Sam Altman

Sam Altman (born 22 April 1985) is an American entrepreneur who has been chief executive of OpenAI since 2019, with a break of a few days in November 2023. Under his leadership the company moved from a small research laboratory to the operator of ChatGPT, one of the fastest-adopted consumer software products in history, and to a central participant in the debate over how artificial intelligence should be governed 1.

Before OpenAI he founded a location-sharing start-up, Loopt, and ran Y Combinator, the accelerator that funded many of the most valuable technology companies of the 2010s. That background shaped his approach at OpenAI: a focus on distribution and scale, a willingness to ship imperfect products, and an argument that deployment itself is a form of safety research 2.

This entry covers his early ventures, his years at Y Combinator, the founding and growth of OpenAI, the board crisis of 2023, the company's restructuring, his stated views on intelligence and economics, and the criticism his record has attracted.

Altman grew up in St. Louis and studied computer science at Stanford University, leaving in 2005 before completing a degree to work on Loopt, a service that shared a user's location with friends. Loopt was among the first companies funded by Y Combinator, and it was sold in 2012 for a sum reported at about 43 million US dollars, modest by later standards but enough to establish him as a founder and investor 1.

He became a partner at Y Combinator in 2011 and its president in 2014, succeeding Paul Graham. During his tenure the accelerator expanded its batch sizes, created a continuity fund to invest in later rounds, and began supporting research-heavy companies in areas such as biotechnology and energy. The network of founders he built there became a lasting source of influence in Silicon Valley 3.

He left the presidency in 2019 to focus on OpenAI, though he remained involved with Y Combinator as a partner for a period. In the same years he became an investor in and adviser to companies working on energy, including the fusion company Helion and the nuclear start-up Oklo, and he co-founded Tools for Humanity, the company behind the Worldcoin identity project 2.

These commitments later became relevant to his OpenAI role, because they involve technologies — energy, identity, computation — that OpenAI also depends on. Supporters describe the overlap as useful expertise; critics describe it as a set of conflicts that are not always disclosed clearly 1.

OpenAI was founded in December 2015 as a non-profit with the stated goal of ensuring that artificial general intelligence benefits all of humanity. Altman was a co-founder and initially shared leadership with others, including Elon Musk; he became chief executive in 2019, the year Microsoft invested one billion US dollars and the company created a capped-profit subsidiary to raise capital 1.

The next three years were productive in research and quiet commercially. GPT-2 was withheld in 2019 over misuse concerns, GPT-3 was released as an API in 2020, and DALL·E and Codex followed. Altman's contribution was less about any single paper than about organisation: recruiting researchers, raising money, and arguing that scale would continue to produce capability 2.

ChatGPT's release on 30 November 2022 changed the company's trajectory. The product reached an estimated hundred million users within two months, and Altman became the public face of generative AI. He testified before the US Senate in May 2023, calling for licensing and safety standards for advanced models, a position that some competitors and open-source advocates regarded as an attempt to entrench the leading labs 1.

Through 2023 the company expanded its model line, signed a deeper partnership with Microsoft, and began building an enterprise business. Altman's public profile grew accordingly, and Time named him CEO of the Year in 2023, shortly before the board moved against him 1.

On 17 November 2023 the OpenAI board removed Altman as chief executive, stating that he had not been consistently candid in his communications with the board and that it no longer had confidence in his ability to lead. The announcement was made without warning to most employees or to Microsoft, the company's largest investor 4.

The reaction was immediate. Most of OpenAI's staff signed a letter threatening to resign unless Altman was reinstated and the board replaced, and Microsoft offered to hire the team. Within days the board that had dismissed him was itself replaced, and Altman returned as chief executive on 21 November under a new initial board chaired by Bret Taylor, with Larry Summers and Adam D'Angelo as members 4.

The episode exposed the tension built into OpenAI's structure. A non-profit board with a mission to govern advanced AI had legal authority over a company whose employees, investors and commercial partners expected conventional governance. The board exercised its authority and then lost the support required to sustain it 5.

An internal review commissioned by the new board was published in March 2024. It found that the prior board had acted in good faith but that its decision had been made without adequate process, and it did not substantiate the specific concerns behind the dismissal. Altman remained chief executive, and the episode became a reference point in debates about who controls frontier AI labs 1.

After the crisis OpenAI raised capital at a rapid pace. A funding round in October 2024 valued the company at about 157 billion US dollars, and a larger round in 2025, led by SoftBank, valued it at 300 billion. Secondary sales later in 2025 implied a valuation of about 500 billion, making it one of the most valuable private companies in the world 1.

The capital went into infrastructure. In January 2025 Altman announced Stargate, a joint venture with SoftBank, Oracle and MGX to build data centres in the United States, with commitments described as reaching 500 billion dollars over several years. The project is an attempt to secure the compute and power that frontier training requires, and it depends on financing and construction schedules that are not yet proven 6.

In October 2025 OpenAI completed a restructuring that converted its for-profit subsidiary into a public benefit corporation governed by the nonprofit foundation, which holds a controlling interest and a majority of board seats. Microsoft received a stake of roughly 27 per cent. The structure is intended to satisfy investors while preserving the foundation's mission control, and it has been criticised as a compromise that does neither job fully 5.

Alongside the money came governance changes and departures. Several senior safety researchers left in 2024, including the co-founder Ilya Sutskever and the alignment lead Jan Leike, and Leike said publicly that safety culture had taken second place to product. Altman acknowledged the criticism and said the company would invest more in safety, while continuing to release models at a faster cadence 1.

Altman's public writing sets out a consistent position. In "Moore's Law for Everything" (2021) he argued that artificial intelligence would drive the cost of many goods and services towards zero, and that the gains would accrue to owners of capital unless governments taxed capital and land and distributed the proceeds, for example through a citizen dividend. The essay is unusual among technology executives for taking redistribution seriously 7.

In "The Intelligence Age" (2024) the emphasis shifted to abundance and infrastructure. Altman argued that machine intelligence could deliver broad prosperity within a decade, that the constraint was compute and energy rather than algorithms, and that the transition would require large-scale construction of data centres and power generation. Critics noted that the essay described benefits more concretely than mechanisms 6.

On safety, his position is that iterative deployment is safer than secrecy. Releasing models gradually, he argues, allows society to adapt and researchers to find problems while capability is still limited; withholding them concentrates power and leaves the public unprepared. Critics reply that the argument is unfalsifiable and conveniently aligned with commercial interest 1.

On regulation he has favoured licensing for advanced models, testing regimes and an international body, while opposing rules that would slow deployment. In testimony and interviews he has compared the moment to the early regulation of aviation and nuclear energy, and has said he expects artificial general intelligence to arrive within a small number of years, a forecast he has repeated since 2023 2.

Altman's influence rests on three things: control of the leading consumer AI product, access to capital on a scale unusual for a private company, and a public profile that makes him a default spokesman for the industry. He speaks at governments and conferences, appears before legislators, and is treated by much of the press as the person who can explain what is happening in AI 1.

His supporters describe him as an unusually effective organiser who combined research ambition with product sense at a moment when most AI labs had neither. They credit him with turning a research agenda into a usable product, with securing the capital needed to compete, and with keeping the company intact through a governance crisis that could have destroyed it 2.

His critics describe a pattern of centralising authority, presenting contested decisions as inevitable, and allowing safety commitments to be revised when they conflict with growth. The 2023 board statement, the safety departures and the restructuring are cited as evidence. Altman has rejected the characterisation, saying that the company's mission requires it to be commercially successful and that criticism of his leadership is often motivated by competitors 5.

He remains a polarising figure in a field where technical reputations are usually quieter. That is partly a matter of temperament and partly structural: no other person is so closely identified with a technology that is simultaneously celebrated as a productivity revolution and feared as a risk to employment, information and political stability 1.

The most documented criticism concerns candour. The November 2023 board statement said Altman had not been consistently candid, and although the later review did not substantiate the underlying concerns, the phrase has followed him. Reporting since then has described instances in which employees or partners learned of decisions late, and critics argue that a leader of a safety-critical organisation should be held to a higher standard of disclosure 1.

The safety departures are the second theme. When Leike left in May 2024 he wrote that safety had been deprioritised in favour of products, and Sutskever's departure removed a co-founder who had been central to the company's research identity. OpenAI responded that it had created new safety bodies and increased investment, but the pattern of departures continued, and each one was treated as evidence by critics 1.

The commercial disputes are more concrete. In May 2024 the actor Scarlett Johansson said OpenAI had approached her about voicing an assistant and then used a voice she found similar without her consent; OpenAI said the voice belonged to a different actor and paused the feature. The company has also been sued by authors and newspapers over training data, and its use of copyrighted material remains unresolved in court 5.

His outside investments have attracted conflict-of-interest questions, particularly in energy companies that could benefit from OpenAI's power needs and in Tools for Humanity's identity token. Altman has said the investments are disclosed where required and that the technologies are complementary. The criticism is less that any single deal is improper than that the boundaries of his commitments are difficult for outsiders to see 2.

The dispute about Altman's leadership is partly factual and partly a disagreement about what counts as evidence. His defenders point to the company's growth, the absence of a substantiated finding of misconduct, and the board review that declined to endorse the original dismissal. Critics point to the original board's stated reasons, the repeated safety departures and the concentration of authority in a single executive, and argue that the review was commissioned by a board that needed him back 4.

His record on safety is read in two ways. One reading is that OpenAI under Altman has published more safety research, run more external red-teaming and deployed more cautiously than most competitors, and that its staged releases set the pattern the industry followed. The other is that the company repeatedly announced safety commitments and then weakened them when they became expensive, and that its own employees said so on the way out 1.

The forecasting dispute is simpler and less consequential. Altman has said AGI may arrive within a few years; many researchers regard that timeline as marketing, and the models he ships, while capable, still fail at tasks a child can do. He has responded that predictions are inherently uncertain and that the point of deploying early is to learn. Whether he believes the timeline or uses it to raise capital is not something the public record can settle 6.

What can be said with confidence is that he has shaped the field's institutions more than its techniques. The governance model he built, the product category he popularised and the capital scale he normalised are his distinctive contributions, and each of them is contested on grounds that are as much political as technical 5.