Nvidia and Meta Platforms said on Tuesday they would deepen their partnership on Meta’s Muse family of artificial intelligence models and agents, with Nvidia investing $3 billion in Meta’s equity offering.
The collaboration will focus on improving AI performance, training and inference efficiency, while reducing operating costs. Meta also plans to expand its Nvidia-powered infrastructure as it develops AI agents for users worldwide.
Muse, known internally as Hatch, is central to CEO Mark Zuckerberg’s ambition to deliver “personal superintelligence” to billions of users.
Modelled on the open-source AI agent OpenClaw, Muse can access services including email, calendars, payments, health, shopping and smart-home applications, with users able to revoke access.
The companies said the models and agents run on Nvidia infrastructure and have recorded strong benchmark results and growing consumer downloads since launch.
Nvidia and Meta are also considering broader cooperation on Nvidia’s OpenShell framework and related tools.
OpenShell is designed to support safer AI-agent deployment, while Sentry monitors agent behaviour. Meta plans to deploy additional Nvidia technology, including Vera Rubin systems, Vera CPUs, Spectrum-X Ethernet networking and BlueField-4 technology, to improve computing and network efficiency.
Meta earlier increased its equity offering to $97.5 billion, the largest fundraising in its history, to accelerate AI infrastructure investment.
Nvidia CEO Jensen Huang said AI agents could change how people use computers by allowing software to understand objectives, reason and complete tasks.
Zuckerberg said the partnership would support the development of advanced personal agents and frontier AI models.
