The $39 Billion Question: Why OpenAI's Losses Don't Matter

OpenAI lost $39 billion last year. Let that number sit for a moment. Thirty-nine billion dollars. That is not a typo, and it is not a preliminary figure pending auditor review. According to the Financial Times, these are audited numbers from 2025—the year the company generated $13 billion in revenue and spent $34 billion keeping the lights on, the GPUs spinning, and Sam Altman's vision of artificial superintelligence funded.

Exclude restructuring charges and other non-cash items, and the loss shrinks to a merely devastating $8 billion. The distinction barely matters. What matters is this: OpenAI is now weeks away from filing confidentially for a U.S. IPO, and Wall Street is treating a $39 billion annual loss the way a casino treats a losing hand at 3 a.m.—as a reason to bet bigger.

This is not incompetence. This is not a company in distress. This is the complete inversion of profitability logic, and it is working precisely as intended.

For two decades, the public markets have operated on a simple principle: losses are bad. Losses are a signal. Losses mean management cannot allocate capital efficiently, cannot scale without hemorrhaging money, cannot survive without external funding. Companies have been punished for losing money. Startups have been killed by venture capital partners who tired of watching cash burn. IPOs have been delayed, repriced, or cancelled because the gap between revenue and expenses grew too wide.

That principle is dead. Not weakened. Not under pressure. Dead.

OpenAI's path to an IPO with a $39 billion loss sits alongside a revenue trajectory that is accelerating faster than most of Silicon Valley has ever seen. The company is on track to generate more than $40 billion in annualized revenue based on current performance—a number that roughly doubles its run rate from the end of 2025. The company expects losses of $14 billion in 2026 alone. Profitability is not forecast until 2030. HSBC analysts estimate OpenAI may need over $207 billion in additional funding by 2030 just to maintain operations.

These are not the financials of a company approaching maturity. These are the financials of a company that is betting everything on being the only company that matters in artificial intelligence. And the market is betting alongside them.

Consider SpaceX as a comparison point—not because the trajectories are identical, but because it offers a useful test case for how the market prices capital intensity and moonshot ambitions. When SpaceX went public, it carried a valuation of $85.7 billion and was burning through $18.4 billion in quarterly capex. The company was profitable. It was also burning cash at a scale that would have terrified most industrial conglomerates. Following its IPO debut, SpaceX's market value reached $1.86 trillion. The market was pricing not current profitability but future market dominance—and the capital required to achieve it.

OpenAI is playing a different game at a different scale. SpaceX builds rockets. OpenAI is attempting to build something that might reshape the entire structure of knowledge work. The company is not losing $39 billion because it is inefficient. It is losing $39 billion because it is spending $19 billion on research and development alone, plus nearly $6 billion on sales, marketing, and infrastructure. That is the cost of staying ahead of Anthropic, staying ahead of Google, staying ahead of whatever comes next.

Wall Street has stopped pretending that profitability proves viability. The new calculus is simpler and far more dangerous: if you are winning the only race that matters, the losses don't just get forgiven. They get celebrated as evidence of commitment.

The Financial Times reports that OpenAI is leaning toward waiting until 2027 to go public, with CEO Sam Altman cautioning that timing is undecided. That patience is revealing. The company does not need an IPO to survive. It needs an IPO because it needs $207 billion. The market is willing to provide it—not because OpenAI is profitable, but because the alternative to funding OpenAI is losing the AI race entirely.

This is not irrational exuberance. This is rational exuberance with a very specific bet underneath it: that in a world where artificial intelligence restructures human productivity, being the dominant player is worth any amount of interim loss. Whether that bet is correct will determine whether $39 billion in annual losses were a reasonable price or merely the opening act of a spectacular collapse. The market has chosen to wait and see. The rest of us are along for the ride.