OpenAI has lowered the target for its latest funding round from $30 billion to $20 billion, according to a Financial Times report published Friday. The company now expects a first close below that figure, with a retail offering for wealthy individuals meant to top up the total.
The Wall Street Journal added detail: OpenAI has secured about $11 billion so far, and large investors responded to the roadshow with what the paper called a "tepid" reception. The round is led by investors from the United Arab Emirates.
This is the first time OpenAI's fundraising has visibly stalled, and it matters because the company is the anchor tenant of the AI boom. If the firm that kicked off the ChatGPT era can't raise at the speed it planned, the capital flowing into AI infrastructure, from data centers to chip orders, slows with it.
What changed between the $30 billion plan and now
OpenAI was earlier reported to be seeking $30 billion at a $1.4 trillion valuation, then at $1.25 trillion, ahead of a planned IPO. The cut to $20 billion lands in the middle of a broad tech selloff driven by worries about heavy AI infrastructure spending.
Wall Street has grown wary of the money required to build and run frontier models. Investors who once treated AI capex as a growth story now want evidence of revenue to match. OpenAI's own costs are part of the problem: training and serving models at scale burns capital faster than the company can bill for it.
The timing hurt. The fundraising push coincided with a stock rout that made large allocators cautious, and Anthropic's own IPO plans reportedly face headwinds too. When two of the biggest private AI companies struggle to price rounds in the same month, the signal is hard to dismiss as a one-off.
The money that has actually landed
Of the $20 billion target, roughly $11 billion is committed, per the Journal. That leaves a gap of about $9 billion to be filled by later closes, including the planned retail offering for wealthy individuals.
The UAE-led investor group is the backbone of the round. Sovereign wealth from the Gulf has been the most reliable source of large AI checks this year, and OpenAI's decision to lean on it reflects both the scale of the need and the limits of traditional venture appetite.
It's not clear this will hold. A first close below target followed by a retail top-up is a structure companies use when institutional demand is thinner than hoped. It works, but it costs time and dilutes the story OpenAI tells before an IPO.
How this compares to earlier rounds
OpenAI's previous raise, completed in early 2026, was oversubscribed and closed quickly. The contrast with this round is stark. Investors who queued up six months ago are now asking for audited revenue projections and detailed compute cost breakdowns.
The shift shows up in term sheets too. Earlier rounds carried light governance conditions. This one, according to people familiar with the talks, includes tighter reporting requirements and milestone-based tranches. That is the kind of structure lenders use, not equity investors chasing a moonshot.
For founders watching from the outside, the lesson is uncomfortable. Even the most hyped company in the sector now has to earn its capital. The days when a pitch deck and a demo were enough to raise $30 billion are gone, at least for now.
Why investors are hesitating
The core worry is infrastructure. Training runs for frontier models require tens of thousands of GPUs, power contracts, and data center leases that lock in spending for years. Investors are asking whether the revenue from ChatGPT subscriptions and enterprise contracts can carry that load.
OpenAI has revenue, and it is growing. But the gap between what the company spends and what it earns is wide enough that even believers are trimming their checks. Some large investors described their response as tepid, a word that carries weight in a market where "oversubscribed" was the norm a year ago.
Anthropic's planned IPO facing similar challenges adds to the unease. If the two most credible AI labs both find the public market less welcoming than expected, the assumption that AI companies can fund themselves at any price starts to crack.
What this means for the AI spending race
The pullback does not mean AI investment is collapsing. Tenstorrent, an AI chip designer, closed a $1.5 billion Series E this year, and funds like MGX have raised tens of billions for AI-focused vehicles. Capital is still moving; it is just moving more selectively.
What changes is the price of ambition. A company that could once raise $30 billion on a vision now has to settle for $20 billion and a retail tranche. That forces trade-offs: fewer training runs, slower data center builds, tighter hiring. Each one feeds back into the model release schedule.
For competitors, the opening is real. Anthropic, Google, and Meta all watch OpenAI's funding closely, and a constrained rival is easier to catch. The race does not stop, but the leader's pace is no longer taken for granted.
The IPO question
OpenAI's planned IPO sits behind all of this. A lower funding round and a tepid institutional response are not the conditions under which a company wants to go public. The valuation, once floated at $1.4 trillion, now rests on a $1.25 trillion figure that itself looks soft.
The path forward likely runs through the retail offering and a second close. If those fill the gap, OpenAI can still fund its 2027 compute buildout and keep its IPO timeline intact. If they fall short, the company faces a harder choice: cut spending or accept a lower valuation.
Neither option is fatal. OpenAI remains the most valuable private company in the world on paper. But the era of unlimited checks for AI infrastructure is over, and the company that defined it is the one learning that first.
See more on the broader AI funding picture in our desk coverage, and read the original reporting from TechCrunch and Reuters.
OpenAI declined to comment to TechCrunch. The next signal will come from the size of the first close and whether the retail tranche fills. If it lands near $20 billion, the narrative of a temporary wobble holds. If it lands well below, expect the valuation and the IPO timeline to move together.