Nvidia backs Meta investment with $3 billion AI deal

Nvidia and Meta said on Tuesday, October 27, 2026, that they will deepen their partnership on the Muse assistant and related models, with Nvidia putting $3 billion into Meta's equity offering. The Meta investment sits alongside an agreement to optimise Muse models and agents running on Nvidia infrastructure for performance and cost. Muse, known inside Meta as Hatch, is central to chief executive Mark Zuckerberg's plan to deliver a personal superintelligence service to people who use Meta's apps every day. According to Reuters, the announcement landed on the same day Meta increased the size of its equity offering.

Earlier in the day, Meta increased its offering to $97.5 billion, the largest fundraising in its history, according to Reuters. Meta had announced on Monday a plan to raise funds in waves of stock offerings to bankroll spending on AI infrastructure. The common shares in the offering were priced at $625 each, according to Dow Jones Newswires. The fundraising shows how much cash the largest technology companies now need to build data centres, buy chips and train models, and it also shows how closely suppliers and customers are now tied together in that spending.

What the expanded partnership covers

Meta intends to expand its deployment of Nvidia infrastructure, including Vera Rubin and Vera central processing units for next-generation AI agents, the companies said in a statement. Meta will also deploy Nvidia Spectrum-X Ethernet networking and BlueField-4 to support network efficiency and throughput. The aim, according to the statement, is maximum performance per watt for agent workloads, which matters because agents run many steps in sequence and need fast responses. The companies intend to work on optimisations across workflows to improve performance and reduce total cost of ownership for Muse models and agents.

Muse models and agents, which run on Nvidia infrastructure, have delivered frontier-level benchmark performance and surging consumer downloads since launch, according to the companies. Muse is modelled on the open-source AI agent OpenClaw and can access email, calendar, payments, health, shopping and smart-home applications, Reuters reported. Users can revoke that access at any time, Reuters added. The access matters because an assistant that can act across apps is more useful than one that only answers questions, but it also raises the stakes for safety and control. The companies said they may also work together on Nvidia's OpenShell framework for safe deployment and use a tool called Sentry to monitor how agents behave.

Nvidia founder and chief executive Jensen Huang described AI agents as a "new computing paradigm" in which software understands goals, reasons, uses tools and completes work on behalf of users. He said optimising computing infrastructure end to end would help Muse serve billions of people and help those people achieve more. The statement frames agents as a shift from operating applications to working with intelligent agents that carry out tasks. That framing explains why Meta wants its assistant to work across messaging, search and commerce, and why Nvidia wants its hardware and networking tuned for agent workloads rather than for single chat answers.

Why the money matters now

For Meta, the offering provides cash for the infrastructure behind its AI plans at a moment when investors are asking how quickly that spending will pay off. Building and running agents at scale requires data centres, power, networking and a steady supply of advanced chips, and Meta has said it will keep spending heavily to stay competitive. For Nvidia, the stake aligns the chipmaker with a customer that is building agents for a mass audience, and it gives Nvidia a direct interest in Meta's fundraising success. Suppliers rarely write cheques of this size into a customer's offering, which is why the commitment drew attention in market coverage on Tuesday.

The deal also fits a wider pattern in which chipmakers, cloud providers and app developers are locking in supply and demand through partnerships rather than arm's-length sales. Meta needs reliable access to Nvidia systems as it expands agent features, while Nvidia needs evidence that its newest systems will be deployed at scale. The companies said in a statement that the announcement reinforces a long-term partnership supporting Meta's AI infrastructure ambitions. Readers following agent technology will recognise the same safety question: agents that act on behalf of users need guardrails, monitoring and a way to stop them when they go wrong.

Markets are also watching the Federal Reserve meeting that began on Tuesday, with investors weighing earnings from large technology companies against interest-rate expectations, according to GenZ NewZ coverage of the rate decision. Against that backdrop, a concrete supply and investment deal stands out because it is not a forecast. It is a signed expansion of work backed by a priced offering.

Sources: Reuters: Nvidia, Meta deepen AI ties and Nvidia and Meta company statement on expanding Muse collaboration.