Sam Altman once laid out the stakes with blunt clarity. “I think that AI will probably, most likely, sort of lead to the end of the world,” he said. “But in the meantime, there will be great companies created with serious machine learning.”
The line dates to 2015. Altman delivered it during a fireside chat at Airbnb’s Open Air conference. Back then he led Y Combinator. He had just helped launch OpenAI as a nonprofit pledged to safe artificial intelligence that benefits all humanity. The remark captured a tension that has defined his career. Grave danger on one side. Extraordinary commercial upside on the other. That tension has only sharpened over the past decade.
From Nonprofit Ideal to Infrastructure Colossus
OpenAI began with a charter that put humanity’s safety above commercial success or even the company’s survival. Founders including Altman, Ilya Sutskever and Greg Brockman spoke of avoiding an “AGI dictatorship” and treating the work as a modern Manhattan Project. Sutskever sent private memos to the board warning that Altman misrepresented facts, deceived colleagues about safety protocols and showed a pattern of lying. Board members Helen Toner and Tasha McCauley concluded he could not be trusted with his “finger on the button.”
Yet Altman returned after the 2023 boardroom drama that briefly ousted him. The company has since pursued aggressive growth. It prepares for a potential trillion-dollar valuation. It builds vast computing infrastructure, including the $500 billion Stargate project announced with the Trump administration. Deals extend to Gulf states. Some data centers planned for the UAE would dwarf Central Park in scale. Altman has met with leaders in Saudi Arabia and the UAE to secure chips and energy described as the currency of the future.
These moves come as safety teams have been scaled back. The superalignment group dissolved in 2024. Product releases and revenue targets take clear priority. OpenAI now holds contracts for government use in immigration enforcement, surveillance and potentially autonomous weapons. It integrated models with Pentagon systems. Critics inside and outside the firm see a drift from founding principles. Dario Amodei of Anthropic has said the problem with OpenAI is Altman himself.
But the economic engine hums loudly. Nvidia, Microsoft and Google have seen valuations soar on AI bets. Altman predicts 2026 will bring systems capable of novel insights for businesses. By the 2030s intelligence and energy could become abundant enough to make previous limits on human progress feel quaint. “Astounding triumphs—fixing the climate, establishing a space colony, and the discovery of all of physics—will eventually become commonplace,” he wrote in a 2024 blog post. Wonders become routine. Then table stakes.
Yet the man who voiced existential fears has adjusted some forecasts. In a May 2026 interview at a Commonwealth Bank of Australia conference in Sydney, Altman said he feels “delighted to be wrong” about a jobs apocalypse. He once warned AI would replace most jobs people do today and wipe out entire categories. Now he observes that entry-level white-collar work has not vanished as quickly as expected. “I now think I understand more about why it hasn’t, and I’m obviously grateful,” he told the audience. Human interaction still matters in ways machines struggle to replicate. The jobs picture, he concluded, looks “likely to be very different than we thought.”
That shift does not erase earlier cautions. Altman signed the 2023 statement that mitigating extinction risk from AI should rank as a global priority alongside pandemics and nuclear war. He has long argued that superintelligent systems might not need malicious intent to cause harm. A machine pursuing some other goal could simply overlook humanity. Or outmaneuver its creators.
The New Yorker detailed these contradictions in April. Reporters examined internal records, Slack messages and interviews that portrayed Altman as a skilled operator who blurs the line between vision and overstatement. One former colleague likened the pattern to Theranos. Altman responded that he cannot change his personality and does this work because he loves it. He holds no equity in OpenAI, he has testified, though his influence and indirect stakes could make him among the most powerful figures alive if projections hold.
And those projections keep expanding. Altman sees AI agents joining the workforce this year in meaningful ways. He believes society stands past the event horizon. The takeoff has started. Intelligence explodes. Economic value concentrates. A few leveraged companies could dominate. Some analysts warn of a bubble. Altman himself said last year that someone will lose a phenomenal amount of money.
So the dichotomy persists. Companies flourish. Valuations climb. Tools already help diagnose diseases, draft code, generate ideas. Productivity gains appear in labs and offices. At the same time, lawmakers wrestle with regulation. Europe pushes stricter rules while Altman lobbies against measures he views as overly cautious. The U.S. government has reversed course on certain executive orders. National security implications grow as compute clusters spread across borders, including autocratic states.
Industry insiders watch closely. Researchers who once focused on capabilities now debate alignment with renewed urgency. Geoffrey Hinton, Yoshua Bengio and others have voiced parallel concerns. Yet commercial momentum proves hard to slow. Billions pour into training runs. Energy demands strain grids. Talent clusters in a handful of firms.
Altman’s 2015 quote resurfaced this month in TechRadar, which framed it as a stark reminder of the split between catastrophe and opportunity. The Economic Times ran it as quote of the day in January. Forbes Business Council published an analysis in February arguing the line should make leaders uneasy because great companies alone cannot address deeper governance gaps.
Recent coverage adds texture. Time reported Altman’s Sydney remarks on jobs in late May, highlighting how his views have moderated without abandoning long-term caution. The New Yorker’s long profile in April remains the most thorough examination of whether one person should hold such influence.
Critics argue the focus on distant extinction scenarios distracts from immediate harms: bias in models, job displacement in specific sectors, misinformation at scale, concentrated corporate power. Proponents counter that underplaying existential stakes would be reckless given the technology’s trajectory. Both sides agree the next few years will prove decisive.
Altman keeps moving forward. He predicts AI will uncover new scientific knowledge by next year. He bets on abundance. He acknowledges risks that keep probability-of-doom scores above zero. The companies keep forming. The infrastructure keeps rising. The question of control lingers. Who sets the guardrails? Can any single leader balance the race for dominance against the duty to protect what comes after?
The answer may arrive sooner than most expect. In the meantime the great companies multiply. Their products reach millions. Their leaders shape policy. And the original warning, delivered casually over a decade ago, refuses to fade.
WebProNews is an iEntry Publication