Nvidia and Meta Forge Extensive Partnership in AI and Infrastructure Sector
Tech giant Nvidia has revealed a comprehensive long-term collaboration with Meta, encompassing on-premises, cloud services, and artificial intelligence infrastructure. This strategic alliance will involve the integration of millions of Nvidia’s latest chips and networking technology into Meta’s operations.
In an official statement, Nvidia disclosed that the partnership entails the widespread deployment of Nvidia’s cutting-edge Blackwell GPUs and upcoming Rubin AI chips, alongside independent installations of the Grace and Vera central processors. Additionally, Meta will incorporate Nvidia’s Spectrum-X Ethernet networking solution into its Facebook Open Switching System platform. Financial specifics of the agreement were not disclosed publicly.
Jensen Huang, Nvidia’s CEO, emphasized the significance of this collaboration, stating, “Through comprehensive collaboration in CPUs, GPUs, networking, and software, we are providing Meta’s researchers and engineers with the complete Nvidia platform to propel advancements in AI.” Mark Zuckerberg, Meta’s CEO, articulated that the partnership’s objective is to “offer personalized superintelligence to individuals worldwide.”
This joint effort marks the premiere large-scale implementation of Nvidia’s Grace CPUs independently. Both companies have observed notable enhancements in performance efficiency within Meta’s data center applications. Furthermore, they are actively working on the deployment of next-generation Vera CPUs, aiming for extensive integration by 2027, as per Nvidia’s official blog.
Meta has embraced Nvidia’s Spectrum-X Ethernet networking technology across its infrastructure to deliver AI-centric networking capabilities with exceptional low-latency performance. Moreover, the company has leveraged Nvidia’s confidential computing technology for WhatsApp, enabling AI-driven functionalities while upholding user privacy.