Nvidia’s deal with meta signals a new era in computing power

Ask anyone what Nvidia makeand they’ll probably say “GPUs” first. For decades, the chipmaker has been defined by advanced parallel computing, and the rise of generative AI and the resulting increase in demand for GPUs has been a blessing for the company.

But Nvidia’s recent moves suggest it wants to lock more customers into the less compute-intensive end of the AI ​​market — customers who don’t necessarily need the best, most. powerful GPUs to train AI models, but instead look for the most efficient ways to run agentic AI software. Nvidia recently spent billions licensing technology from a chip startup focused on low-latency AI computing, and it also began selling standalone CPUs as part of its latest superchip system.

Yesterday, Nvidia and Meta announced that the social media giant agreed to buy billions of dollars worth of Nvidia chips to provide computing power for its massive infrastructure projects—with Nvidia’s CPUs as part of the deal.

The multi-year agreement is an extension of a cozy ongoing partnership between the two companies. Meta previously estimated that it would be purchased by the end of 2024 350,000 H100 chips from Nvidia, and that by the end of 2025 the company will have access to 1.3 million GPUs in total (although it wasn’t clear if these would all be Nvidia chips).

As part of the latest announcement, Nvidia said that Meta will “build hyperscale data centers optimized for both training and inference in support of the company’s long-term AI infrastructure roadmap.” This includes a “large-scale deployment” of Nvidia’s CPUs and “millions of Nvidia Blackwell and Rubin GPUs.”

Notably, Meta is the first tech giant to announce a large-scale purchase of Nvidia’s Grace CPU as a stand-alone chip, something Nvidia said would be an option when it revealed the full specs of its new Vera Rubin superchip in January. Nvidia has also emphasized that it offers technology that connects multiple chips, as part of its “soup-to-nuts approach” to calculating power, as one analyst puts it.

Ben Bajarin, CEO and principal analyst at technology market research firm Creative Strategies, says the move signaled Nvidia’s realization that a growing range of AI software must now run on CPUs, much in the same way that conventional cloud applications do. “The reason the industry is so bullish on CPUs inside data centers right now is agentic AI, which is putting new demands on common CPU architectures,” he says.

ON recent report from the chip newsletter Semianalysis underlined this point. Analysts noted that CPU usage is accelerating to support AI training and inference, citing one of Microsoft’s data centers for OpenAI as an example, where “tens of thousands of CPUs are now needed to process and manage the petabytes of data generated by the GPUs, a use case that would otherwise not have been required without AI.”

However, Bajarin notes that CPUs are still only one component of the most advanced AI hardware systems. The number of GPUs that Meta buys from Nvidia is still more than the CPUs.

“If you’re one of the hyperscalers, you’re not going to run there of your inference computation on CPUs,” says Bajarin. “You just need whatever software you’re running to be fast enough on the CPU to communicate with the GPU architecture that’s actually the driving force of that computer. Otherwise, the CPU becomes a bottleneck.”

Meta declined to comment on its extended deal with Nvidia. During a recent earnings call, the social media giant said it plans to dramatically increase its spending on AI infrastructure this year to between $115 billion and $135 billion, up from $72.2 billion last year.