Pivot to AI
Pivot to AI

20260508 - OpenAI may not be able to IPO in 2026

6 min
•May 8, 202622 days ago
Listen to Episode
Summary

OpenAI's IPO plans face significant delays due to deteriorating financials, missed revenue targets, and internal disagreements between CEO Sam Altman and CFO Sarah Fryer over compute spending. The episode examines how OpenAI and SpaceX are burning venture capital on inflated valuations while struggling with actual business fundamentals, raising questions about the sustainability of the AI bubble.

Insights
  • OpenAI's IPO timeline has shifted from 2026 to mid-late 2027 at earliest, driven by missed internal targets and revenue shortfalls against competitors like Anthropic
  • Public alignment statements between CEO and CFO on capital allocation are red flags for pre-IPO companies, suggesting underlying strategic disagreements
  • The AI funding model relies on venture capital chasing valuation growth rather than profitability, creating unsustainable burn rates that require eventual public market exits
  • Index rule changes being made ad-hoc to accommodate SpaceX and OpenAI IPOs represent systemic risk, with retail investors bearing the downside of speculative bets
  • SpaceX's leaked S-1 filing shows disconnect between Elon Musk's public claims and actual business fundamentals, yet investor enthusiasm remains driven by brand rather than metrics
Trends
AI startup IPO delays extending into 2027-2028 as profitability and unit economics deteriorateIncreasing tension between venture capital valuations and public market readiness in AI sectorIndex operators modifying listing criteria to accommodate high-profile but financially weak AI companiesCompute spending arms race among AI leaders outpacing revenue growth and creating cash flow crisesVenture capital concentration risk as OpenAI absorbs disproportionate share of US VC fundingRegulatory and index governance changes being driven by specific company needs rather than market healthCompetitive pressure from Anthropic forcing OpenAI to increase compute spending despite financial constraintsCEO-CFO misalignment becoming public, signaling deeper governance and strategy disagreements
Companies
OpenAI
Primary subject; facing IPO delays to 2027+ due to missed revenue targets, compute spending disputes, and deteriorati...
Anthropic
Competitor gaining ground on OpenAI, driving pressure for increased compute spending and capital allocation
SpaceX
Parallel IPO candidate with leaked S-1 filing showing weak fundamentals but strong investor enthusiasm based on Elon ...
PitchBook
Venture capital news site cited for analyst Harrison Rolfe's assessment of optimal OpenAI IPO timing
People
David Gerard
Podcast host and primary commentator analyzing OpenAI and SpaceX IPO prospects and financial challenges
Sarah Fryer
CFO expressing concerns about revenue growth, compute spending sustainability, and IPO readiness; publicly disagreein...
Sam Altman
CEO pushing for maximum compute spending ($1.4T data center plan) despite CFO concerns about financial sustainability
Harrison Rolfe
Venture capital analyst predicting OpenAI IPO should occur mid-late 2027 after quarters of clean execution
Elon Musk
SpaceX leader whose public claims about orbital data centers are walked back in leaked S-1 filing
Amy Caster
Co-wrote this episode after working on additional AI story research
Quotes
"She is worried the company might not be able to pay for future computing contracts if revenue doesn't grow fast enough"
David Gerard (reporting on Sarah Fryer's concerns)•Early in episode
"We are totally aligned on buying as much compute as we can and working hard on it together every day"
Sam Altman and Sarah Fryer (joint statement)•Mid-episode
"Strategic alignment on capital allocation is not something that healthy pre-IPO companies need to publicly assert"
Harrison Rolfe (PitchBook analyst)•Mid-episode
"Sam has one trick be bigger than everyone. But it turns out this costs actual money"
David Gerard•Early-mid episode
"You and I know it's fairy gold, but the story of the AI bubble is setting real dollars on fire to put imaginary valuation dollars on the books"
David Gerard•Mid-episode
Full Transcript
Hi, I'm David Gerard and this is Pivot to AI, coming to you Monday to Friday. Today, the IPO for the company with the worst financials on earth might not be a hapner. The trouble with doing a stock market initial public offering is you need to bear yourself to the world and show everything about your horrible business numbers. All of OpenAI's numbers have always been hilariously terrible. And just a month ago, Sarah Fryer, the chief financial officer, thought the numbers were bad enough, the IPO should be delayed till next year. Just this week, Fryer's been saying this a bit louder. OpenAI missed a pile of its internal numbers for users and revenue, and it's losing ground to Anthropic. Quote. She is worried the company might not be able to pay for future computing contracts if revenue doesn't grow fast enough, according to people familiar with the matter. Board directors have also more closely examined the company's data centre deals in recent months and questioned Chief Executive Sam Altman's efforts to secure even more computing power despite the business slowdown. Unquote. Sam has one trick be bigger than everyone But it turns out this costs actual money Altman and Fryer issued a statement Quote, We are totally aligned on buying as much compute as we can and working hard on it together every day. Unquote. Yeah, but one of you's worrying about how to pay for it. Altman told the world he was going to spend $1.4 trillion on data centers, and Fryer had to tell the investors the plan was $600 billion, which OpenAI isn't going to make either. Thing is, the analysts agree. Like Harrison Rolfe's from PitchBook, the news site for venture capital, a site you should all be reading. Rolfe's thinks a good time for OpenAI to do an IPO is, Mid to late 2027. It was a few quarters of clean execution and not floundering. Good luck with that one. Rolfs is particularly worried about Friar and Altman's public statements that they're totally getting on fine. Quote. Strategic alignment on capital allocation is not something that healthy pre-IPO companies need to publicly assert. When the CEO and CFO feel compelled to call the report ridiculous, the report is usually close to accurate." The problem is OpenAI doesn have the time It doesn have till late 2027 It be dead It won just run out of the money it has now It'll run out of all the future money. OpenAI is sucking up all the venture capital in the US right now, because it's the valuation number that's going up. And that's the one thing investors are desperate for. You and I know it's fairy gold, but the story of the AI bubble is setting real dollars on fire to put imaginary valuation dollars on the books. OpenAI has to do an IPO so they can dump on the public markets when they finally run out of other people's money. Right now, they have an imaginary valuation of $852 billion, apparently one to IPO for $1 trillion. This is nuts. Urban AI's other problem is SpaceX, which is now also ex-AI. SpaceX is doing their IPO real soon, maybe next month. Whoever goes first sets the tone for AI startup IPOs, and that's going to be SpaceX. SpaceX's draft S1 filing leaked, because of course it did. It walks back all the big claims Elon Musk's made, like orbital data centers. The company's numbers are trash. That doesn't matter, because anyone putting money into SpaceX is betting on Elon magic not tawdry details like a coherent business plan And Elon cults are precisely the audience for AI Who going to buy this stuff Well, you are. Your pension fund, the S&P 500 and the Nasdaq Index are both changing their rules ad hoc specifically to let SpaceX and OpenAI join the index and all the funds by the indexes. When times look trepidatious, change the rules to make numbers go faster. Nothing can go wrong. And you are the sucker of last resort. Thanks for tuning in to Pivot to AI. And thank you to Amy caster who co-wrote this episode after we spent all afternoon working on another AI story that isn't cooked yet. We'll get there. Don't forget, send this episode to just one other person. Spread the word. Hit like and subscribe on YouTube. Leave a nice review in your podcast app. And if you want Pivot2AI to keep coming out daily, you should send $5 to the Patreon linked in the show notes. It really helps. Thank you all. I'll see you on Monday and bye for now.