NVIDIA to Invest $3 Billion in Meta Offering as AI Financing Reaches Record
Key Facts
NVIDIA said it would invest $3 billion in Meta’s stock offering, as the companies announced on October 26 that they were expanding their collaboration on Muse models and the Muse family. The joint work aims to improve training and inference performance, while Meta plans to expand its deployment of NVIDIA’s AI infrastructure. The investment announcement represented a chip supplier participating in the financing of a major customer, alongside Meta’s growing reliance on the company’s technology. The announcement did not specify that NVIDIA’s funds would directly finance Meta’s infrastructure plans. The following day, Meta raised the amount it was seeking through equity to $97.5 billion, putting the cost of its AI expansion at the center of the news.
On October 26, Meta announced a concurrent offering comprising about $22.5 billion of Class A shares and about $12.5 billion of mandatory convertible preferred depositary shares. The company then said on October 27 that it had increased the value of its common stock offering from $22.5 billion to $25 billion, priced at $625 per share. It also priced 250 million depositary shares at $50 each, bringing total announced financing to $97.5 billion. The financing structure includes nonvoting Class C shares, allowing the company to raise capital while preserving the existing voting rights of Mark Zuckerberg and other shareholders with voting rights. According to Investing.com, citing Reuters, proceeds from the offerings will be used for general corporate purposes, including investment in AI infrastructure.
The financing comes alongside high capital spending: Meta raised the lower end of its 2026 outlook to $135 billion from $130 billion, while keeping the upper end at $145 billion. The company had forecast capital spending of between $135 billion and $145 billion in its filing. In the third quarter of 2026, revenue was $64.601 billion, up 26% year over year. But free cash flow turned negative at $2.12 billion, after reaching positive $10.63 billion in the third quarter of 2025, according to Reuters. By contrast, operating cash flow totaled $97 billion in the first nine months of 2026, and the company held about $82 billion in cash and marketable securities at the end of September.
Meta’s results show its user base expanding alongside this investment drive: the company said its family of apps averaged 3.64 billion daily active users in September 2026, up 3% from a year earlier. Revenue and active-user growth give the company substantial operating resources, while projected spending and negative free cash flow put infrastructure financing under investor scrutiny. In the October 26 session, the stock fell by as much as 4% before reversing course to rise 1%, according to a report attributed to Reuters and published by Investing.com. Meta shares closed at $630.56 that day, while NVIDIA shares ended the session at $216. The closing price alone does not establish investors’ view of the financing, but it followed notable volatility during the session.
Shares of both companies came under pressure on October 22, when Meta fell 7%, its biggest decline since July, and NVIDIA dropped 4% amid concerns about the cost of AI expansion and uncertainty over returns. Reuters quoted analyst Matthew Maley as saying that the question of returns on spending had begun to drive selling, adding: "It’s becoming clearer and clearer to people that this is going to take a lot longer." After the financing announcement, William Blair analysts said investors would continue to question the sources of financing and the possibility of raising additional capital. They added that the plan reflects Meta’s confidence in AI returns, as well as the rising capital intensity required to keep competing in the field. The debate is therefore centered on whether business growth can offset the cost of computing and its financing.
On the technology front, Meta plans to deploy more NVIDIA systems, including Vera Rubin and Vera processors, as well as Spectrum-X Ethernet and BlueField-4 networking. The collaboration combines expanding computing capacity with work on Muse models, with a stated focus on training and inference performance. The companies said they would explore further collaboration opportunities on OpenShell for secure deployment and Sentry for monitoring agent behavior. The announcement did not specify a timeline for launching these opportunities or details on their implementation. The next clear step is to execute the announced offerings and expand the technology infrastructure, while markets monitor the impact of spending on cash flows and returns.