Nvidia Puts $3 Billion Into Meta’s Record $97.5 Billion Share Sale — and Into Making Its Muse AI Assistant Faster
The world's biggest chipmaker is now also one of the biggest backers of its own customer's expansion. Nvidia confirmed on Tuesday that it will invest $3 billion in Meta's equity offering and work more closely on Muse, the AI assistant at the centre of Mark Zuckerberg's "personal superintelligence" push — hours after Meta lifted its record fundraising to $97.5 billion. The partnership is real, the money is real, and so is the question hanging over both: how much of the AI boom is the industry funding itself?
Highlights
- Nvidia and Meta announced on Tuesday, October 27, 2026 that they will deepen their partnership on the Muse AI assistant and related models, with Nvidia investing $3 billion in Meta’s equity offering
- Meta raised its equity offering the same day to $97.5 billion — the largest fundraising in the company’s history — to accelerate investment in AI infrastructure
- The offering combines roughly $35 billion of common and preferred stock with a planned at-the-market programme of up to $60 billion tied to Meta’s new nonvoting Class C shares, expected to begin trading in early 2027
- Nvidia, AMD and Broadcom had earlier indicated interest in buying up to $5 billion of the shares between them; Nvidia’s confirmed $3 billion is the largest single slice disclosed
- Analysts at William Blair say they “expect investors to continue to question the financing and potential for future capital raises”, while believing Meta “will ultimately produce a profitable AI business”
The announcement: money and engineering in one deal
Nvidia and Meta said on Tuesday they would deepen their partnership on the Muse AI assistant and related AI models, with Nvidia investing $3 billion in the Facebook parent’s equity offering (Reuters, via The Indian Express). Meta’s Muse models and agents, which already run on Nvidia infrastructure, will work more closely with the chipmaker to improve performance and reduce costs (Reuters, via The Indian Express). TechSpot’s account of Nvidia’s confirmation adds the technical scope: the expanded partnership will focus on improving training and inference performance across Muse and its family of models while reducing the cost of running them, and Meta plans to expand its deployment of Nvidia’s Vera Rubin systems and Vera CPUs alongside Spectrum-X Ethernet networking and BlueField-4 technology (TechSpot).
The two companies may also explore working together more on Nvidia’s OpenShell framework, including using it to roll out software safely, and on a tool called Sentry to monitor how AI agents behave (Reuters, via The Indian Express).
What Muse is, and why Nvidia wants it running well
Muse, known internally as Hatch, is central to CEO Mark Zuckerberg’s plan to deliver “personal superintelligence” to the billions of people who use Meta’s services daily (Reuters, via The Indian Express). Modelled on the open-source AI agent OpenClaw, Muse can access a user’s email, calendar, payments, health, shopping and smart-home applications, with users able to revoke access at any time (Reuters, via The Indian Express). TechSpot describes the same product from the user’s side: an agent that can take action across everyday services, from sending emails to booking travel, with users deciding how much access to grant it, and notes the companies’ claim that Muse has delivered strong benchmark results and surging downloads since launch — a claim published without figures, and one readers should treat accordingly (TechSpot).
For Nvidia, the engineering logic is straightforward: an assistant designed to sit inside the daily routines of billions of users is an inference workload of a scale few products on earth can match, and every efficiency gain in running Muse is a gain on Nvidia hardware rather than a rival’s. That is also why the deal does not stand alone — in February, Meta and Nvidia announced a multi-year partnership covering AI GPUs, CPUs and networking technology (Asianet Newsable).
The $97.5 billion raise the $3 billion sits inside
Earlier on Tuesday, Meta increased its equity offering to $97.5 billion, the largest fundraising in its history, as it accelerates investment in AI infrastructure (Reuters, via The Indian Express). The structure, as reported by QNC News from Meta’s disclosures, enlarges the common stock offering unveiled on Monday to $25 billion, up from $22.5 billion, with shares priced at $625 each, alongside depositary shares priced at $50 apiece — together putting the public offerings of common and preferred stock at roughly $35 billion (QNC News). Asianet Newsable reports the same pieces with the quantities filled in: Meta sold $12.5 billion of preferred shares at $50 apiece and $25 billion of common stock at $625 per share (Asianet Newsable).
On top of that sits a separate at-the-market programme worth as much as $60 billion, which will only begin once Meta’s new nonvoting Class C shares start trading — expected in the first quarter of 2027 — and whose timing and terms are not yet settled (QNC News). Meta intends to use the proceeds to support its investment in AI infrastructure and its consumer AI tools (Asianet Newsable). Before Nvidia’s solo confirmation, Meta had said Nvidia, AMD and Broadcom had agreed to buy as much as $5 billion of the common shares in aggregate (Asianet Newsable) — so Nvidia’s $3 billion is the majority of the chipmakers’ combined interest, now confirmed rather than indicated.
The circular-money question, asked out loud
Supplier backing of this size has fed a worry that has followed the AI build-out all year: money moving in a circle through the sector, with chip companies helping to fund one of their largest buyers, whose spending then returns to them as orders (QNC News). It is the same pattern visible in Nvidia’s agreement to invest up to $100 billion in OpenAI as it deploys at least 10 gigawatts of AI data centres using Nvidia systems (The Motley Fool), and in the $475 billion-plus of compute commitments disclosed in Anthropic’s IPO filing.
How analysts and the market reacted
The scepticism is on the record rather than whispered. Analysts at William Blair said in a note that they “expect investors to continue to question the financing and potential for future capital raises”, adding that they believe Meta “will ultimately produce a profitable AI business”, though “the path may not be linear” (QNC News). Retail sentiment was “bullish” for both META and NVDA on Tuesday, according to Asianet Newsable, though the tape itself was more muted: NVDA was down 0.6% in overnight trading ahead of Tuesday while META shares were flat, with large-cap tech names mixed following Anthropic’s IPO filing (Asianet Newsable).
The bottom line: strip away the record-breaking headline and the deal has two separable halves. The engineering half — Nvidia optimising Muse’s training and inference on Vera Rubin systems, Vera CPUs and its own networking — is the kind of partnership both companies would plausibly sign with no money attached. The financial half, a supplier writing a $3 billion cheque into its customer’s record $97.5 billion raise, is what investors will keep circling: it binds the two companies closer, and it makes the AI boom’s funding loop a little more visible each time it happens.
