OpenAI is in early talks to raise at least 30 billion dollars at a pre-money valuation of around 1.4 trillion dollars, according to Bloomberg. The discussions were reported on 29 September and the terms could change.
The round is described as a bridge, providing capital in place of an initial public offering. Sam Altman had said earlier in September that the company would not go public in 2026, citing concerns related to AI safety.
Six months earlier, in March, OpenAI closed a round with 122 billion dollars of committed capital at an 852 billion dollar valuation. That one was widely characterised at the time as the last private raise before a listing.
So the valuation has risen roughly 64 per cent in half a year, and the last round has been followed by another round.
Both halves of that sentence matter and they point in different directions.
The upward half is grounded in something real. Annualised revenue run rate is reported to be approaching 70 billion dollars, having risen more than 70 per cent since the start of the third quarter, against a figure just past 40 billion in mid-August. The company has disclosed 1.2 billion weekly users and 2.5 million enterprise customers. Those are not projections. A valuation moving on numbers like that is not a mood.
The other half is the word bridge. A bridge round exists to get a company from where its money runs out to the event that refills it. When the event is postponed, the bridge has to be extended, and the honest reading of a second raise six months after a final raise is that capital is going out faster than the previous plan assumed. In this case that is not mismanagement, it is physics: compute and infrastructure costs for frontier model development escalate with ambition, and the ambition has not been reduced.
Any founder who has raised money will recognise the smaller version of this with some discomfort.
You raise what you are confident is the last round. You say so to the team, to the bank, and to yourself, because the alternative framing is demoralising and because at the time you believe it. Then the market moves, or an opportunity appears that cannot be refused, or the build takes nine months longer than the model said, and you are back in front of investors having described the previous conversation as final.
The reputational cost of that is lower than most founders fear, provided one condition holds. The business has to have grown into the previous raise. Investors forgive a revised plan attached to a revenue chart. What they do not forgive is a second raise at the same scale of business as the first, which is the situation where bridge means something else entirely.
OpenAI is raising more money than it said it would, at a far higher price, against a revenue line that has nearly doubled in a quarter.
The bridge is longer because the far bank moved.
