Are AI giants rushing to go public essentially a “getaway game” for capital, leaving ordinary people to pick up the tab?

```

Anthropic recently submitted a secret IPO application; OpenAI quickly followed suit, with both AI giants racing to enter the public market. However, behind the impressive revenue figures, a more concerning issue is emerging: is this IPO frenzy an extension of technological dividends to the public, or a carefully planned exit by early investors transferring risk to ordinary people?

According to a recent analysis by The Algorithmic Bridge columnist Alberto Romero, Anthropic's annualized revenue run rate is close to $50 billion, and it may achieve its first profitable quarter; OpenAI’s revenue is slightly lower and has no timeline for profitability—both companies are in an ongoing cash burn phase. At their current scale, continued losses are no longer a sustainable business model.

Michael Hartnett, Bank of America's Chief Strategist, makes a sharp observation: this IPO wave is less an investment opportunity for the public market, and more an exit channel for early investors to transfer years of accumulated risk onto the public market. Bridgewater founder Ray Dalio once said, "Popping the bubble is the process of converting wealth to cash"—for insiders, regardless of whether a bubble exists, they are winners.

Meanwhile, the commercial reliability of AI technology is facing ever more real scrutiny. Microsoft this month announced the withdrawal of Claude Code authorization, switching to internal tools due to unsustainable costs; Uber set a hard cap of $1,500 per month on employee spending for AI programming tools; Starbucks quit North American AI tool deployment just nine months in, citing reliability issues. Even OpenAI CEO Sam Altman admits AI costs have become “a huge problem.”

IPO window narrows suddenly, giants scramble to get ahead

Anthropic recently announced in a blog post that it has submitted a confidential S-1 draft to the SEC; once approved, the company will have the option to formally launch its IPO. OpenAI's listing process is also accelerating. Alberto Romero points out in his analysis that if Musk’s SpaceX acquisition of xAI is counted in, the market is actually facing an intensive “triple IPO” shock.

This timing is intriguing. Both companies are currently unprofitable, yet their valuations are sky-high, with massive capital expenditures, unclear returns, and enormous long-term spending commitments. Alberto Romero argues that the essence of this race is not a technical contest between the two companies, but a capital game against time—the IPO window is narrowing not because AGI (Artificial General Intelligence) is about to arrive, but because the narrative that ‘AGI is about to arrive’ is itself rapidly losing its shelf life.

Risk transfer logic: privatized gains, socialized losses

Alberto Romero cites Michael Hartnett’s view that the core mechanism of these IPOs is systematic risk transfer. Early investors use IPOs to sell off high-risk exposures built up over years to pension funds, index funds, and the broader public.

Using Ray Dalio’s framework: whether there’s a bubble or not, insiders fare about the same—if there’s no bubble, they keep operating and get rich eventually; if there is a bubble, they exit early and cash out immediately. Ordinary investors in the secondary market bear the tail risk.

Alberto Romero summarizes this logic as ‘privatized gains, socialized losses’. He notes that if the market recognizes the bubble in current valuations before IPO, Anthropic and OpenAI’s opportunity to cash out at the top will be greatly reduced; but if the correction comes only after IPO, losses will spread throughout the economy and be shared by ordinary people.

Enterprise trust shaken, AI commercialization faces real-world interrogation

Underlying these concerns are real signals from the enterprise side. Alberto Romero lists several recent cases: Microsoft canceled Claude Code authorization, switched to internal tools due to unsustainable costs; Uber capped monthly employee AI programming tool spending at $1,500; Starbucks shut down AI tool deployment in North America after only nine months, citing unreliable performance.

These cases point to a core issue: no clear positive correlation currently exists between AI expenditure and AI returns. Alberto Romero believes similar stories will keep emerging, as this predicament is widespread—not a unique challenge for any single firm. He particularly notes that the hallucination problem in AI models remains unsolved, which makes him distinctly skeptical that AI can withstand the test of time like the internet or electricity did.

Whose future is deeply tied to AI?

Alberto Romero presents a somewhat ironic paradox: founders and early shareholders of AI giants, whether AI ultimately succeeds or not, have already locked in their gains—if successful, they reap both fame and fortune as technology revolutionaries; if not, they’ve cashed out through the IPO. In contrast, it is ordinary people whose fate becomes deeply bound to AI’s prospects.

He also acknowledges that if AI technology ultimately delivers on its promise, IPOs will extend what was private gains to public wealth, which is how capital markets foster broader prosperity. But he remains pessimistic, because: with reliability in doubt and hallucination issues unsolved, whether this technology can truly stand the test of history is still an open question.

Alberto Romero adds that he does not see this as a premeditated conspiracy. In his view, Silicon Valley’s information bubble is closed enough that those founders who truly believe ‘superintelligence by 2027’ may be sincere. But he emphasizes that subjective intent does not change objective results—regardless of motivation, the transfer of risk is happening for real.

Risk Disclosure and DisclaimerThe market has risks; invest cautiously. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial circumstances, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are applicable to their situation. Investments based on this are at your own risk. ```