Nvidia Rules Out $100B OpenAI Bet as IPO Nears

The largest private funding round in the history of technology closed last week, and today its most prominent participant walked back the headline number attached to his name.

Speaking at the Morgan Stanley Technology, Media and Telecom Conference on Wednesday, Nvidia CEO Jensen Huang confirmed that his company's $30 billion commitment to OpenAI - part of a record $110 billion round announced on February 27 - will probably be the last direct equity check Nvidia writes to the AI startup. The $100 billion figure the two companies attached to a September infrastructure pact? "Not in the cards," Huang said. The reason is simple: OpenAI is going public.

History's Biggest Private Round

The numbers from the February 27 announcement deserve to be stated plainly. OpenAI raised $110 billion in a single financing round, placing it among the most heavily capitalized private companies in history. The round closed at a $730 billion pre-money valuation - $840 billion post-money - making OpenAI worth more than most of the S&P 500 by market capitalization.

Three investors put up the entire sum:

None of these commitments are simple cash transfers. Amazon's headline number is front-loaded toward a $15 billion check - the remaining $35 billion doesn't flow unless OpenAI either completes an IPO or reaches what the term sheet defines as artificial general intelligence. SoftBank's $30 billion is financed initially through bridge loans. As reported separately on this site, the Amazon deal also bundles an exclusive cloud distribution agreement that makes AWS the sole third-party platform for OpenAI Frontier and commits OpenAI to consuming 2 gigawatts of Trainium capacity.

The infrastructure obligations on all sides are as large as the financial ones. OpenAI commits to consuming gigawatts of compute from its new partners.

Why Huang Is Pulling Back

Nvidia entered the $110 billion round in an unusual position. In September, the company announced an infrastructure deal with OpenAI that carried a very different number: up to $100 billion in equity investment, to be rolled out gradually as each gigawatt of shared data center capacity came online, with a total target of 10 gigawatts.

Today, Huang put that figure to rest.

"We're going to invest $30 billion in OpenAI. I think the opportunity to invest $100 billion in OpenAI is probably not in the cards and the reason for that is because they're going to go public."

• Jensen Huang, Morgan Stanley TMT Conference, March 4, 2026

The logic is straightforward. Once OpenAI lists on a public exchange, Nvidia can buy shares on the open market like any other investor. A massive pre-IPO equity stake - the kind that made sense when OpenAI needed private capital to operate - becomes unnecessary, and potentially complicated for a public company with its own shareholders to explain.

Huang made the same point about Nvidia's $10 billion investment in Anthropic, saying that commitment would also likely be its last round of direct private investment in the rival AI lab.

This isn't a falling out. Nvidia's commercial relationship with OpenAI - 3 gigawatts of dedicated inference capacity and 2 gigawatts of Vera Rubin training - remains intact. The infrastructure deal runs regardless of the equity structure.

Counter-Argument: The Partnership Is What Actually Matters

It is tempting to read Huang's statement as a vote of reduced confidence in OpenAI. The opposite reading is more accurate.

Nvidia's real exposure to OpenAI's success is through hardware sales and the 10-gigawatt infrastructure agreement, not equity. If OpenAI grows, it buys more GPUs. Every dollar OpenAI spends on compute flows through Nvidia's product lines. The equity stake is almost incidental to that relationship.

The infrastructure commitment - deploy 10 gigawatts of Nvidia systems, with Nvidia investing up to $100 billion steadily per gigawatt - was always partly aspirational. The revised $30 billion covers the first few gigawatts and reflects where OpenAI actually is in its buildout today, not a decade-long projection. Huang isn't retreating from OpenAI; he's acknowledging that the venture equity model has a natural endpoint when a company files for an IPO.

For Nvidia, an OpenAI IPO is a net positive. It locks in OpenAI's capital requirements, gives Nvidia a liquid asset to hold or sell, and removes the ambiguity around future funding rounds. The $30 billion isn't a consolation prize - it's a clean final position before the IPO converts all of this into tradeable stock.

OpenAI's advisers are targeting a valuation of up to $1 trillion at IPO, up from the $730B pre-money set in last week's round.

What the Market Is Missing

The enthusiasm around the $110 billion round has obscured some structural details that matter for anyone paying attention to the long game.

OpenAI's 2025 revenue came in at approximately $13 billion against operating costs that pushed the company to a cash burn of around $9 billion for the year. The 2026 forecast is better - $30 billion in revenue - but costs are scaling faster. Inference costs alone are projected to hit $14 billion in 2026, and total cash burn is expected to reach $17 billion. OpenAI's own projections don't show cash-flow positive operations until 2030.

At $730 billion pre-money and $30 billion in projected 2026 revenue, the implied multiple is roughly 24 times forward revenue - elevated even by the standards of AI infrastructure plays.

As covered in our February VC record report, OpenAI's round was the single largest driver of February's record $189 billion startup funding month globally. That concentration of capital in one company is either visionary or alarming depending on your assumptions about the pace of AGI development and OpenAI's ability to maintain its market share as models commoditize.

The IPO math is equally conditional. Amazon's $35 billion doesn't transfer unless the company actually lists - or hits an AGI threshold that no one has formally defined. SoftBank's installments run through October. If the IPO timeline slips beyond the end of the year, as some advisers have suggested it might, the full $110 billion number becomes harder to defend.

None of this is a knock on the deal itself. OpenAI built a product used by 700 million people per week and is growing fast enough to attract three of the most capital-disciplined investors in tech. Huang's statement today does not change that picture. It just makes the arithmetic more honest.

Sources:

• Nvidia will not be able to invest US$100 billion in OpenAI due to IPO, CEO Jensen says - BNN Bloomberg
• OpenAI raises $110B in one of the largest private funding rounds in history - TechCrunch
• OpenAI secures another $110 billion in funding from Amazon, Nvidia and SoftBank - Engadget
• OpenAI and NVIDIA Announce Strategic Partnership to Deploy 10 Gigawatts - NVIDIA Newsroom
• 2026 IPO Watchlist: OpenAI, SpaceX and Other Tech Giants - Built In
• OpenAI's Annual Recurring Revenue Tripled to $20 Billion in 2025 - PYMNTS