Sam Altman's massive bets: OpenAI eyes $100B round at $750B valuation

Inside his San Francisco office, Sam Altman keeps a slender stick of depleted uranium-238 standing upright on his desk. To prove it is completely harmless, he casually waves a Geiger counter over it.

That uranium rod is not alone. As detailed in a wide-ranging Forbes profile, the 40-year-old OpenAI CEO recently hauled a duffel bag to work filled with artifacts wrapped in bathroom towels. Among them: a 40,000-year-old stone hand ax, a 3,500-year-old bronze sword, a Concorde jet compressor blade, and an old GPU chip that helped train the earliest versions of ChatGPT.

Altman uses these items to illustrate how technology compounds across generations. But the modern numbers around OpenAI are just as dizzying. ChatGPT now claims more than 800 million weekly active users, OpenAI cleared over $13 billion in revenue last year, and the company was recently valued at $500 billion.

Walking back the holy grail

OpenAI is currently in talks to raise a mind-boggling $100 billion round that could value the startup at $750 billion or more. Around the tech world, major players are projected to pour $500 billion into AI data centers and chips this year alone.

Altman, who holds no direct equity in OpenAI and built his estimated $3 billion fortune on outside investments, has built an aura around his vision. As Y Combinator founder Paul Graham put it, Altman is exceptionally good at convincing people of things and getting them to do what he wants.

Yet that salesman instinct recently hit a bumpy patch. During the interviews, Altman boldly claimed that OpenAI had basically built AGI, or was at least very close to it.

Days later, according to The Verge, he walked it back. Altman clarified that he meant it as a spiritual statement rather than a literal one, conceding that reaching real AGI will still require many medium-sized breakthroughs rather than one giant leap.

Why it matters

Altman's verbal backtrack exposes the razor-thin line OpenAI is currently walking between visionary acceleration and pure promotional theater.

When you are seeking to raise $100 billion at a $750 billion valuation, you cannot afford to sound incremental. You have to convince sovereign funds and tech titans that artificial general intelligence is just around the corner. Claiming AGI has been achieved spiritually looks like a tell: the ceiling of pure model scaling is becoming harder to ignore, and the engineering gaps remain stubborn.

For developers and enterprises, this is a clear sign to calibrate expectations. Groundbreaking agentic workflows and human-level autonomy are not arriving overnight. The near term will be about grinding out practical efficiencies rather than relying on an imminent, god-like intelligence.

Altman resembles an Edison or a Steve Jobs more than a laboratory scientist. He drives the industry by pulling massive capital and infrastructure forward through sheer conviction. But when the tech world is spending half a trillion dollars a year building data centers on that conviction, spiritual milestones will soon need to deliver very real, balance-sheet-grade returns.