NVIDIA and Meta said they will deepen work on Muse and related models, while Meta plans broader NVIDIA hardware deployment and NVIDIA invests $3 billion.
NVIDIA and Meta said they will expand their collaboration on Muse and the Muse family of models, a move the companies said is aimed at improving inference and training performance for AI agents. The announcement also included plans for broader infrastructure deployment and a separate $3 billion NVIDIA investment in Meta's equity offering, underscoring the scale of their long-running partnership.
In the companies' statement dated Oct. 27, 2026, NVIDIA said the expanded work on Muse and the Muse family of models reflects demand for AI agents and is intended to optimize workflows, improve performance and reduce total cost of ownership. NVIDIA chief executive Jensen Huang said the companies are working to optimize computing infrastructure end to end for AI agents, while Meta chief executive Mark Zuckerberg said Muse is intended to bring personal superintelligence to users.
The announcement said Meta intends to expand deployment of NVIDIA infrastructure, including Vera Rubin and Vera CPUs for next-generation AI agents, and will also deploy NVIDIA Spectrum-X Ethernet networking and BlueField-4 to support network efficiency and throughput. The companies also said they are considering additional collaboration on NVIDIA OpenShell Reference Architecture, including NVIDIA OpenShell for safe deployment and NVIDIA Sentry for monitoring agent behavior.
The update follows a series of Muse-related product releases from Meta in 2026. Meta AI Research said on Oct. 15 that it released Muse Spark Open, a downloadable multimodal model under the Muse Model License with limited commercial terms.
Source link







