Photo: Art Streiber/August Image
In December 2015 a small group of Silicon Valley investors and two large technology companies announced OpenAI, a nonprofit research lab with a single stated aim: to steer artificial intelligence toward outcomes that help people rather than harm them. The headline number was $1 billion, pledged jointly by Elon Musk, Sam Altman, Greg Brockman, Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon Web Services, Infosys and YC Research.
That figure was a commitment, not a cheque. Tax filings later showed that only about $133 million had actually arrived by 2021, and Musk, the best-known name on the list, left the board in 2018. What began as a charity-style lab has since become one of the most valuable AI companies in the world. Here is what was promised, why it mattered, and how the story turned out.
What was announced in December 2015
OpenAI went public on 11 December 2015. Altman, then president of the startup accelerator Y Combinator, and Musk were named co-chairs. Ilya Sutskever, a research scientist who had been working at Google, became research director, and Greg Brockman, the former chief technology officer of the payments company Stripe, joined to run the organisation day to day.
The founding team was small: around a dozen researchers, including Andrej Karpathy, John Schulman and Wojciech Zaremba, with advisers such as the deep-learning pioneer Yoshua Bengio and the computer scientist Alan Kay. The launch post said the lab would try to advance digital intelligence in the way most likely to benefit humanity as a whole, and that it would be free of the need to generate a financial return.
That last point was the selling idea. Most cutting-edge AI work at the time happened inside companies such as Google, Facebook and IBM, where research ultimately served products and shareholders. A nonprofit, the founders argued, could publish openly, share code and put safety ahead of profit.
Why Musk and others were worried about AI
Musk had spent much of 2014 and 2015 warning that advanced AI could be dangerous. Speaking at a Massachusetts Institute of Technology aeronautics symposium in October 2014, he compared building powerful AI to summoning the demon
, a phrase that travelled around the world within hours.
He was not alone. In late 2014 the physicist Stephen Hawking told the BBC that fully developed artificial intelligence could, in the worst case, end the human race, because a system able to improve its own design might outpace slow biological evolution. Around the same time, thousands of researchers signed an open letter organised by the Future of Life Institute calling for research to make AI systems robust and beneficial.
These fears were partly about science-fiction scenarios, but also about nearer problems: automated decision-making in finance, transport and weapons; the concentration of power in a handful of firms; and the difficulty of predicting when machines might match human performance on general tasks. The founders said that because no one could forecast when human-level AI would arrive, it was worth having a well-funded lab ready to prioritise good outcomes when it did.
How much of the $1 billion actually arrived
Many early news reports described the money as a "$1 billion fund" from Musk. That was never accurate. The billion was a combined pledge from several backers, and the founders said at the start that they expected to spend only a small part of it in the first few years.
Public filings later told a more modest story. According to OpenAI's own tax returns, the nonprofit had received roughly $133 million in total by 2021. OpenAI has said Musk's personal contributions came to less than $45 million. Large-scale AI research turned out to need far more computing power, and money, than charitable donations could supply.
Musk's departure and the shift to a profit-making model
In February 2018 Musk stepped down from OpenAI's board. The organisation said the move avoided a potential future conflict of interest as Tesla became more focused on AI for its self-driving cars, and that he would continue to donate and advise. Later accounts from both sides suggest there were also disagreements about control and direction.
A year later OpenAI changed shape. In 2019 it created a "capped-profit" company under the control of the nonprofit, allowing it to raise investment while limiting investors' returns to 100 times their stake. Microsoft invested $1 billion that year, added about $2 billion more by 2022, and announced a further multi-year investment, reported at around $10 billion, in January 2023.
The turning point for public awareness came at the end of November 2022, when OpenAI released ChatGPT. More than a million people signed up within days, and the product made the lab a household name.
Boardroom drama, lawsuits and the 2025 restructuring
The tension between the nonprofit mission and commercial pressure became visible in November 2023, when OpenAI's board removed Altman as chief executive, saying it had lost confidence in him. After an employee revolt and pressure from investors, he returned within days and a new board was formed, chaired by Bret Taylor.
Musk, by then running his own AI company, sued OpenAI and Altman in February 2024, arguing that they had abandoned the founding promise to work for the public benefit rather than for profit. OpenAI countersued and published emails suggesting Musk himself had once favoured a for-profit structure under his leadership. In May 2026 a jury found for OpenAI and its co-defendants, concluding that Musk had brought his claims too late; his lawyer said he would appeal.
Meanwhile, in October 2025, OpenAI completed a restructuring approved by the attorneys general of California and Delaware. Its business became OpenAI Group PBC, a public benefit corporation, and the original nonprofit became the OpenAI Foundation, which keeps a large minority stake and formal control. Microsoft also holds a large stake.
What the original promise still means
Looking back, the 2015 announcement captured a real dilemma that has not gone away. The people most worried about powerful AI were also among those racing to build it, on the theory that it is safer for careful, mission-driven groups to lead than to leave the field to others. Critics argue that this logic simply speeds up the race.
A few points are worth keeping in mind when you read about OpenAI's founding today:
- The billion was pledged, not paid. Actual donations were a fraction of the headline figure.
- Musk was a co-founder, not the sole funder. Altman, Brockman, Thiel, Livingston, Hoffman, AWS, Infosys and YC Research were all part of the pledge.
- "Open" has narrowed. OpenAI published code and papers freely in its early years but now keeps its most capable models' details private, citing safety and competition.
- Safety debates are mainstream. Questions that sounded speculative in 2015, such as who controls advanced AI and how to test it, are now discussed by governments and regulators.
Whether OpenAI has stayed true to its founding statement is still argued in courtrooms, boardrooms and newsrooms. What is clear is that a nonprofit announced with a modest research team in 2015 ended up shaping how the whole world thinks about artificial intelligence.
Sources: TechCrunch, 11 December 2015: OpenAI launches with backing from Elon Musk and Sam Altman; Wikipedia: OpenAI (funding, restructuring and lawsuit history); CNBC, 21 February 2018: Elon Musk is leaving the board of OpenAI; MIT AeroAstro Centennial Symposium, Elon Musk interview (video); first reported on this site via True Activist. Lead photo: Art Streiber/August Image.
Related:
COMMENTS