OpenAI

OpenAI is an American artificial-intelligence research organisation founded in December 2015 by Sam Altman, Elon Musk, Ilya Sutskever, Greg Brockman, and others under the banner of "artificial general intelligence that benefits all of humanity". It began as a non-profit, added a capped-profit subsidiary (OpenAI LP) in 2019, and in the recapitalisation completed on 28 October 2025 became a non-profit OpenAI Foundation in control of a public benefit corporation, OpenAI Group PBC. Through the GPT series and ChatGPT (2022) it sat at the centre of the generative-AI boom; on 31 March 2026 it closed a funding round of US$122 billion in committed capital at a US$852 billion post-money valuation.

From non-profit to public benefit corporation

At its founding OpenAI emphasised open publication; in 2016 it released the OpenAI Gym reinforcement-learning environment, and in 2017 popularised the PPO algorithm. Following the Transformer paper (Google Brain, 2017), it scaled that lineage stepwise — GPT-1 (2018), then GPT-2 (2019), whose staged weight release on stated safety grounds drew significant controversy.

In 2019 OpenAI grafted a capped-profit subsidiary, OpenAI LP, onto the non-profit and took a US$1 billion strategic investment from Microsoft, making Azure its compute platform. GPT-3 in 2020 established the commercial API business; GitHub Copilot followed in June 2021, and on 30 November 2022 ChatGPT was released as a free research preview. In January 2023 Microsoft announced a further multiyear, multibillion-dollar investment — it disclosed no figure; reporting put the total in the order of US$10 billion. GPT-4 came in March 2023, GPT-4V and DALL·E 3 in September, GPT-4o in May 2024, the o1 "reasoning model" family in September 2024, the SearchGPT prototype in July 2024, GPT-5 in 2025, GPT-5.4 in March 2026 and GPT-5.5 in April 2026.

Organisationally, in November 2023 the board removed Altman as CEO and he was reinstated five days later — the Altman firing crisis. The recapitalisation of 28 October 2025 left the OpenAI Foundation holding a 26 per cent equity stake in OpenAI Group PBC — worth roughly US$130 billion at the then-current valuation — plus a warrant that pays out in additional shares if the share price rises more than tenfold within fifteen years. Microsoft's holding was valued at approximately US$135 billion, about 27 per cent on an as-converted diluted basis; Microsoft states that, excluding the effect of the recent funding rounds, it had held 32.5 per cent of the for-profit. On 8 June 2026 OpenAI announced that it had submitted a confidential S-1 to the SEC, while saying the timing of any listing remained undecided.

ChatGPT, the API, and the public-sector tier

Key events

• February 2018: Elon Musk leaves the board.
• July 2019: OpenAI LP formed; Microsoft invests US$1 bn.
• June 2020: GPT-3 released, commercial API opened.
• 30 November 2022: ChatGPT launched.
• January 2023: Microsoft's further investment, amount undisclosed.
• March 2023: GPT-4 released. On 31 March Italy's data-protection authority orders ChatGPT suspended there; the order is lifted at the end of April.
• November 2023: Altman fired and reinstated within five days.
• May 2024: GPT-4o announced. Sutskever and Jan Leike depart and the Superalignment team is dissolved.
• July 2024: SearchGPT prototype.
• September 2024: o1 reasoning family announced.
• December 2024: Partnership with the defence company Anduril announced.
• January 2025: ChatGPT Gov launched for US government agencies; OpenAI joins the Stargate programme.
• June 2025: A US Department of Defense CDAO contract worth up to US$200 m.
• 28 October 2025: Recapitalisation completed — the OpenAI Foundation controls OpenAI Group PBC; Microsoft holds roughly 27 per cent.
• 31 March 2026: US$122 bn raised at an US$852 bn post-money valuation.
• 18 May 2026: Musk's suit against Altman and OpenAI fails on the statute of limitations; Musk says he will appeal.
• 8 June 2026: Confidential S-1 submitted to the SEC.

Key metrics

• Valuation: US$852 bn post-money, 31 March 2026 (OpenAI's own figure)
• Revenue: US$13.1 bn for calendar 2025; running at US$2 bn per month as of March 2026, an annualised rate of roughly US$24 bn
• ChatGPT: more than 900 million weekly active users (February 2026)
• Microsoft: post-recapitalisation holding valued at about US$135 bn, roughly 27 per cent as-converted diluted. OpenAI has contracted to buy an incremental US$250 bn of Azure services
• Compute: the Stargate programme (OpenAI / Oracle / SoftBank / MGX) targets up to US$500 bn of data-centre buildout

The founding promise against the speed of commercialisation

ChatGPT turned large language models into a consumer commodity and simultaneously rewrote Google's search business and Microsoft's cloud strategy. But the tension between the founding promise of safe AGI for everyone and the speed of commercialisation through the for-profit arm has destabilised the company repeatedly. The November 2023 board episode was read as the surfacing of a split between the safety wing and the commercial wing. Since then, safety-oriented founding-era staff have tended to leave: Sutskever started Safe Superintelligence in 2024 and Mira Murati founded Thinking Machines Lab in 2025. The Amodei siblings had already left to found Anthropic in 2021, before that crisis.

Litigation is heavy. The New York Times sued OpenAI and Microsoft for copyright infringement in December 2023. Elon Musk filed a "violation of founding mission" suit in 2024, dropped it, and refiled; on 18 May 2026 a federal court in Oakland rejected it as time-barred under a three-year statute of limitations, without reaching the merits of the charitable-trust claims. Musk has said he will appeal to the Ninth Circuit. Under the EU AI Act, the systemic-risk obligations placed on general-purpose AI fall first on frontier models like OpenAI's.

On labour and ethics, TIME reported in January 2023 that the labelling of toxic content for ChatGPT's safety filters had been outsourced through Sama to Kenyan workers paid less than US$2 an hour. On military use, OpenAI removed the "military and warfare" prohibition from its usage policy in January 2024, announced the Anduril partnership that December, and took a DoD CDAO contract worth up to US$200 m in June 2025. In June 2026 ChatGPT's share of the conversational-AI market was reported to have fallen below 50 per cent for the first time, as competition with Anthropic and Google intensified.