OpenAI Chief Executive Sam Altman declared at a fireside chat in Chapel Hill, North Carolina on 2 September that artificial intelligence has collapsed the startup development cycle from three months to as little as 17 minutes, predicting the technology will trigger the largest wave of entrepreneurship the world has ever witnessed.

The 17-Minute Claim

Speaking alongside US Commerce Secretary Howard Lutnick, Altman pointed specifically to Codex — OpenAI's coding agent — as the tool reshaping the economics of early-stage company building. 'The work that we expected a startup to accomplish in three months is now probably doable in, like, 17 minutes with Codex,' he said. He added that the broader world had not yet grasped the scale of what this shift would mean: 'This is just amazing. The world is not caught up with what this is going to mean.'

Barriers That AI Is Dismantling

Altman, who previously served as president of Y Combinator — the Silicon Valley accelerator behind companies such as Airbnb and Stripe — argued that promising ideas had long been strangled by resource constraints. 'If you had a great idea, but you didn't have the resources or the ability to get enough capital together or hire experts to work on it or to figure out what it's gonna take to invent new science or manage a complex supply chain, you can do that with AI,' he said. The implication is direct: the traditional gatekeepers of startup success — venture capital access, technical hiring, scientific expertise — are losing their grip.

The Laundromat That Started It All

To illustrate AI's real-world reach beyond Silicon Valley, Altman recalled an early ChatGPT user: a laundromat operator who used the tool to draft marketing materials, navigate legal contracts, and manage vendor relationships. 'He was really the first person I met that had figured out how to use AI to effectively run a business,' Altman said. The anecdote is notable because it predates the current generation of far more capable models, suggesting the productivity gains Altman describes today are compounding on an already-elevated baseline.

2025 Acceleration and What Comes Next

Altman noted that 2025 marked a visible inflection point: 'In 2025, we saw a real acceleration of this with coding agents and white collar work agents.' He predicted the next frontier would be long-term AI agents — persistent virtual colleagues capable of working continuously alongside individuals and organisations, rather than responding to one-off prompts. He framed this not merely as a productivity upgrade for large corporations and governments, but as a structural democratisation: 'It could enable more people to build independent businesses instead of becoming small parts of large corporations.'

Global Policy Dimension

Addressing G20 representatives at the same event, Altman acknowledged that countries could adopt different regulatory postures toward AI and choose whether to build their own data centres or lease computing capacity. However, he was unequivocal that participation was not optional: every economy would need to deploy the technology to remain competitive and sustain business growth. This comes amid intensifying geopolitical competition over AI infrastructure, with the US, China, and the European Union each pursuing distinct regulatory and investment strategies.