🔥 Trade with Pros on Discord → 21 Days Free (No Card)JOIN FREE

Nvidia to supply Meta with Blackwell, Rubin GPUs and Grace/Vera CPUs under multiyear agreement

In this post:

  • Nvidia signed a multiyear deal with Meta to supply Blackwell and Rubin GPUs, along with Grace and Vera CPUs, for AI data centers.
  • Meta will deploy millions of GPUs, GB300 systems, and Spectrum-X networking to power training and inference at hyperscale.
  • Meta is expanding Grace CPU use now and plans potential large-scale Vera CPU deployment starting in 2027.

Nvidia has signed a multiyear, multigenerational partnership with Meta that covers on-premises systems, cloud deployments, and full AI infrastructure.

The deal locks in shipments of Nvidia Blackwell and Rubin GPUs, along with Nvidia Grace CPUs and future Vera CPUs. The agreement spans training systems, inference systems, and networking across Meta’s global data centers.

Meta said it will build hyperscale data centers designed for both training and inference as part of its long-term AI roadmap.

The plan includes deploying millions of Nvidia Blackwell and Rubin GPUs. It also includes large-scale rollout of Nvidia CPUs. On the networking side, Meta will integrate Nvidia Spectrum-X Ethernet switches into its Facebook Open Switching System platform.

Meta expands AI clusters with Blackwell and Rubin

Jensen Huang, founder and CEO of Nvidia, said, “No one deploys AI at Meta’s scale — integrating frontier research with industrial-scale infrastructure to power the world’s largest personalization and recommendation systems for billions of users.”

Jensen also said, “Through deep codesign across CPUs, GPUs, networking and software, we are bringing the full NVIDIA platform to Meta’s researchers and engineers as they build the foundation for the next AI frontier.”

Mark Zuckerberg, founder and CEO of Meta, said, “We’re excited to expand our partnership with NVIDIA to build leading-edge clusters using their Vera Rubin platform to deliver personal superintelligence to everyone in the world.”

Meta will deploy industry-leading GB300-based systems as part of this rollout. The company plans to build one unified architecture that connects its on-premises data centers with Nvidia Cloud Partner deployments. The goal is to keep operations simple while scaling performance across regions.

See also  Sam Altman steps down from OpenAI’s Safety Commission amid scrutiny

Meta has adopted the Nvidia Spectrum-X Ethernet networking platform across its infrastructure footprint. The networking system is designed for AI-scale traffic. It is built to deliver predictable, low-latency performance while maximizing hardware use and improving power efficiency.

Meta rolls out Grace and Vera CPUs while Nvidia exits Arm

Meta and Nvidia are continuing their work on Arm-based Grace CPUs inside Meta’s production data centers. The Grace chips are designed to improve performance per watt.

This collaboration represents the first large-scale Grace-only deployment. The companies invested in codesign and software optimization across CPU ecosystem libraries to improve efficiency with each generation.

The two companies are also working on deploying Vera CPUs. Large-scale deployment could begin in 2027. Vera is expected to extend Meta’s energy-efficient AI compute footprint and support the broader Arm software ecosystem.

Separately, Nvidia has sold the remainder of its stake in Arm, unloading 1.1 million shares valued at about $140 million based on Arm’s closing price Tuesday. The sale took place in the fourth quarter of last year and reduces Nvidia’s stake in Arm to zero.

The sale ends a long chapter. In 2020, Nvidia agreed to buy Arm for $40 billion. The deal faced opposition from regulators and industry players soon after it was announced. Arm’s chip technology supports most advanced semiconductors worldwide, and its independence was viewed as critical. In February 2022, both sides terminated the agreement.

See also  Epic games legal victory against pro player proves cheaters never win in Fortnite

Arm, majority-owned by SoftBank, later moved forward with plans to sell shares to the public.

SoftBank dumped its entire Nvidia stake in October. The sale was quiet but huge. It unloaded about $5.8 billion worth of shares. The goal was simple. Free up cash and double down on OpenAI. Clean break. No leftovers.

Since that disclosure, Nvidia has slipped around 7%. Now it heads into its February 26 earnings report with analysts still projecting revenue growth of 67%.

Masayoshi Son does not hedge. He goes all in. Then he goes all in again on something else. Each new bet usually requires selling the last one. He thinks in bold swings, not small trims. And the numbers he deals with rarely look normal. They look unreal.

This is the second time he has fully exited Nvidia. The first time was in 2019. That exit turned into one of those stories people bring up when they talk about regret. SoftBank had bought a 4.9% stake in the chipmaker in 2017 for roughly $4 billion. The position later generated about $3 billion in profit. Then crypto mining collapsed. Nvidia shares fell about 50%. SoftBank sold. At the time, it looked rational.

Join a premium crypto trading community free for 30 days - normally $100/mo.

Share link:

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Most read

Loading Most Read articles...

Stay on top of crypto news, get daily updates in your inbox

Editor's choice

Loading Editor's Choice articles...

- The Crypto newsletter that keeps you ahead -

Markets move fast.

We move faster.

Subscribe to Cryptopolitan Daily and get timely, sharp, and relevant crypto insights straight to your inbox.

Join now and
never miss a move.

Get in. Get the facts.
Get ahead.

Subscribe to CryptoPolitan