Meta has expanded its AI infrastructure partnership with Nvidia, deploying millions of GPUs and advanced CPUs across data centers to support large-scale AI workloads and next-generation applications.
Meta Platforms has announced a major expansion of its AI infrastructure partnership with Nvidia, marking a significant step in the company’s efforts to scale next-generation artificial intelligence workloads. The multiyear collaboration will see Nvidia supplying millions of GPUs, including the latest Blackwell models, as well as Grace and Vera CPUs, to support both training and inference operations across Meta’s global data centers.
As part of the agreement, Meta will integrate Nvidia’s hardware into a unified AI compute architecture, combining on-premises servers with cloud environments to increase efficiency, performance, and scalability. The partnership also extends to networking, with deployment of Nvidia’s Spectrum-X Ethernet platform to enhance low-latency connections for AI workloads, enabling faster processing across platforms.
The expanded deal includes the rollout of Confidential Computing solutions to enable privacy-enhanced AI processing, starting with applications in WhatsApp and other Meta services. This integration reflects Meta’s commitment to leveraging advanced AI infrastructure while maintaining security and data privacy standards.
Nvidia’s role in the partnership has grown from previous collaborations, positioning it as a central supplier of Meta AI chips and systems for large-scale projects. Analysts note that the combination of high-performance GPUs, CPUs, and advanced networking will support Meta’s growing portfolio of AI models, including foundation models used across multiple applications.
In addition to hardware deployment, the companies plan joint optimization of AI workloads to maximize utilization and reduce energy consumption, ensuring the infrastructure scales efficiently as AI demands increase. Meta engineers will work closely with Nvidia to tailor solutions for both current and future AI initiatives.
This latest expansion reinforces the strategic nature of the Meta-Nvidia partnership, building on years of collaboration in GPU compute, AI training, and network infrastructure. By integrating Nvidia’s latest hardware and software tools, Meta aims to create a flexible, high-performance foundation capable of supporting the next generation of AI applications.
As AI adoption accelerates across Meta’s platforms, this collaboration will provide the computational backbone necessary for large-scale AI workloads, supporting development, deployment, and experimentation across multiple teams and products.
