Meta announced an expanded partnership with Nvidia on Tuesday to deploy millions of chips in its artificial intelligence data centers, marking a significant expansion of a decade-long collaboration.
The multiyear agreement includes Nvidia’s new standalone Grace central processing units, next-generation Vera Rubin rack-scale systems, and Spectrum-X Ethernet networking technology. Meta will utilize these components to support AI features on WhatsApp and its broader infrastructure. The deal represents a major commitment within Meta’s capital expenditure plans, with chip analyst Ben Bajarin of Creative Strategies estimating it to be “certainly in the tens of billions of dollars.” This agreement is part of Meta’s overall commitment to spend $600 billion in the United States by 2028 on data centers and infrastructure.
Meta becomes the first company to deploy Nvidia’s Grace CPUs as standalone chips in data centers, rather than incorporated alongside GPUs in a server. Nvidia described this as the first large-scale deployment of Grace CPUs on their own. Bajarin noted that Meta doing this at scale affirms Nvidia’s infrastructure strategy across CPU and GPU systems. The deal secures supply for Meta of Nvidia’s current Blackwell GPUs, which have been on backorder for months, and the next-generation Rubin GPUs, which recently entered production. Meta plans to deploy the next-generation Vera CPUs in 2027. The agreement also includes Nvidia’s security capabilities for AI features on WhatsApp.
Engineering teams from both companies will work together to optimize and accelerate AI models. Meta is developing a new frontier model called Avocado as a successor to its Llama AI technology. The most recent version of Llama, released last spring, failed to excite developers, according to a previous report. Meta has plans for 30 data centers, 26 of which will be based in the U.S. Its two largest AI data centers under construction are the Prometheus site in New Albany, Ohio, and the Hyperion site in Richland Parish, Louisiana. In January, Meta announced plans to spend up to $135 billion on AI in 2026.
Shares of Meta and Nvidia climbed in extended trading on Tuesday. Advanced Micro Devices stock sank about 4% on the news. In November, Nvidia stock fell 4% on reports that Meta was considering using Google’s tensor processing units in its data centers in 2027. Meta CEO Mark Zuckerberg stated that the expanded partnership continues the company’s push to deliver personal superintelligence to everyone. Meta also develops in-house silicon processors and utilizes chips from AMD. Meta’s stock saw its worst day in three years in October after the company announced ambitious AI spending, then popped 10% in January after reporting stronger-than-expected sales guidance.
