# OpenAI Delays Its IPO as the AI Bubble Debate Splits the Tape - The VC Read - Week of June 26 to July 3, 2026

> The VC Read for the week of June 26 to July 3, 2026. OpenAI reportedly delays its IPO rather than price below $1 trillion and SpaceX trades soft after a record debut, igniting the widest AI-bubble bull-bear split on the tape, from Grantham and Zitron to Bloom Energy's KR Sridhar and IPO scholar Jay Ritter.

## The VC Read

### Week of June 26 to July 3, 2026: OpenAI Delays Its IPO as the AI Bubble Debate Splits the Tape

---

*The week the AI-bubble argument stopped being a dinner-party abstraction: a delayed mega-IPO, a $2 trillion rocket wobbling out of the gate, and the widest bull-bear split anyone on the mics can remember.*

Quick note on the tape: this was not a thin week. Two hard catalysts, OpenAI reportedly pushing its IPO to next year, and SpaceX trading soft after its record debut, dragged the whole VC/macro conversation onto the same battlefield. The structural venture debates (fund size, DPI, capital concentration) were there but got drowned out by the noise.

## The Big Debate: Is This the Top, and Was OpenAI Blinking the Tell?

The news that set it off: the New York Times reported OpenAI is leaning toward delaying its IPO rather than accept anything under the $1 trillion valuation Sam Altman is holding out for (it was $850B in March). On Pivot, Scott Galloway argued the delay is really a confession. He called it "the great flippening," Anthropic overtaking OpenAI faster than "Avis overtake Hertz," and said the bankers, not the market, forced the pause: an S-1 "side by side, would show a company that is losing momentum and is spending way too much money and losing too much money," with leaked losses up "nearly 8x in 2025 and spending hitting $34 billion," against an Anthropic "projecting that it'll break even by 2030." (Pivot, "Comcast Splits, OpenAI Weighs IPO Delay, and Buttigieg Targeted," June 30, 2026.)

The bears smelled blood. Jeremy Grantham called it the most expensive market in American history and put a number on the unwind: "getting back to trend from here is closer to a 70% decline than a 50%." On SpaceX at ~$2 trillion, 90% of the story riding on AI extrapolation, he was withering: history will see it as "one of the defining peaks of all time." His frame: "They lost their shirts in the railroad... They lost their shirts in the Internet... And they will lose their shirts in AI." (Squawk Pod, "Jeremy Grantham: The Most Expensive Market in American History 6/26/26," June 26, 2026.) Ed Zitron supplied the plumbing for that view: Meta selling "excess compute" is the giveaway, because "80% of all AI compute is owned by or used by OpenAI or Anthropic," "60% of organizations are token minimizing," and "over 100 gigawatts of compute capacity under construction" chases maybe "six gigawatts" of real demand. His kicker: "unlike the dot-com bubble, there is no useful infrastructure here." (Prof G Markets, "The AI Trade Just Got A Warning From Meta," July 2, 2026.)

The steelman on the other side isn't a hand-wave. KR Sridhar of Bloom Energy told 20VC to separate the tape from the trend: AI is "almost a hockey stick on a hockey stick," and "for the first time in human history, we are manufacturing intelligence. When was the last time any person anywhere in any civilization said, we have too much intelligence, let's stop? Never." His advice: "don't short the US, don't short Silicon Valley." (20VC, "Why We Are Not in an AI Capex Bubble... with KR Sridhar, Bloom Energy," June 29, 2026.) And the sober middle came from the man they call Mr. IPO, Jay Ritter: yes, SpaceX priced at a price-to-sales ratio "over 90" and such names usually underperform, but as a timing signal, giant IPOs mark tops with only "about 51% accuracy," and remember Shiller's "irrational exuberance" landed "more than 3 years before the US market peaked at a much higher level." (Monetary Matters, "Top IPO Scholar on Unprecedented IPO Wave & Why IPOs Underperform the Market | Jay Ritter," June 30, 2026.)

The honest read: everyone agrees AI is real and the infrastructure spend is unprecedented. The fight is entirely about price and timing, and this week, for the first time, a marquee issuer flinched first.

## Signals

1) **The IPO window is a keyhole, not a door.** Ritter: 2026 will host "the largest IPOs in the history of the world" (SpaceX roughly twice the prior record), yet the count of companies going public is "nowhere near the 311 operating companies that went public in 2021." Software and biotech "dried up," partly on "a widespread view that AI is threatening the business models" of SaaS. The counterweight: Nasdaq Vice Chairman Bob McHughie says 2026 is poised to be the largest capital-raise year in history. (Monetary Matters, June 30, 2026; Brew Markets, "Fighting for the Biggest IPOs (ft. Nasdaq Vice Chairman)," June 29, 2026.)

2) **Exits are quietly thawing, take the call.** On the This Week in Startups VC roundtable, Aileen Lee's advice for founders got emphatic: if an outsider "make[s] a real serious approach, you should take that very, very, very, very seriously," because "your first offer is probably your best offer." The panel flagged SmarterDX's ~$1B exit (a seed win) and Intercom going "thin and then sell for a pretty quick number"; meanwhile Italian roll-up Bending Spoons (Vimeo, Evernote, Eventbrite) is filing to go public near an $18B valuation. M&A is coming back, just not at 2021 marks. (This Week in Startups, "Why the VC Hype Cycle Always Gets It Wrong | VC Roundtable | E2307," July 1, 2026.)

3) **"SaaS is dead" is overstated, but the pricing model isn't safe.** Box CEO Aaron Levie made the incumbent's case: agents need "deterministic software... the right permissions, the right workflow design," so "we actually see an increase in usage because agents are now roaming around accessing all of this data," and "a plumber in Missouri is [not] going to make their own CRM." The real question is monetization, which he expects shifts toward consumption. That squares with Ritter's read that the market fears exactly this, hence no SaaS IPOs. (My First Million, "I dropped out of college and built a $3.6B company from scratch," July 2, 2026.)

4) **Mega-seeds up top, orphans in the middle.** The roundtable ran the tape on $100M+ early rounds, StarCloud ($170M Series A), General Intuition ($320M), Scale of Cognition and Scout AI ($100M each). But Aileen Lee warned the mirror image: later-stage investors "are not stepping up," leaving "orphan companies" with boards where people are "asleep at the wheel," a byproduct of turnover at the giant funds. Capital is concentrating at both the hot top and the barbell's ends, the messy middle is where value leaks. (This Week in Startups, July 1, 2026.)

5) **The DPI reckoning nobody wants to say out loud.** Hamilton Lane's Miguel Luenya (co-head of Global VC, $1T+ AUM) noted venture has "historically returned the least capital in terms of DPI" of any private asset class, and that continuation vehicles only appear when GPs face LP pressure raising the next fund. It's why LPs on the TWiST panel said they need "the SpaceX's and Stripes and OpenAI's and Anthropics to go public" to feed real money back, small funds returning $10-15M don't move a large institution. (How I Invest with David Weisburd, "E398: Hamilton Lane ($1T AUM) on Venture Capital, AI, and Private Markets," July 2, 2026.)

## Quote of the Week

> "They lost their shirts in the railroad. Brilliant idea. They lost their shirts in the Internet... And they will lose their shirts in AI."

Jeremy Grantham, on why a world-changing technology and a portfolio-ending bubble are not mutually exclusive. (Squawk Pod, "Jeremy Grantham: The Most Expensive Market in American History 6/26/26," June 26, 2026.)