Meta Platforms has forged a major multi-year partnership with Nvidia to deploy millions of advanced processors across its AI data centres.
The deal introduces standalone Grace CPUs at scale for the first time, alongside next-generation Vera Rubin systems planned for 2027.
Chief executive Mark Zuckerberg said the collaboration advances Meta’s goal “to deliver personal superintelligence to everyone in the world,” a vision unveiled last July. Though financial details are undisclosed, analysts value it in tens of billions, matching Meta’s $135 billion AI spending plan for this year.
“The deal is certainly in the tens of billions of dollars,” said Ben Bajarin of Creative Strategies. “We do expect a good portion of Meta’s capex to go toward this Nvidia build-out.”
Evolving from a decade of Nvidia GPU use, the pact spans processors, networking and software, with Grace optimised for inference tasks.
“They’re really designed to run those inference workloads, run those agentic workloads, as a companion to a Grace Blackwell/Vera Rubin rack,” Bajarin noted. “Meta doing this at scale is affirmation of the soup-to-nuts strategy that Nvidia’s putting across both sets of infrastructure: CPU and GPU.”
It backs Meta’s $600 billion U.S. data centre push by 2028, including 30 sites—mostly domestic—like the 1GW Prometheus in Ohio and 5GW Hyperion in Louisiana. Spectrum-X switches and WhatsApp AI security features are also included.
Shares of Meta and Nvidia rose after hours Tuesday, while AMD fell 4 per cent, as Meta diversifies with Google TPUs, in-house chips and AMD amid Nvidia supply strains.