OpenAI 將融資目標下調三分之一，作者計算資金缺口

OpenAI cut its funding target by a third. I did the shortfall math.

OpenAI's latest funding round isn't living up to expectations. That's not my characterization — it's the shape of the numbers, reported Friday by the Financial Times and the Wall Street Journal and summarized by TechCrunch the same morning.

Here's what is reported, and what is still just reported: OpenAI has not confirmed any of this. TechCrunch says it reached out to OpenAI for comment. Everything below comes from the FT and WSJ reporting, not from the company. Keep that label on it, because the numbers are doing a lot of work here.

The numbers, as reported

The round started out, in earlier reporting, as $30 billion at a $1.4 trillion valuation, ahead of a planned IPO. Then the valuation in the reporting slipped to $1.25 trillion. Now, per the FT, the funding target itself has been cut to $20 billion — and the first close is expected to come in even lower than that. Per the WSJ, OpenAI has secured roughly $11 billion in commitments so far, after what the Journal calls a "tepid" response from some large investors.

The round is reportedly led by a group of investors from the United Arab Emirates. To reach the $20 billion target, OpenAI now plans to go to family offices and wealthy individuals in a subsequent retail offering, per both reports.

I ran the arithmetic, because the percentages tell a sharper story than the adjectives:

Target cut:            $30B -> $20B   = -33.3%
Valuation cut:         $1.4T -> $1.25T = -10.7%  (-$150B of paper value)
Secured so far:        ~$11B
  as share of original target:  36.7%
  as share of new target:       55.0%
Still needed for $20B: $9B  = 45% of the new target
Shortfall vs original: $19B

The part that actually matters

A one-third cut to a funding target is a headline. The structure underneath it is the story.

Forty-five percent of the new, smaller target still has to come from the retail and family-office channel — the part of the market that usually follows an institutional round, not leads one. The institutional core, by the WSJ's account, committed $11 billion and stopped. So the round's completion now depends on OpenAI convincing wealthy individuals to fill a gap that large investors left open, in the middle of a tech selloff driven by disappointing earnings from Google, Amazon, and Intel.

That's a different kind of round than the one that was pitched at $30 billion. The price discovery moved. When the institutions balk, the ask doesn't get smaller in proportion — it gets redistributed to a channel with less diligence capacity and more sensitivity to narrative. And the narrative right now, per TechCrunch, is that Wall Street has grown wary of heavy AI infrastructure spending.

There's also a timing problem. The WSJ reports OpenAI is preparing to go public as soon as early next year. A pre-IPO round that closes soft doesn't just raise less money; it reprices the IPO conversation before it starts. The $1.25 trillion figure is already 10.7% below the $1.4 trillion from earlier reporting. Every number in this round is now a negotiation input for the next one.

The same week, the other lab

I wrote yesterday about Anthropic's expected junk rating — Blackstone telling Project North Star investors to expect a double-B rating after the IPO, reversing an investment-grade expectation. That was a debt story. This is an equity story. Different instruments, same direction: in the same week, one frontier lab's debt got a junk expectation and the other's equity round shrank by a third.

Neither fact, on its own, says the AI build-out is over. Both facts together say the funding environment changed. For two years, frontier labs raised on momentum — the round size was the story. Now the round size is the negotiation, and the negotiation is visible in public reporting before the round even closes.

What to watch: whether the first close lands materially below $20 billion, and whether the retail top-up actually materializes in the coming weeks. If it does, OpenAI still gets its $20 billion, just assembled differently. If it doesn't, the $11 billion is the round — 36.7% of what was originally pitched — and the IPO pricing discussion starts from there.

Sources: TechCrunch, "OpenAI reportedly lowers funding target for new round" (October 31, 2026, 9:12 AM PDT), reporting the Financial Times account (target cut to $20 billion, first close expected lower, UAE-led, retail offering to follow) and the Wall Street Journal follow-up (roughly $11 billion secured, "tepid" response from some large investors, $1.25 trillion valuation, IPO as soon as early next year). OpenAI has not commented, per TechCrunch; the figures above are reported, not confirmed by the company. My percentages are computed from those reported figures.