OpenAI called March its last round. Now it wants $30 billion more.

If you've been using ChatGPT this week — a recipe, a cover letter, help with your kid's homework — the company behind it is quietly asking investors for another $30 billion. Not instead of an IPO. In addition to it. And the IPO just moved to 2027.

Bloomberg reported on September 29 that OpenAI is in talks to raise at least $30 billion more at a valuation of roughly $1.4 trillion. TechCrunch confirmed the details, including a run-rate revenue figure of around $40 billion in August. This is a good-news data point and a concerning one at the same time.

What's actually happening

Here's an object lesson, because the numbers can be fiddly.

Think about what happened when Netflix first launched: DVDs by mail, cheap subscription, growing fast. Then came streaming, and the content budget went from hundreds of millions to hundreds of billions. The economics of "we'll spend whatever it takes to grow first" are familiar. Netflix got there eventually. But there were years in between where it raised capital faster than it earned it.

OpenAI is in that in-between phase, except compressed and at a scale Netflix never touched.

The company did about $40 billion in annualized run-rate revenue as of August — that's an estimate based on monthly numbers, not a final audited figure. That's genuinely impressive. But running frontier AI models at the scale ChatGPT operates requires server farms, energy, and research budgets that dwarf even those numbers. OpenAI's 2025 books (audited figures leaked in June) showed $13 billion in revenue against $34 billion in costs. The product grows. The gap closes. It has not closed yet.

The $30 billion is called a bridge financing — a term that means "capital to get us to our next milestone." The milestone is an IPO, the public stock offering that OpenAI has been planning since at least 2025. The IPO got pushed from 2026 to 2027. The bridge is how it pays the bills in the meantime.

Two sovereign funds are reportedly in talks to participate: SoftBank Group and MGX (Abu Dhabi's state-managed AI investment vehicle), which is reportedly considering up to $10 billion. No term sheet has been signed as of Bloomberg's September 29 report.

Why the IPO keeps sliding

Sam Altman described the March 2026 raise — that was $122 billion at an $852 billion valuation — as the last round before the company went public. That was six months ago.

There are two ways to read the delay, and I think both are partially true.

The favorable read: private investors are still willing to fund OpenAI at favorable terms, and going public requires extensive disclosures — detailed financials, legal risk summaries, details competitors can read. If you can avoid that and still get the capital, why not?

The less-favorable read: a company that spent $34 billion to earn $13 billion in 2025 is hard to take public without the numbers looking much better. And a $1.4 trillion valuation — for context, Apple was trading around that number for much of 2022, but Apple was profitable — prices in a lot of assumptions about the gap closing on schedule. A stumble on revenue, or a competitor who undercuts on price, changes the math fast.

Adding competitive pressure: Anthropic filed confidentially for its own IPO in June at a reported $965 billion valuation, with a public offering expected mid-November 2026. Both companies are competing for the same pool of institutional and sovereign investors. Whoever gets there first gets the better terms.

Source spread

• Bloomberg — OpenAI Targets $30 Billion in Funding at $1.4 Trillion Value — hype. Primary report; describes the talks as early-stage, terms subject to change.
• TechCrunch — OpenAI reportedly in talks to raise $30B round at $1.4T valuation — builder. Confirms the $40B August run-rate and notes no lead investor is signed.
• Where's Your Ed At — OpenAI 2025 audited financials — skeptic. The June 2026 look at OpenAI's actual books: $13B revenue, $34B costs, $17B going to Microsoft.
• CoinCentral — Inside OpenAI's $30 Billion Funding Push — builder. Covers the MGX angle and the Anthropic IPO competitive dynamic.

What's real, what deserves a side-eye

What's real:

• The revenue growth is genuine. Going from near-zero to $40 billion in run-rate in three years is exceptional by any measure. The product has users — probably you among them.
• The structural funding gap is also real. Running models at ChatGPT's scale costs more than the subscription revenue currently recovers.
• Private capital is still available. Sovereign funds from the Gulf and Asia are deploying heavily into AI infrastructure. That doesn't last forever, but it hasn't stopped.

What deserves a side-eye:

• Altman's "last round before IPO" statement aged poorly. He said it six months ago and it was already not accurate. That's worth remembering next time there's a statement about IPO timing.
• Revenue figures conflict across reports. TechCrunch says $40B run-rate; Reuters says "nearing $70B annualized." A 75% spread between two unnamed-source estimates means at least one is inaccurate, and neither has been independently verified.
• The $1.4 trillion valuation is what OpenAI is asking for — not what investors agreed to pay. A term sheet (the legal document where both sides agree to the terms) has not been signed. Until it is, the number is a proposal, not a fact.

What to do about it

Not much requires immediate action, but three things worth keeping in mind.

• If you pay $20/month for Plus, you're fine for now. OpenAI has consistently kept paid tiers cleaner and better-featured than the free tier. That price hasn't moved since launch. But "hasn't moved yet" and "won't move" are different things.
• If you're on the free tier, expect the product to keep changing. Ads in May were step one. The free tier has always been a funnel toward paid, and that dynamic gets stronger as revenue pressure increases, not weaker.
• Cross-check decisions that matter. This was always good advice. But a company raising $30 billion has financial incentives to keep you engaged with the product, and "engaged" and "well-served" aren't always the same thing. For health, financial, or major life decisions: use it as a starting point, then verify with a human.

Further reading

• Bloomberg — OpenAI Targets $30 Billion in Funding — primary report
• TechCrunch — OpenAI in talks for $30B — confirms revenue figure and deal status
• Where's Your Ed At — OpenAI 2025 financials — the cost vs revenue picture in detail
• Fortune — Anthropic confidentially files for IPO — competitive context