Meta Platforms is investing billions of dollars in Nvidia, significantly expanding their existing partnership to incorporate millions of additional AI chips and standalone central processing units (CPUs) into Meta’s data center infrastructure, according to announcements made this week.
The expanded agreement, confirmed by both companies, will see Nvidia supply Meta with a substantial volume of its GPUs over multiple years. While the precise financial terms remain undisclosed, sources indicate the deal is worth billions of dollars. This investment is intended to bolster Meta’s capacity for artificial intelligence workloads, powering its existing platforms – Facebook, Instagram, and WhatsApp – and supporting the development of new AI-driven features, and applications.
Beyond GPUs, the deal now includes the provision of standalone CPUs from Nvidia, a relatively new component of the partnership. This suggests Meta is seeking to build more customized and integrated AI systems, moving beyond solely relying on graphics processing for its AI tasks. CNBC reported that this expansion signals a deeper integration of Nvidia’s technology into Meta’s core infrastructure.
The move comes as Meta continues to prioritize AI as a key strategic area. The company has been actively integrating AI into various aspects of its operations, from content recommendation and ad targeting to developing generative AI tools for its users. The increased computing power provided by Nvidia’s chips will be crucial for training and deploying increasingly complex AI models.
For Nvidia, the deal represents a major win, securing a significant revenue stream and solidifying its position as a leading provider of AI infrastructure. The Financial Times noted that the agreement arrives as Nvidia faces increasing competition from rivals in the chip market. The partnership with Meta provides a stable and substantial customer, helping to offset potential challenges from competitors.
Nvidia highlighted the collaboration as a key component of building AI infrastructure, emphasizing the scale of the deployment with “millions” of its GPUs. The company did not provide a specific timeline for the delivery of the chips, but indicated that the rollout would occur over the coming years.
As of Wednesday, February 18, 2026, neither Meta nor Nvidia have publicly detailed the specific applications for the increased AI capacity, leaving the full scope of the investment’s impact on their respective product roadmaps unclear.
