NVIDIA and Meta said they will deepen their collaboration on Muse and the Muse family of models while NVIDIA commits $3 billion to Meta's equity offering. In an Oct. 26, 2026 newsroom release distributed via GlobeNewswire on Oct. 27, the companies framed the work around inference and training performance for AI agents—a pairing of capital markets news and infrastructure strategy that lands a day after Meta outlined a massive equity raise to fund AI buildout. CNBC and the Wall Street Journal reported the investment on Oct. 27.
Key facts
| Item | Detail |
|---|---|
| Investment | NVIDIA will invest $3 billion in Meta's equity offering (NVIDIA newsroom); no share class, price, or timing disclosed |
| Partnership focus | Expand work on Muse and the Muse family of models for inference and training; optimize workflows for performance and total cost of ownership |
| Infrastructure | Meta intends to expand NVIDIA deployment: Vera Rubin, Vera CPUs, Spectrum-X Ethernet, BlueField-4 |
| Under consideration | Additional collaboration on NVIDIA OpenShell Reference Architecture, including OpenShell for safe deployment and NVIDIA Sentry for monitoring agent behavior—not yet agreed |
| Context | Meta on Oct. 26 announced plans to raise up to $95 billion for AI infrastructure; NVIDIA's release says it "will invest" $3B, vs. earlier non-binding cornerstone interest (see our Oct. 26 coverage) |
What NVIDIA and Meta announced
The release says NVIDIA and Meta will partner to expand their work on Muse and the broader Muse model line to improve inference and training performance, reflecting popular demand for useful AI agents. Meta's Muse models and agents run on NVIDIA infrastructure; according to the release, they have delivered frontier-level benchmark performance and surging consumer downloads since launch. The companies intend to partner on workflow optimizations to maximize performance and improve total cost of ownership.
NVIDIA CEO Jensen Huang said agents are "a new computing paradigm — software that understands our goals, reasons, uses tools and gets work done on our behalf," and that the companies are optimizing infrastructure "end to end for capable, efficient and secure AI agents." Meta CEO Mark Zuckerberg said Muse is how Meta brings personal superintelligence to everyone and that Meta is excited to keep working with NVIDIA on personal agents and frontier models.
The $3B investment and Meta's $95B raise
The NVIDIA release states: "In addition, NVIDIA will invest $3 billion in Meta's equity offering." It does not say whether that amount relates to the up-to-$5.0 billion aggregate non-binding cornerstone interest in Class A common stock that Meta's Oct. 26 release attributed to NVIDIA, AMD, and Broadcom. Treat the $3 billion as NVIDIA's own stated plan, not automatically part of the earlier indication.
CNBC reported the investment after Meta on Oct. 26 announced plans to raise $95 billion for AI infrastructure—initially $35 billion in underwritten public equity offerings and an additional $60 billion via an at-the-market program, with NVIDIA, AMD, and Broadcom interested in up to $5 billion in shares. See Meta Plans Up to $95B Stock Sale to Fund AI Buildout for how Meta split concurrent offerings from the Class C ATM it does not expect before Q1 2027. Meta's cornerstone language was non-binding; NVIDIA's release says it will invest $3 billion.
What Meta is deploying
Meta intends to expand NVIDIA infrastructure for next-generation AI agents, including Vera Rubin and Vera CPUs for performance per watt, plus Spectrum-X Ethernet and BlueField-4. That aligns with Meta's heavy AI capex plans outlined in its Oct. 26 financing release. For Muse context, see Meta Bets Up to $145B on AI as Muse Tops 10M Accounts, Meta Opens Muse Spark Weights: 816B Model Runs on Two Macs, and NVIDIA DSX POD Cuts AI Factory Deployment to 14 Days.
Agent safety tools under consideration
The companies will consider additional collaboration on the NVIDIA OpenShell Reference Architecture, including OpenShell for safe deployment and NVIDIA Sentry for monitoring agent behavior. That language is exploratory, not a signed deal—but it signals that deployment safety and observability may ride alongside Muse scale.
Why it matters
The story links consumer agents, massive AI capex, and strategic equity between Meta and a key chip supplier. NVIDIA's $3 billion commitment backs Meta's financing while the two plan joint optimization work on how Muse runs on Vera-era systems—not only how many chips Meta orders.
What to watch next
Prospectus filings may clarify timing and structure for NVIDIA's $3 billion stake; the release gives no price, share class, or schedule. Watch whether AMD and Broadcom issue their own investment announcements, and whether OpenShell and Sentry move from "consider" to production guardrails for Muse agents.
Sources
- NVIDIA Newsroom: NVIDIA and Meta to Expand Collaboration on Muse and Muse Family of Models Amid Robust Growth — Oct. 26, 2026
- GlobeNewswire: NVIDIA and Meta to Expand Collaboration on Muse and Muse Family of Models Amid Robust Growth — Oct. 27, 2026
- CNBC: Nvidia to invest $3 billion in Meta as tech giants deepen AI partnership — Oct. 27, 2026
- Wall Street Journal: Nvidia to Invest $3 Billion in Meta's Equity Offering as They Expand AI Partnership — Oct. 27, 2026