Loading page…
Meta intends to deploy more Nvidia computing and networking systems. The companies aim to improve Muse’s performance and costs, while security collaboration remains under consideration.
Listen to this story
Meta is seeking $95 billion to fund AI infrastructure, and Nvidia has committed to a proposed $3 billion purchase in the offering. The companies are also widening their work on Meta’s Muse models and agents, covering training and inference as Meta plans to deploy Nvidia computing and networking systems. The stated aim is better performance per watt and lower operating costs; the announcement does not report achieved efficiency gains, and the share purchase has not closed.
Meta’s financing is split between $35 billion in underwritten public equity offerings and a further $60 billion through an at-the-market program.
Meta plans to use Vera Rubin systems, Vera CPUs, Spectrum-X Ethernet and BlueField-4 in the expanded infrastructure.
Nvidia and Meta will consider additional work on OpenShell for safe agent deployment and Sentry for monitoring; neither is described as a confirmed deployment.
Nvidia is putting money behind a customer that also intends to use more of its technology. On October 27, 2026, the chipmaker said it would invest $3 billion in Meta’s equity offering, alongside an expanded partnership to improve the computing behind Meta’s Muse AI models and agents.
The investment follows Meta’s October 26 announcement that it planned to raise a total of $95 billion to support AI infrastructure spending. That financing has two parts: an initial $35 billion in underwritten public equity offerings, followed by another $60 billion through an at-the-market program.
Meta had said Nvidia, AMD and Broadcom were interested in purchasing up to $5 billion in shares between them. Nvidia’s Tuesday announcement supplies a specific commitment from one of those prospective buyers. It is a proposed investment, not a disclosure that the share purchase has already closed.
AI agents represent a new computing paradigm — software that understands our goals, reasons, uses tools and gets work done on our behalf.
Jensen Huang, Nvidia founder and CEO, in the companies’ announcement
Muse launched in September, so this is a follow-on infrastructure partnership rather than the agent’s introduction. The companies said their expanded work would address both training the Muse family of models and inference—the computing used to run those models after training.
Meta intends to expand its use of Vera Rubin systems and Vera CPUs for next-generation agents. The stated objective is to increase performance per watt: getting more computing work from the electricity consumed. That is the companies’ deployment goal, not a measured improvement delivered by this announcement.
The buildout also includes Spectrum-X Ethernet networking and BlueField-4, which Meta plans to deploy to support network efficiency and throughput. The scope goes beyond processors: the companies are pairing computing systems with networking technology intended to help move data through the infrastructure.
Alongside that hardware deployment, Meta and Nvidia intend to optimize workflows to improve performance and total cost of ownership—the overall cost of owning and operating the systems. Their stated target is not simply more capacity, but more efficient operation of the infrastructure already underlying Muse’s models and agents.
The security portion carries a different level of commitment. Rather than announcing a deployment, Meta and Nvidia said they would consider additional collaboration involving Nvidia’s OpenShell Reference Architecture. They identified two technologies and their intended roles:
Loading discussion...
Join the conversation
Explain whether a supplier’s financial stake increases your confidence or gives you pause.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.