The AI IPO Moment
Two companies. Two trillion dollars. One question that nobody wants to answer.
Before the end of 2026, OpenAI and Anthropic are expected to enter public markets within months of each other - together representing the largest simultaneous debut of private AI companies in history.
At their current valuations, these are not IPOs in any conventional sense. They are a public referendum on whether artificial intelligence is a genuine civilisational shift or the most expensive hype cycle in financial history.
Bubble or Revolution? The Only Question That Matters
Most IPO analysis focuses on multiples, comparables, and path to profitability. For OpenAI and Anthropic, that framework is largely beside the point. OpenAI is targeting a listing at approximately $1 trillion — 65 times trailing revenue — and does not expect to reach profitability until 2030.
Anthropic is raising at a valuation approaching $900 billion, having been valued at $380 billion just three months ago. These are not prices derived from discounted cash flows. They are bets on the shape of the next several decades.
The bubble case is coherent and deserves to be stated plainly. AI models are improving rapidly, but the value accruing from that improvement is flowing overwhelmingly to users, not to the companies producing it.
The internet provided a clear monetisation layer — advertising, which became a $200 billion annual industry. But nobody pays for a browser. Nobody pays for email.
If AI follows the same trajectory — ubiquitous, transformative, but ultimately commoditised — the correct analogy is not the iPhone. It is Google Search: one or two winners, enormous societal value, and a graveyard of companies that built the infrastructure and captured none of the returns.
The consequences of that scenario would be severe. In Q1 2026 alone, Google and Amazon combined spent $130 billion on capital expenditure — more than three times the inflation-adjusted cost of the Manhattan Project, in a single quarter. They plan to spend nearly $700 billion this year. If AI demand plateaus before that infrastructure is utilised, the write-downs would represent a capital destruction event without modern precedent. Private credit markets, heavily exposed to software company valuations, would face a correction that makes the 2001 dot-com bust look contained.
There is also a structural distortion embedded in the current earnings cycle that is worth naming. A Fortune investigation in late April found that nearly half of Google's record $62.6 billion Q1 profit — roughly $28.7 billion — came not from search, cloud, or any of its products, but from marking up its equity stake in Anthropic following Anthropic's latest funding round.
Amazon reported a similar dynamic. Big Tech is booking extraordinary profits by investing in AI companies whose valuations rise partly because Big Tech keeps investing in them. The circularity of this mechanism is not evidence of fraud.
But it should be understood for what it is.
The counter-case being that this is not a bubble receives far less airtime, possibly because its implications are more unsettling.
If AI is a genuine general-purpose technology on the order of electricity or the internet, the world in twenty years will be structurally unrecognisable from the one we inhabit today. In that scenario, OpenAI and Anthropic are not overvalued. They are the early anchors of an entirely new economic order, and current valuations will look modest in retrospect.
The AGI scenarios are a topic for a separate paper. But any honest analysis must acknowledge that the non-bubble case is internally consistent — and is not being adequately priced into mainstream conversation.
Buying these IPOs at listing is, in either case, a vote. Not on earnings — on the nature of the next fifty years. That is not inherently irrational. But it should be described accurately.
The Succession: Anthropic vs OpenAI
Assuming the revolution case holds, the more granular question is which company captures the most value from it. And here, the conventional wisdom is quietly being overturned.
OpenAI remains the brand. ChatGPT has more than 900 million weekly active users — a consumer reach no competitor is close to matching. It launched in November 2022, reached one million users in five days, and single-handedly created the public understanding of what AI could be. Its annualised revenue has crossed $25 billion, up from $6 billion at the end of 2024 — one of the fastest revenue ramps in technology history. That cultural and commercial imprint is not easily displaced.
But the lead has slipped. Anthropic's annualised revenue stood at $1 billion at the end of 2024. By April 2026 it had reached $30 billion — a 30-fold increase in sixteen months. That growth rate has no clean precedent in enterprise software. The driver is Claude: specifically Claude Code, which has emerged as the dominant tool in AI-assisted software development, and a broader product suite that has embedded Anthropic deeply into enterprise workflows in a way that creates genuine switching costs. Anthropic is not winning on consumer reach. It is winning on the customers who pay the most and stay the longest.
The valuation trajectory confirms the shift. Anthropic was valued at $380 billion in February 2026. By early May it was targeting a new round at approximately $900 billion — a 2.3x increase in under three months. On secondary markets, it has begun trading above OpenAI on an implied valuation basis. This is a remarkable inversion for a company most of the public still cannot name.
OpenAI's risks are structural as well as competitive. The company is projected to burn $14 billion in 2026 and does not expect profitability until 2030. Its CFO has warned internally that the company may not be ready for a public listing this year if compute spending continues to outpace revenue.
And the ongoing Musk litigation — a legal dispute over OpenAI's restructuring from non-profit to public benefit corporation — adds governance uncertainty that institutional investors do not welcome ahead of a listing.
The deeper question is whether OpenAI's consumer dominance is durable or illusory. ChatGPT's 900 million weekly users are largely free users. Enterprise — where pricing power, retention, and margin actually live — is where Anthropic has the stronger position.
If the AI industry consolidates around one or two providers, as enterprise software historically tends to do, the current momentum suggests Anthropic is better placed to be one of them. OpenAI still has the reach advantage. But reach without monetisation is a distribution story, not a business one.
The Quiet Winners
The most underappreciated dimension of the AI IPO moment is not about the companies going public — it is about the companies already holding them.
Microsoft owns approximately 27% of OpenAI, having invested $13 billion across three rounds since 2019. At OpenAI's current $852 billion valuation, that stake is worth roughly $228 billion — a 17x paper return on capital that Satya Nadella committed before most boardrooms had heard of ChatGPT.
Google owns approximately 14% of Anthropic — the maximum it is contractually permitted to hold — with a stake now approaching $130 billion in paper value and rising.
Amazon has committed up to $33 billion in Anthropic across a series of deals, carrying a mid-teens percentage stake currently valued at over $70 billion on its books.
Three of the five largest companies in the world have made what may prove to be the most consequential venture bets in corporate history — and all three positions are coming liquid in the same year.
The scale of the embedded gains is striking. Microsoft's $13 billion has become $228 billion on paper. Amazon's $8 billion original outlay is now carried at over $70 billion. Google invested $3 billion beginning in 2023 and is now committing up to $40 billion more — at valuations that make even that enormous follow-on look cheap relative to where Anthropic is heading.
These are not passive financial bets. Each company is simultaneously Anthropic's or OpenAI's largest cloud customer, its primary infrastructure partner, and its biggest outside shareholder. The lines between investor, customer, and competitor have dissolved entirely.
The irony is pointed. If the AI revolution plays out as its proponents believe, Microsoft, Google, and Amazon may ultimately profit more from it than the companies actually building the technology.
The pioneers rarely capture all the value they create. In this cycle, the incumbents made sure of it — years before the public markets had a chance to weigh in.
What This Means for Private Markets
For PE investors, the OpenAI and Anthropic listings matter beyond the AI sector. PE exit volumes fell 36% in Q1 2026 — not because assets are impaired, but because sponsors are waiting for a functioning public window.
The mega-IPOs could be the catalyst that reopens it. If SpaceX prices well in June and OpenAI follows in Q4, the confidence effect on the broader IPO pipeline could unlock exits across sectors that have nothing to do with AI.
The alternative is that these listings absorb so much institutional capital that they crowd out everything else for the remainder of the year — and that any pricing disappointment triggers a risk-off move that closes the window again.
The outcome depends almost entirely on the question we started with.
We do not have a confident answer on whether AI is bubble or revolution. We are not sure anyone does.
What we are confident about is this: the second half of 2026 will force the market to stop avoiding the question.
Disclaimer: This Market Note is produced by NEEM Group for informational purposes only and does not constitute investment advice or a
solicitation to buy or sell any security. The views expressed reflect NEEM's analysis at the time of writing and are subject to change. Readers
should conduct their own due diligence before making any investment decisions.
© 2026 NEEM. All rights reserved. | info@neemgroup.com | neemgroup.com