The Nerd Uncle › Business

OpenAI raised a third. Of what it asked for.

The Wall Street Journal reports the company has raised about a third of its target, at a valuation below the one it wanted. Here is what that does and does not tell you.

31 OCTOBER 2026 · 4 min read

The Pioneer Building in San Francisco's Mission District. — The Pioneer Building in San Francisco, which housed OpenAI for its first several years. Photograph: HaeB (CC BY-SA 4.0)

OpenAI set out to raise $30 billion. It has raised about $11 billion. That is the Wall Street Journal's reporting, picked up this morning by The Decoder, and OpenAI has not commented.

The Pioneer Building in San Francisco, which housed OpenAI for its first several years. Photograph: HaeB (CC BY-SA 4.0)

Start with the caveats. This is one newspaper's account, attributed to unnamed people, about a round that is not finished. Fundraising closes in stages over months; a company can be well short at one point and whole at another.

What the figures say

The round is reported to value the company at $1.25 trillion, against the $1.4 trillion it was said to be seeking. It is now said to be looking for up to $20 billion more from family offices and wealthy individuals rather than the large institutions it started with.

That last detail is the interesting one. Family office money is slower and more patient — and typically approached once the obvious doors have been knocked on.

Wall Street. A valuation is the arithmetic of one deal, not a price anyone has paid. Photograph: Alex Proimos (CC BY 2.0)

A number is not a price

A valuation of $1.25 trillion does not mean anyone thinks OpenAI is worth $1.25 trillion. It is arithmetic: an investor pays a given sum for a given slice, you multiply out, and that is the headline. It describes what one buyer agreed with one seller on one day.

So the gap between $1.4 trillion and $1.25 trillion is not $150 billion of lost value. It is a change in the terms of a deal that has not closed.

The comparison everyone is making

Anthropic filed to go public last week and may list as early as November, on $16.8 billion of third-quarter revenue. Two companies, two routes to the same thing: both need enormous capital, and one is asking the public markets while the other asks private ones.

The idea to take away

Model research is cheap compared with running models at scale. This money is overwhelmingly for computers and the electricity to run them, which is why the fundraising never stops and why a cold month shows up so fast. When a lab raises or fails to raise a sum, the story is rarely about the research. It is about the power bill.

■