Meta and NVIDIA have announced a multiyear, multigenerational partnership to develop AI infrastructure spanning on-premises, cloud, and data center deployments. The collaboration will involve deploying NVIDIA CPUs and millions of NVIDIA Blackwell and Rubin GPUs, alongside NVIDIA Spectrum-X Ethernet switches for Meta’s Facebook Open Switching System platform.
According to the press release, Meta will build hyperscale data centers optimized for both AI training and inference, and the companies will expand deployment of Arm-based NVIDIA Grace CPUs, marking the first large-scale Grace-only implementation. NVIDIA Vera CPUs may also see large-scale deployment in 2027.
Jensen Huang, founder and CEO, NVIDIA, said, “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.”
Mark Zuckerberg, founder and CEO, Meta, said, “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.”
Meta will also implement NVIDIA GB300-based systems to unify operations across on-premises and cloud environments, while deploying Spectrum-X Ethernet networking for low-latency, high-efficiency performance. NVIDIA Confidential Computing will be used for WhatsApp to enable AI-powered features while keeping user data secure, with plans to expand these capabilities across Meta’s platforms.
The companies are collaborating on optimizing next-generation AI models, combining NVIDIA’s platform with Meta’s production workloads to improve performance and efficiency at scale.