Millions of Blackwell and Rubin GPUs Deployed
Multi-Year Data Center Agreement Signed

Meta has signed a large-scale artificial intelligence (AI) infrastructure deal with Nvidia that extends to central processing units (CPUs).


The two companies announced on the 17th (local time) that they had entered into a multi-year agreement to build AI data centers equipped with millions of Nvidia’s latest and next-generation GPUs, including “Blackwell” and “Rubin.”

Zuckerberg Joins Hands With Jensen Huang... Large-Scale AI Infrastructure Deal View original image

In particular, Meta has decided to deploy Nvidia’s “Grace” CPU as a standalone chip in its data centers. Among major data center operators, Meta is the first to adopt Nvidia CPUs exclusively for server use. Given that the data-center CPU market has until now been dominated by Intel and AMD, analysts say Nvidia has begun to crack the existing oligopoly structure.


Meta also said it is applying Nvidia’s networking platform “Spectrum-X” across its infrastructure to improve operational and power efficiency, and that it has implemented AI features in the WhatsApp messenger app while maintaining user data security by leveraging Nvidia’s confidential computing technology.



Meta is currently building the 1-gigawatt (GW) “Prometheus” data center in Ohio and the 5 GW “Hyperion” facility in Louisiana. Meta had previously shown signs of hedging by considering the adoption of Google’s Tensor Processing Unit (TPU), which competes with Nvidia in the AI chip space, but with this deal it will retain its status as one of Nvidia’s core customers.

Zuckerberg Joins Hands With Jensen Huang... Large-Scale AI Infrastructure Deal View original image

Mark Zuckerberg, CEO of Meta, said, “We are pleased to expand our partnership with Nvidia to build a leading cluster using the Vera Rubin platform,” adding, “Through this, we will deliver personalized SuperIntelligence to everyone around the world.” Nvidia CEO Jensen Huang said, “No company deploys AI at the scale that Meta does,” and added, “Through deep co-design across CPUs, GPUs, networking, and software, we will provide Meta with the full Nvidia platform to help it build the foundation for the next-generation AI frontier.”


This content was produced with the assistance of AI translation services.

© The Asia Business Daily(www.asiae.co.kr). All rights reserved.

Today’s Briefing