Artificial Intelligence – Bubble Scenario

AI increasingly looks reset-ready

  • The news that OpenAI and Anthropic are continuing to raise huge sums at increasingly eye-watering valuations means that when the IPOs come, the probability of an IPO failure triggering a market reset is rising rapidly.
  • OpenAI is raising a quick $50bn at $750bn while Anthropic is thought to be raising $20bn+ at $350bn, even though it only had a revenue run rate of $9bn at the end of 2025.
  • By any stretch of the imagination, these numbers are outlandish, and it is only because almost all of these companies remain private that a proper forensic examination of the business case has not taken place.
  • In the world of private investments, it is only after you have stated your interest and signed an NDA that you are permitted to have a look at the forecasts.
  • If you have the temerity to question the forecasts, then you will be sent to the back of the queue, and the next person will take your allotment of highly-priced paper.
  • This means that “Luddites” like me, who like profits and cash flow, don’t get to see the numbers and our views are deeply out of favour with the AI glitterati.
  • This works until it doesn’t.
  • The problem at the moment is that there is way too much money chasing too little paper, and it is not until that balance is restored that OpenAI, Anthropic, et at will be forced to reckon with reality.
  • It is when the money stops flowing that the reset will occur, and I have been looking for signs of this for 2 years.
  • In the last couple of months, the first cracks have started to appear, but like any bankruptcy or crisis, it happens very slowly and then all at once, by which time it is too late to get out.
  • The biggest red flag for me remains the bond market, where the cost to insure oneself against Oracle going bankrupt has increased 2.8x since it signed its $300bn compute deal with OpenAI.
  • Furthermore, the same insurance to protect oneself against a CoreWeave bankruptcy is also rising and already stands at 0.5%, which is roughly 3x Oracle and 10x a relatively normal company like IBM (0.05%).
  • This is the bond market telling investors that it in uncomfortable with the business model of compute and does not understand what makes OpenAI, Anthropic et al going concerns.
  • 2026 is expected to see 3 monster IPOs, each of which is expected to raise $50bn+ from public market investors, which are OpenAI, Anthropic and SpaceX.
  • I suspect that the SpaceX IPO won’t go ahead and will instead become part of a restructuring of Elon Musk’s companies into a structure that is like Alphabet, leaving the two big LLM makers.
  • These are the same sorts of investors who will be perusing the F1 statements and wondering how they will make money from investing in OpenAI at a valuation that is now likely to be $1tn or more.
  • I think that the short answer is that they won’t simply because, at a valuation of $1tn, OpenAI has to win the race of the consumer AI ecosystem (see here) and win it convincingly.
  • In practice, this means it has to find a way to earn in excess of $20 per user per month just to stay ahead of the depreciation charge of $280bn per year on the $1.4tn in investments that are planned.
  • This figure needs to be much higher if the company wants to make profits or generate cash, which every equity valuation that is higher than $0 demands.
  • This, on its own, is a big challenge, and OpenAI has the added problem of Google breathing down its neck, which has both a technical advantage in cost and almost unlimited resources.
  • Hence, I think that there is a rising probability that the public market rejects the valuation it is being asked to pay for OpenAI, which would cause a loss in confidence in the business model for AI.
  • As all of these companies are now connected to one another, this could easily set off a domino effect where the valuations of all concerned take a big hit, and some go out of business or are acquired.
  • This is exactly how resets have happened in the past, and I am becoming increasingly concerned that in 2026, past will be prologue.

RICHARD WINDSOR

Richard is founder, owner of research company, Radio Free Mobile. He has 16 years of experience working in sell side equity research. During his 11 year tenure at Nomura Securities, he focused on the equity coverage of the Global Technology sector.