OpenAI Raised Half the Money at an 11 Percent Lower Mark

Abhishek Dash11 min read
Two blocks of tightly stacked blank paper rest on a dark wooden desk, the upper block offset from the lower so their cut edges form a step, beside a blank brass tag on a braided cord, with a taller stack standing out of focus in the top right corner.

In short

OpenAI had raised 15 billion dollars at a 1.25 trillion dollar valuation as of 12 November 2026, CNBC reported, citing David Faber. In late September the company was reported to be seeking up to 30 billion dollars at a 1.4 trillion dollar valuation, so the round came in at half the capital and roughly 11 percent below the planned mark. OpenAI has not confirmed the figures.

Key takeaways

  • CNBC reported on 12 November 2026 that OpenAI had raised 15 billion dollars at a 1.25 trillion dollar valuation, citing David Faber.
  • Bloomberg reported on 29 September 2026 that OpenAI was targeting up to 30 billion dollars at a 1.4 trillion dollar valuation, so the reported round is about half the capital at a mark roughly 11 percent lower.
  • Bloomberg reported on 6 November 2026 an 11 billion dollar first close at 1.25 trillion, inside a round that could total about 20 billion dollars and had not closed.
  • TechCrunch reported on 11 November 2026 that OpenAI's quarterly revenue crossed 10 billion dollars for the first time, which is about 40 billion dollars annualised and roughly 31 times the 1.25 trillion dollar mark.
  • Nvidia has invested at least 30 billion dollars in OpenAI, per CNBC, and agreed to buy 10 billion dollars of Anthropic stock in its IPO, which is expected to trade on 19 November 2026.
  • OpenAI has published no funding announcement and did not respond to TechCrunch, so every figure here is reported rather than filed.

Fifteen billion dollars at a $1.25 trillion valuation. CNBC reported those two figures for OpenAI on 12 November 2026, attributing them to David Faber. In late September, Bloomberg reported that the same company was seeking up to $30 billion at a $1.4 trillion valuation.

That is half the capital and a mark about 11 percent lower than the plan.

OpenAI has confirmed none of it. The company is private, it has published no funding announcement, and it did not respond to TechCrunch when that outlet reported the earlier close on 6 November. Every number in this post rests on unnamed people speaking to CNBC or Bloomberg. For the largest private financing of the year, that is a thin floor, and it belongs at the top.

The business went the other way. TechCrunch reported on 11 November 2026 that OpenAI's quarterly revenue crossed $10 billion for the first time. Capital became more expensive in the same fortnight that the company became larger.

That is the shape of it. Not a collapse, and not a triumph.

What CNBC actually said

The Faber item ran inside CNBC's live coverage of the Nvidia board story, not as a standalone article, and the standalone version did not surface in search when this post was written. The wording is that OpenAI had been originally planning to raise up to $30 billion at a $1.4 trillion valuation before pulling back, and that it had raised $15 billion at $1.25 trillion.

Two verbs carry the story. Planned and raised. The first is the September intention as Bloomberg reported it on 29 September 2026. The second is what CNBC reported six weeks later.

Between them sits $15 billion of capital that did not arrive, at a price $150 billion below the September plan if you subtract one valuation from the other. They are comparable in the only sense that matters here: both are numbers the reporting put on the same company eight weeks apart, one as a target and one as a result.

The difference between them is the most informative thing published about AI valuations this quarter.

Six days earlier, the total looked bigger

Bloomberg's 6 November 2026 report described an $11 billion first close at a $1.25 trillion valuation, inside a round that had not closed and could total about $20 billion. TechCrunch relayed it the same day, alongside the September target.

Read the two reports together and the revision is in the capital, not the price. The valuation held at $1.25 trillion from 6 November to 12 November. What moved was the expected total, from about $20 billion to a reported $15 billion, a difference of a quarter.

There are two readings. Either the round closed smaller than buyers expected six days earlier, or the outlets are counting different things: commitments, closed capital, or a headline size that folds in a slice raised before. The reports do not settle it and this post cannot.

The distinction is not pedantry. A total that shrinks after the first close says something about demand from the marginal investor. A definitional difference says something about how the round was described.

The reported figures, in order

Date Outlet What was reported Capital Valuation
29 September 2026 Bloomberg Target for the round up to 30 billion dollars sought 1.4 trillion dollars
6 November 2026 Bloomberg First close, round still open 11 billion dollars closed, about 20 billion expected 1.25 trillion dollars
12 November 2026 CNBC Reported total, per David Faber 15 billion dollars raised 1.25 trillion dollars

The last row is the one to quote. The middle row is the one that shows how fast the expectation moved, and the first row is the one that makes the last row a shortfall rather than a number on its own.

A smaller round is not a down round

$1.25 trillion is not a markdown in the sense that phrase usually carries. The company was not repriced below its previous private valuation. It was priced below the valuation it had hoped for. Those are different events, and the difference accounts for most of the commentary.

What the round does show is that the marginal buyer in November would not pay the September price. At $1.4 trillion, a $30 billion check was available. At $1.25 trillion, half that capital was.

An 11 percent difference in price produced a 50 percent difference in capital. Either the demand curve for OpenAI equity is steep, or the binding constraint was never price but the size of check any single investor was willing to write into one private company. The reporting does not separate those, and the arithmetic cannot.

Who is actually paying

Nvidia has invested at least $30 billion in OpenAI, according to CNBC, and has agreed to buy $10 billion of Anthropic stock in its initial public offering. The company selling the accelerators is now one of the largest financiers of the labs buying them.

That is legal, and it is disclosed at the aggregate level in Nvidia's own filings. It matters because of what it does to the price signal. A private round priced mostly by strategic investors who also sell to the company is not an arm's length mark, and the buyers have interests that run past the equity.

The $30 billion is CNBC's figure. It was not reconciled to a line item in Nvidia's quarterly filing, which discloses guarantees and investments without breaking out this one by name. Treat it as reported, not as a disclosed position.

The consequence is narrow. $1.25 trillion is a number negotiated between parties with interests beyond the shares, not what a diversified public buyer would pay for the same cash flows. That is a limit on the figure, not an accusation.

The Anthropic print lands on 19 November

The cleanest comparison arrives within the week. Anthropic is expected to trade on 19 November 2026. Bloomberg reported on 6 November that potential IPO buyers believed its valuation could be as low as $1.1 trillion, against the $2 trillion investors were said to expect about a month earlier. Anthropic's own range implies up to roughly $1.4 trillion fully diluted.

The $1.1 trillion is a buyer's whisper, not a cleared price. Nobody has paid it. It still describes the same shift in the same fortnight: expectations for a frontier lab falling by close to half with no reported change in the business.

If Anthropic clears near the top of its range, OpenAI's $1.25 trillion looks conservative and this post's framing is too dark. If it clears near the bottom, the private mark is running ahead of the public market. That is a test with a date on it, which is more than most valuation arguments get.

Why the mark moved

The mechanical explanation is that private marks lag public ones. The AI trade sold off through October and into November 2026 on rate and overspending concerns, and the midterm result and data centre ballot measures added political risk to the buildout story.

That is a reading of the tape, not something either outlet reported as the cause of this round. Nobody at CNBC or Bloomberg is quoted saying rates did it. The mechanism is plausible; the causation is unestablished.

What can be said from the reporting is narrower and still useful. The direction across six weeks of coverage is down. A valuation target trimmed about 11 percent. A capital target cut in half. A peer's expected IPO valuation halved in the same window. Three separate prints pointing the same way is not proof of a cause. It is enough to say the market changed its mind about duration.

What $10 billion a quarter does to the arithmetic

If a quarter at that rate repeats four times, which is arithmetic and not a forecast, OpenAI is running at roughly $40 billion a year. At $1.25 trillion that is about 31 times annualised revenue. At the September target of $1.4 trillion it would have been 35 times.

Both multiples are high for a company of that age. Both are far below the ones that show up in arguments about an AI bubble, which usually rest on revenue that does not exist yet. The reported revenue is the strongest fact in this post, and it is still a TechCrunch report rather than an audited statement.

The revenue print also removes the easy explanation. Demand did not fall in any of the reporting. The company did not lose customers or cut prices. It grew faster than it had before and raised half of what it wanted.

A company can raise less because investors lost interest, or because it needed less, or because the price of the capital it was offered stopped being worth paying. The first is contradicted by the revenue print. The second and third are compatible with it, and they imply very different things about what happens in 2027.

The 2027 IPO is the next print

Reuters reported on 5 November 2026 that OpenAI is expected to go public in 2027, and that a majority of its sales now come from enterprises, according to president Brad Brockman. That timeline makes the private mark a staging post rather than a final price.

Everything downstream of $1.25 trillion is now set against it. Employee option marks. Secondary sales. Strike prices in new hire packages. Whatever the company eventually tells underwriters about its last private round. A number reported by anonymous sources and confirmed by nobody becomes the reference point for all of it, because there is nothing else to use.

That is the part a later filing will not correct so much as replace. The prospectus, when it exists, will have to describe this round. Until then the reported figures are the record, and the record is other people's sources.

What is documented and what is not

Claim Where it comes from Grade
OpenAI raised $15 billion at $1.25 trillion, about half the target CNBC, David Faber, 12 November 2026 reported
The original plan was up to $30 billion at $1.4 trillion Bloomberg, 29 September 2026 reported
An $11 billion first close at $1.25 trillion, round open, about $20 billion expected Bloomberg, 6 November 2026 reported
Quarterly revenue crossed $10 billion for the first time TechCrunch, 11 November 2026 reported
Anthropic IPO buyers eye a valuation as low as $1.1 trillion Bloomberg, 6 November 2026 reported
Nvidia has invested at least $30 billion in OpenAI CNBC secondary
OpenAI expects to IPO in 2027, enterprise is now the majority of sales Reuters, 5 November 2026 reported
OpenAI has confirmed none of the above OpenAI news index, no funding announcement primary, by absence

Where this could be wrong

The largest risk is that $15 billion is a partial close rather than a final total. On 6 November the round was explicitly open. CNBC's 12 November figure could be a second close on the way to a larger number, in which case the shortfall is smaller or temporary. The accurate phrasing is that OpenAI had raised $15 billion as of 12 November 2026, not that it closed at $15 billion.

There is no primary document. OpenAI is private, there is no S-1 or annual report, and the company has published nothing about this round. Every figure is attributed to people described as familiar with the matter. The gap between what those people told two outlets and what actually sits in OpenAI's accounts is unknowable from outside.

Raise, round and total are used inconsistently across the reporting. Committed capital, closed capital and headline round size are three different quantities, and the sources do not always say which one they mean. Part of the apparent $20 billion to $15 billion revision may be that ambiguity rather than a change in demand.

The revenue figure is also reported. $10 billion a quarter is TechCrunch's number, not an audited statement, and the multiple of 31 is arithmetic performed on two unverified inputs. If either is off, the multiple is off by the same proportion.

The Anthropic comparison is sentiment. $1.1 trillion is what potential buyers were said to believe, not a price anyone paid. The 19 November debut will produce a real number and may not resemble the whisper.

Nvidia's $30 billion was not reconciled to its own filing. That filing discloses guarantees and investments but does not break out this line by name, so the figure is CNBC's characterisation of a position rather than a disclosed figure. The 10-Q also covers the quarter before this round, so it cannot show what Nvidia committed in November.

Paywalled originals were not read. The Bloomberg and Reuters accounts are cited through relay coverage and the dossier's retrieval record. Their sourcing, their caveats and their exact wording were not verified directly, and CNBC's standalone funding article did not surface, so the numbers come from a live blog item that has since been scrolled past by newer entries.

Motive is inferred throughout. More than one reading of the same six weeks of reporting fits the evidence: a market repricing, a company needing less capital, or a round still in progress. If OpenAI announces the round, if Anthropic prices at the top of its range, or if a later close raises the total, parts of this post will need revising rather than defending.

On this page

Sources

  1. Nvidia live updates: OpenAI raises $15 billion at a $1.25 trillion valuationCNBC, 2026
  2. OpenAI reportedly raises an additional $11 billion at a $1.25 trillion valuationTechCrunch, 2026
  3. OpenAI completes $11 billion portion of funding at $1.25 trillion valueBloomberg, 2026
  4. OpenAI targets $30 billion in new funding at $1.4 trillion valueBloomberg, 2026
  5. Anthropic IPO Buyers See Valuation as Low as $1.1 TrillionBloomberg, 2026
  6. OpenAI quarterly revenue reportedly crosses $10B for the first timeTechCrunch, 2026
  7. OpenAI ready for IPO in 2027, now counts majority sales enterprises, Brockman saysReuters, 2026
  8. OpenAI news index, with no funding announcement as of 12 NovemberOpenAI, 2026

Frequently asked questions

How much did OpenAI raise in November 2026?

CNBC reported on 12 November 2026 that OpenAI had raised 15 billion dollars at a 1.25 trillion dollar valuation, citing David Faber. In late September 2026 Bloomberg reported the company was seeking up to 30 billion dollars at 1.4 trillion. OpenAI has not confirmed either figure and no filing or press release exists for the round.

What is OpenAI worth after this round?

The reported private valuation is 1.25 trillion dollars, unchanged from the 11 billion dollar first close Bloomberg described on 6 November 2026. That is about 11 percent below the 1.4 trillion dollar figure OpenAI was reported to be targeting in September 2026.

Why did OpenAI raise less than it planned?

The reporting does not give a reason. What it shows is that an 11 percent lower price came with roughly half the capital, at a company whose reported quarterly revenue crossed 10 billion dollars in the same fortnight. Whether the constraint was price, or the size of check any single investor would write, is not established by CNBC or Bloomberg.

Has OpenAI confirmed the 15 billion dollar raise?

No. OpenAI has published no funding announcement, and it did not respond to TechCrunch when that outlet reported the 6 November 2026 close. The company is private, so there is no S-1, 8-K or equivalent document to check the numbers against.

How does Anthropic's IPO compare with OpenAI's valuation?

Anthropic is expected to trade on 19 November 2026. Bloomberg reported on 6 November that potential IPO buyers believed its valuation could be as low as 1.1 trillion dollars, against 2 trillion dollars investors were said to expect about a month earlier. That is buyer sentiment, not a cleared price.

When is OpenAI expected to go public?

Reuters reported on 5 November 2026 that OpenAI is expected to IPO in 2027, and that a majority of its sales now come from enterprises, according to president Brad Brockman. No filing has been made and no date has been set.