Sam Altman says OpenAI going public in 2026 would be ‘ill-advised’
Developed over time first seen 2 months ago
Sam Altman has ruled out an OpenAI initial public offering in 2026, telling Fortune it would be "ill-advised" given the current focus on safety issues. Speaking in a 45-minute interview, the OpenAI chief executive said the company feels no pressure to float and will only do so "when we're ready", a notably cautious stance from a firm otherwise known for rapid growth and expansion.
Altman also touched on wider safety concerns during the conversation, including the recent Hugging Face hacking incident and the risk that recursive self-improvement could produce AI systems beyond human control. He said this outcome was "absolutely" possible but pledged to intervene if necessary, including pausing training, stating that "there are risks we should not be able to incur on behalf of humanity."
- Sources
- 2
- Outlets
- 2
- Left · Centre · Right
- 1 · 0 · 0
In short
- Altman rules out an OpenAI IPO happening in 2026
- Cites safety concerns as reason to delay going public
- Says AI beyond human control is "absolutely" possible
How this story developed
- 2 months ago Sam Altman has ruled out an OpenAI initial public offering in 2026, telling Fortune it would be "ill-advised" given the current focus on safety issues.
- 2 months ago OpenAI chief executive Sam Altman has said the company will not pursue an initial public offering in 2026, describing such a move as "ill-advised" given ongoing safety concerns.
Full account
Sam Altman has ruled out an initial public offering for OpenAI in 2026, telling Fortune magazine that current concerns around artificial intelligence safety make this the wrong moment for the company to list on public markets. Speaking in a wide-ranging interview, the OpenAI chief executive said the firm was not under any financial pressure to float and would only do so once it judged the timing right.
Altman told Fortune that OpenAI was 'not rushing into an IPO', arguing that, given everything happening around safety, going public right now would be an ill-advised step. Pressed on whether a listing might instead come in 2027, he said only that it would not be 2026, pointing to the amount of work still needed on safety, alignment and coordination between industry and governments. The same conversation touched on other sensitive topics, including a hacking incident affecting the AI platform Hugging Face and the prospect of 'recursive self-improvement' in AI systems. Asked whether an AI could ultimately slip beyond human control, Altman said this was 'absolutely' possible, but insisted OpenAI would act to prevent it, including halting training runs if necessary, saying there were risks humanity should not be exposed to.
The remarks land against a backdrop of mounting political and industry unease about the pace of AI development. Lawmakers from both major US parties have been pushing for tighter regulation after stark warnings from researchers at rival lab Anthropic that unchecked progress could eventually threaten humanity's survival, alongside reports of AI agents behaving unpredictably and of safety staff resigning from AI companies over risk concerns. Anthropic's chief executive, Dario Amodei, published an essay calling on the industry to deliberately slow the rate at which model capabilities are improved; Altman responded publicly that he agreed with the need to 'pace the frontier', saying the issue had been under active discussion inside OpenAI in recent weeks, and suggesting that a broader industry pact on safety coordination could be close.
OpenAI's caution is not entirely new: reports emerged earlier this year that the company was already weighing whether to delay a potentially trillion-dollar flotation until 2027, partly in light of a volatile reception to SpaceX's own listing. Notably, safety concerns of this kind do not appear to be slowing every AI firm's market ambitions — Anthropic itself is reportedly pressing ahead with preparations for its own IPO, with marketing said to be planned for as early as mid-October and a listing expected before the US midterm elections in November, underlining a divergence in strategy even among labs voicing similar safety concerns.
Where outlets differ
From the full account, which is AI-written from the sources:
One account centres on the breadth of Altman's Fortune interview, treating the IPO comments as one element alongside his remarks on the Hugging Face hack, recursive self-improvement and the risk of AI escaping human control; the other treats the IPO delay itself as the central story.
The second account embeds the IPO decision in wider political context — lawmakers' regulatory push, Anthropic researchers' warnings of existential risk, AI agents 'going rogue', and safety staff departures — none of which feature in the first.
Only the second account references Dario Amodei's essay urging the industry to slow capability gains and Altman's public reply agreeing to 'pace the frontier', plus the reported prospect of a broader industry safety pact.
The second account also adds background not in the first: a earlier report that OpenAI was considering delaying its IPO to 2027 amid SpaceX's volatile listing, and reporting that Anthropic is nonetheless pressing ahead with its own IPO timeline.
The first account includes additional direct quotation nuance, such as Altman saying OpenAI is 'not rushing' and that training could be paused if risks became unacceptable, which is absent from the second.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Supporters of Altman's caution argue that a company building frontier AI systems has a responsibility to prioritise safety and organisational readiness over the pressures of quarterly earnings and shareholder demands. They see public markets as poorly suited to long-term, high-stakes research, where premature listing could create incentives to rush deployment or downplay risks such as loss of control over increasingly capable systems. From this view, refusing to be pushed into an IPO by market enthusiasm, and being willing to pause training if serious risks emerge, reflects a genuinely responsible stance that values humanity's interests above growth or valuation.
The case against
Sceptics counter that avoiding public markets also avoids the disclosure, governance and accountability that come with being a listed company, allowing OpenAI's leadership to make consequential decisions about powerful technology with less external scrutiny. They may note that framing delay as a safety measure is convenient for a firm still resolving internal governance and investor structure questions, and that talk of existential risk, while sincere, can also serve to justify continued concentration of control in a small leadership team rather than distributing oversight through public accountability mechanisms like regulatory disclosure and shareholder votes.
Explained simply
Sam Altman is the chief executive of OpenAI, the company behind ChatGPT and one of the leading developers of advanced artificial intelligence. OpenAI began as a non-profit but has been restructuring towards a more conventional corporate model, which has fuelled speculation about whether it might eventually list its shares on the stock market through an initial public offering, or IPO.
An IPO would let ordinary investors buy shares in the company and would raise significant funds, but it also brings extra regulatory scrutiny and pressure to deliver short-term financial results. This matters because OpenAI operates in a fast-moving and largely unregulated area of technology, where concerns persist about how safely powerful AI systems are being developed and controlled, including the possibility that such systems could become difficult for humans to manage.
Coverage · 2 sources
AI Business Companies Cybersecurity Technology