They filed the papers, then hit the brakes
OpenAI is not for sale. Not this year. Sam Altman, the company’s chief executive, says a stock market debut in 2026 would be ‘ill advised’. He is worried about safety. He is worried about creating an artificial intelligence beyond human control, telling Fortune magazine there are certain risks his company should not be allowed to incur on humanity’s behalf. He even promised to pause development if things got too dangerous. Strong words. The responsible choice.
It is a message of profound caution from the leader of the world’s most watched technology firm. In an interview published on Saturday, Altman insisted there was ‘a lot of stuff to do’ before he could contemplate exposing OpenAI to the pressures of public ownership. He feels no pressure to float. He is not rushing. This all sounds very sensible.
There is just one problem. A very big one. Before this public declaration of restraint, before the philosophical warnings about existential risk, OpenAI confidentially filed the legal paperwork for an initial public offering. They filed the papers. This is not the action of a company merely thinking about a flotation one day, but the formal first step in a complex, costly, and highly regulated process designed to bring a company to the public markets within months, not years. It is a move that involves lawyers, bankers, and a clear intention to sell shares.
A strange reversal. This private action sits in direct opposition to the chief executive’s public pronouncements. A confidential filing is the starting gun for an IPO. It signals that the internal machinery was already in motion and that the company was serious enough about a 2026 listing to begin the formal proceedings with American regulators. Something, or someone, slammed the brakes on a process that was already well under way. The story about safety is the public explanation for a very private and very sudden change of plan. The question is what really happened.
A company that does not need the cash
Most technology startups follow a brutal, predictable path. They burn through cash. They chase growth at all costs. They race towards a stock market flotation because it is the only way to raise the vast sums needed for global expansion and to give their early, risk taking investors a payday. An IPO is the destination. The entire journey is mapped towards it.
OpenAI is not on this journey. It has no need for public money. It has Microsoft.
The Redmond technology giant has ploughed billions of pounds into OpenAI, a staggering investment that grants it privileged access to the AI models while providing Sam Altman with a financial cushion that other founders can only dream about. That money changes everything. It removes the desperate, existential pressure to please public markets. It allows OpenAI to fund its huge computing costs, which run into tens of millions of pounds every single day, without having to sell a single share to a pension fund in the City of London.
Altman can afford to talk about pausing progress. He can philosophise about alignment and existential risk. A typical chief executive, beholden to quarterly earnings reports and the whims of institutional investors, would be fired for such talk. They would be accused of destroying shareholder value. Altman, backed by Microsoft’s deep pockets, has the freedom to prioritise his version of safety over short term profit. He can afford the delay. This is not a company starving for cash. It is a company that is already funded by one of the richest corporations on the planet, a situation that makes the secret IPO filing even more confusing.
What 'safety' means to a City analyst
City investors are simple creatures. They want predictable returns. They build financial models to map the future, translating a company’s strategy into neat columns of pounds and pence that stretch out for years. This is not art. It is arithmetic. An analyst working in a glass tower in Canary Wharf needs to know that revenues will climb, that costs are controlled, and that the chief executive’s primary job is to make the numbers on a spreadsheet come true. A public company promises this predictability in a legally binding prospectus, a solemn pledge of reliable, uninterrupted growth.
Sam Altman does not speak this language. He is not talking about profit margins. He is talking about humanity. In his recent interview with Fortune, he spoke of 'recursive self improvement' and the genuine possibility of creating an AI that is 'beyond human control'. These are subjects for a university philosophy department, not for the pre flotation roadshow presentation to a room full of pension fund managers in Edinburgh. For an analyst paid to calculate risk, these are not inspiring visions of the future. They are red flags. Huge red flags.
The true terror for any potential investor lies in a single, devastating promise Altman made. He said he would take action to prevent an uncontrollable AI, 'even if it meant pausing training'. This is not a metaphor. It is a direct threat to the entire commercial logic of a publicly listed company. Pausing training means telling the research labs to shut down the experiments that create the next generation of products, the equivalent of Ford stopping its production lines indefinitely or GSK halting all drug trials because of a moral quandary.
The consequences would be catastrophic. They are also unquantifiable. A chief executive who reserves the right to halt all commercial progress on ethical grounds, with no notice and for an indeterminate period, makes financial forecasting impossible. How can anyone price shares in a company that might voluntarily stop doing business tomorrow? The entire edifice of shareholder value, which is built on the assumption of continuous, relentless commercial activity, simply collapses. Competitors would not pause. They would race ahead. Revenue would stop. The share price would plummet. Sam Altman’s public statements on safety are, from a City perspective, a promise to investors that he is willing and able to destroy their capital at any moment for reasons they cannot model or control. No public market would ever accept that deal.
The capped-profit problem
Even if the safety issues vanished tomorrow, a deeper, structural problem makes a public listing almost impossible. The problem is the company itself. OpenAI is not a normal business. It is a legal and philosophical maze built from two warring parts, a non profit parent with a humanitarian mission controlling a for profit subsidiary that exists to make money. This hybrid model was designed to hoover up billions in capital from backers like Microsoft while supposedly protecting the company from the corrupting influence of pure commerce. It is this very structure that makes a standard flotation on the London or New York stock exchanges a legal nightmare.
The first issue is the profit. It is capped. Early investors signed up to a deal where their returns are limited to a multiple of their original stake, with all excess wealth generated by the artificial intelligence flowing back to the non profit parent for the good of humanity. This is a revolutionary idea. It is also utterly alien to the logic of public markets, which are built on the simple, powerful premise of unlimited upside for shareholders. Nobody buys shares in a company that has a built in ceiling on their success. The entire mechanism of a stock market, from analyst valuations to pension fund mandates, is designed to chase infinite growth, not to fund a charity once a certain threshold is reached.
The second issue is control. This is far worse. The ultimate authority at OpenAI does not rest with its chief executive or its shareholders. It rests with the board of the non profit parent. This is not theoretical. We saw it happen. In November 2023, that board fired Sam Altman. The decision, made to protect the company's mission, plunged the world's most important technology firm into chaos and was only reversed after immense pressure from employees and investors. The episode provided a brutal, public demonstration of where the real power lies. The board’s legal duty is to its founding mission, not to commercial partners or any future public shareholders.
An initial public offering would force an immediate and unwinnable conflict between these two duties. A publicly listed company has a fiduciary duty to its shareholders, an obligation enforced by regulators like the Financial Conduct Authority. OpenAI’s charter creates a supreme, overriding duty to its abstract mission. It is impossible to serve both. Any prospectus for a flotation would have to explain to potential investors that their capital could be rendered worthless at any moment by a board they do not elect and cannot control, acting on principles that have nothing to do with financial return. It is an unsellable proposition. The company cannot go public without destroying the very legal structure that defines it.
This is a strategic retreat, not a panic
This is not a panic. It is a calculation. A flotation right now would be a profound strategic error, forcing OpenAI to fight a war on two fronts. One front is the City. The other is government. It cannot win both.
Politicians are moving. From Washington to Brussels and London, lawmakers are beginning to draft the rules that will govern artificial intelligence. The dire warnings are real. These regulations will define everything from data privacy to national security protocols, and they will arrive with the full force of the state behind them. A flotation now would subject OpenAI to the unforgiving scrutiny of public market regulators at precisely the moment when politicians and civil servants are drafting this entirely new and unpredictable set of laws designed to constrain the very technology the company sells. It would be a catastrophic mistake.
By delaying the initial public offering, Sam Altman buys himself time and influence. He remains a private actor. He is not a public company chief executive, shackled by the short term demands of quarterly reporting and shareholder calls. This freedom is critical. It allows him to engage directly with governments not as a supplicant asking for permission, but as a partner offering solutions. He can travel to Downing Street. He can walk the halls of the Capitol. He can meet with European Commissioners. He can do all of this without the constant pressure of a share price ticking away in the background, judging his every move.
The goal is simple. Shape the rules. It is far better to help define what 'safe AI' means in legislation than to have a hostile definition imposed by politicians who may not fully grasp the complexities of the technology. By positioning OpenAI as the responsible adult in the room, the company can influence the legal framework to its own advantage, creating a regulatory moat that competitors will find difficult to cross. This is not about retreating from the public square. It is about choosing the right battlefield. The real prize is not a blockbuster stock market debut in 2026. The real prize is writing the rulebook for the next fifty years. That is the game. The IPO can wait.
Who wins from the wait?
So who wins? The answer is clear. Sam Altman wins. He keeps his grip on the company he nearly lost, cementing his authority without the meddlesome interference of public shareholders demanding predictable quarterly earnings. He gets to play the long game. Microsoft wins too. The technology giant has already committed billions of pounds to OpenAI, and this delay ensures its prize asset remains a close, semi private partner rather than a freewheeling public competitor beholden to the wider market. Redmond has no desire for a quick return. It wants dominance. This delay delivers it.
The losers are just as obvious. They are the early investors. Think of the venture capital funds and the long serving employees who own shares that are, for now, just numbers on a page. Their paper fortunes are immense, but they remain unrealised. An initial public offering is the traditional exit. It is the moment when years of risk are converted into hard cash. That moment has now been postponed indefinitely. They are trapped. They must wait for Altman to decide when they can finally get paid, and every new private funding round potentially dilutes the value of their stake. Their patience is not infinite.
So what happens next? Look for the signals. Watch the corporate structure. The biggest single barrier to a stock market listing is OpenAI’s bizarre capped profit model, which subordinates the commercial entity to a non profit board. Any move to simplify or dismantle this arrangement is the strongest possible sign that a float is being prepared in earnest. Follow the money. Another giant private funding round would suggest the delay is for the long haul, kicking the can far down the road. Also, keep an eye on Brussels. The final text of the European Union's AI Act will create the legal reality that all technology companies must inhabit, and its terms will directly influence when, and how, OpenAI finally decides to face the public markets. The game is about control. The IPO is just one move. The waiting continues.
Sources. Guardian Business: OpenAI IPO will not happen in 2026 amid AI safety fears, Sam Altman says. The Verge: Sam Altman says OpenAI going public in 2026 would be ‘ill-advised’. TechCrunch: OpenAI’s Sam Altman says it would be ‘ill-advised’ to go public in 2026. City AM: OpenAI delays IPO as Altman says safety comes first.
Analysis. Drafted with AI assistance from the sources listed above and reviewed by an editor before publication. Jnews links to the organisations it writes about.
