineering and product development teams.

Tom Brown, Anthropic's co-founder and chief compute officer, said access to compute is central to "keeping Claude at the frontier and meeting demand from our customers." AMD Chair and CEO Lisa Su said AMD would deploy Helios with Anthropic at gigawatt scale.

The numbers

GPU capacity Anthropic will deploy
Up to 2 gigawatts (MI450 Series)
First gigawatt deployment begins
First half of 2027
AMD equity investment commitment
Up to $5 billion
Unnamed investment commitments in AMD's 10-Q (linked to Anthropic by The Motley Fool)
Up to $5.0 billion through fiscal year 2028, subject to contingencies

Why CEOs should care

For CFOs and analysts tracking AMD's AI business, both headline figures are ceilings: up to 2 gigawatts and up to $5 billion. Revenue from the first gigawatt depends on deployments that begin in the first half of 2027, and the purchase value was not disclosed. The useful questions are how quickly deployments convert into recognized revenue, what contingencies govern the equity payments, and how much of AMD's AI growth rests on a few large customers.

For enterprise buyers of Claude, the deal adds another hardware supplier behind a model many companies already use. Brown said running across diverse hardware lets Anthropic match workloads to suitable chips, as reported by Tom's Hardware. Procurement and security teams should ask Anthropic, or the cloud providers that resell Claude, whether service levels, data-handling commitments and security attestations apply the same way whichever chips serve their requests.

For boards, the structure is the lesson. A supplier taking an equity stake in a customer that is buying its hardware can lock in demand, but it also ties the supplier's results to that customer's financial health and funding. Directors of companies weighing similar supplier-financing arrangements should ask how contingencies, concentration risk and any related disclosures will be handled before signing.

The bigger picture

The deal shows how AI hardware agreements increasingly come with financial ties between supplier and customer. Tom's Hardware noted that the structure differs from AMD's earlier agreement with OpenAI, under which AMD granted OpenAI warrants that could amount to about a tenth of AMD's shares; this time AMD has committed to take a stake. The outlet also reported that Anthropic's compute already spans Nvidia (NVDA) GPUs, Amazon (AMZN) Trainium chips, Google TPU capacity that Broadcom (AVGO) is to supply starting in 2027, and Nvidia-based capacity that Anthropic leases at SpaceX's Colossus 1 data center, and that OpenAI's first gigawatt of AMD capacity is due in the second half of 2026.

What happened next

On July 23, 2026, at its Advancing AI 2026 event, AMD listed Anthropic among the companies deploying Helios at scale. On August 4, 2026, AMD's second-quarter results highlighted the partnership. AMD's quarterly report, filed August 5, 2026, said that after June 27 the company entered into investment commitments of up to $5.0 billion, subject to certain contingencies, expected to be made through fiscal year 2028. The filing did not name the recipient; The Motley Fool, reporting on September 5, linked the commitments to Anthropic and said neither company has disclosed the contingencies.

The next milestone is the start of the first gigawatt in the first half of 2027. TechCrunch reported on September 28, 2026, on Anthropic's IPO prospectus, including plans to spend $518 billion on cloud, computing and infrastructure in the coming years. TechCrunch's report did not mention AMD, and whether the prospectus discloses terms of the AMD agreement has not been confirmed.

What “Fact-checked” means

Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.

What we checked
Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
How
A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
Who
The checks are made with our newsroom’s technology tools, as steps kept separate from the writing, under rules set by our editor, . A story the checks still flag is not published automatically; it is held for the editor, who decides whether it is fixed, published or dropped.
Archive story
It was written after the event. Its dates were checked against the date of the event, including the dates of any later developments it reports.
If something is wrong
“Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error

How we fact-check →

Companies in this story

AMDAnthropicAI infrastructureInstinct GPUs

Earlier coverage of AMD

All AMD coverage →

Written by

Editor · Technology & Business Writer

Hussein is a writer and business technology enthusiast focused on the intersection of technology, entrepreneurship, finance, artificial intelligence, and digital innovation.

CoversAICybersecurityBig TechSaaSStartupsFintech

How this story was made. Researched and written using our newsroom’s technology tools and fact-checked before publication.

Published by Tech CEO Daily, an independent publication. Masthead · Editorial standards

Free newsletters

The technology briefing for people running businesses.

Daily, weekly, bi-weekly or monthly. You choose.

How often

The Daily Brief · Weekdays, 6 a.m. ET

Free forever. One click to unsubscribe. We never sell your email.