'No one deploys AI at Meta’s scale': Meta signs up Nvidia to power its next big AI projects — so what exactly do Mark Zuckerberg and Jensen Huang have planned?

• Meta and Nvidia launch multiyear partnership for hyperscale AI infrastructure
• Millions of Nvidia’s GPUs and Arm-based CPUs will handle extreme workloads
• Unified architecture spans data centers and Nvidia cloud partner deployments

Meta has announced a multi-year partnership with Nvidia aimed at building hyperscale AI infrastructure capable of handling some of the largest workloads in the technology sector.

This collaboration will deploy millions of GPUs and Arm-based CPUs, expand network capacity, and integrate advanced privacy-preserving computing techniques across the company’s platforms.

The initiative seeks to combine Meta’s extensive production workloads with Nvidia’s hardware and software ecosystem to optimize performance and efficiency.

Unified architecture across data centers

The two companies are creating a unified infrastructure architecture that spans on-premises data centers and Nvidia cloud partner deployments.

This approach simplifies operations while providing scalable, high-performance computing resources for AI training and inference.

“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,” said Jensen Huang, founder and CEO of Nvidia.

“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.”

Nvidia’s GB300-based systems will form the backbone of these deployments. They will offer a platform that integrates compute, memory, and storage to meet the demands of next-generation AI models.

Meta is also expanding Nvidia Spectrum-X Ethernet networking throughout its footprint and aims to deliver predictable, low-latency performance while improving operational and energy efficiency for large-scale workloads.

Meta has begun adopting Nvidia Confidential Computing to support AI-powered capabilities within WhatsApp, allowing machine learning models to process user data while maintaining privacy and integrity.

The collaboration plans to extend this approach to other Meta services, integrating privacy-enhanced AI techniques into multiple applications.

Meta and Nvidia engineering teams are working closely to codesign AI models and optimize software across the infrastructure stack.

By aligning hardware, software, and workloads, the companies aim to improve performance per watt and accelerate training for state-of-the-art models.

Large-scale deployment of Nvidia Grace CPUs is a core part of this effort, with the collaboration representing the first major Grace-only deployment at this scale.

Software optimizations in CPU ecosystem libraries are also being implemented to improve throughput and energy efficiency for successive generations of AI workloads.

“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,” said Mark Zuckerberg, founder and CEO of Meta.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.