Meta expands Nvidia deal to use millions of AI chips in data center build-out, including standalone CPUs

Meta has expanded its partnership with Nvidia in a multi-billion dollar agreement to deploy millions of GPUs and new standalone CPUs across AI data centers, continuing the tech giant's aggressive expansion into artificial intelligence infrastructure. The expanded deal represents one of the largest AI hardware partnerships in the tech industry, with analysts estimating the value could reach tens of billions of dollars over several years. Meta's commitment to acquiring millions of Nvidia's latest chips underscores the company's ambitious plans to enhance its AI capabilities, which power everything from recommendation algorithms to virtual reality experiences. This move comes as Meta faces increasing competition in the AI space from companies like Google, Microsoft, and OpenAI, all of whom are investing heavily in AI infrastructure. The inclusion of new standalone CPUs in the deal is particularly noteworthy, as it suggests Meta is looking to optimize both AI processing and general computing tasks in its data centers. Nvidia's Grace Hopper superchip, which combines CPU and GPU capabilities, appears to be a key component of this strategy. By leveraging these advanced chips, Meta aims to improve the efficiency and performance of its AI workloads while potentially reducing energy consumption and operational costs in its rapidly expanding data center footprint.