Meta has entered a multi-year partnership with NVIDIA to expand its artificial intelligence (AI) infrastructure, which will include a large-scale rollout of the company’s CPUs and GPUs across both on-premises and cloud-based data centers. This collaboration aligns with Meta’s long-term goals to enhance its AI capabilities.
The agreement will facilitate the construction of hyperscale data centers dedicated to AI training and inference, deploying millions of NVIDIA Blackwell and Rubin GPUs. Additionally, NVIDIA Spectrum-X Ethernet switches will be integrated into Meta’s Facebook Open Switching System platform to support these advanced operations.
A notable aspect of this partnership is Meta’s increased use of Arm-based NVIDIA Grace CPUs in its data center applications, aimed at improving performance per watt. This marks the first significant deployment of NVIDIA Grace CPUs, which will be complemented by joint efforts in software optimization and enhancements to CPU ecosystem libraries, designed to boost efficiency with each generation.
Looking ahead, both companies are preparing for the potential large-scale adoption of NVIDIA Vera CPUs set for 2027, which are anticipated to further enhance the energy-efficiency profile of Meta’s AI compute resources. NVIDIA founder and CEO Jensen Huang remarked, “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.”
As part of this unified architecture, Meta will also implement NVIDIA GB300-based systems across its on-premises facilities and through partnerships with NVIDIA Cloud Partners. This move aims to maximize scalability and performance while streamlining operational processes.
In a bid to meet the specific demands of AI-focused networking, Meta has already deployed the NVIDIA Spectrum-X Ethernet networking platform throughout its infrastructure, targeting low-latency connectivity and optimized power usage.
Meta CEO Mark Zuckerberg expressed enthusiasm about the partnership, stating, “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.”
Moreover, Meta has incorporated NVIDIA Confidential Computing into WhatsApp’s processing framework, thereby enhancing user privacy while enabling advanced AI capabilities within the messaging service. The companies plan to extend these confidential computing features across more products in Meta’s portfolio.
Engineering teams from both organizations are closely collaborating to co-design next-generation AI models that aim to improve performance and efficiency for key workloads utilized globally. This joint effort merges Meta’s extensive production-scale requirements with the advancements in NVIDIA’s technology stack.
With this partnership, both Meta and NVIDIA are positioning themselves to lead in the burgeoning field of AI, emphasizing energy efficiency, scalability, and user privacy. As AI continues to evolve, the implications of this collaboration could significantly shape the future of AI technology and its applications across various sectors.
See also
Google and Nvidia Announce Major AI Investments at New Delhi Summit, Targeting $200B in Deals
Anterior Secures $40M Funding to Boost AI Adoption in Health Plans, Totaling $64M
Scout AI Unveils Lethal AI Agents for Military Drone Strikes in California Demo
GVSU Secures $1M for Trustworthy AI Development as Blue Dot Tech Hub Launches
Nvidia and Meta Expand Partnership with Billions for AI Infrastructure and CPUs
