CNBC·2026-02-17 21:17Core Insights - Meta has announced a significant deal to utilize millions of Nvidia chips in its AI data centers, including standalone CPUs and next-generation Vera Rubin systems, as part of its vision to deliver personal superintelligence globally [1][2] - The financial terms of the deal remain undisclosed, but it is expected to be in the tens of billions of dollars, contributing to Meta's planned $135 billion investment in AI by 2026 [2] Group 1: Partnership Details - The partnership between Meta and Nvidia is an expansion of their long-standing relationship, with Meta having used Nvidia GPUs for over a decade [2] - The deal represents a broader technology collaboration, marking a significant step in Meta's infrastructure development [2] Group 2: Technological Advancements - Meta will be the first to deploy Nvidia's Grace CPUs as standalone chips in its data centers, which is a notable advancement in their technology strategy [3] - The Grace CPUs are designed to handle inference workloads and will work alongside the Vera Rubin systems, affirming Nvidia's comprehensive infrastructure strategy [4] Group 3: Investment and Infrastructure Plans - Meta's multiyear deal is part of a larger commitment to invest $600 billion in the U.S. by 2028 for data centers and necessary infrastructure [5] - The company plans to establish 30 data centers, with 26 located in the U.S., including two major AI data centers currently under construction in Ohio and Louisiana [5]