OpenAI announced this week that it had cracked parts of the Navier-Stokes equations—a 90-year-old problem in fluid dynamics—in 88 hours of compute time. The internet erupted in wonder. Investors, meanwhile, were doing different math: how to justify another funding round when the company burns through cash faster than its servers burn through electricity.
Let’s be clear about what happened here. OpenAI’s AI did something genuinely clever with differential equations. That is real. The problem is that solving a century-old physics puzzle is exactly the kind of headline a company needs when it has just spent $6.6 billion in a single year and still does not have a clear path to profitability. It is the corporate equivalent of showing off your new car while your mortgage is three months late.
The Navier-Stokes problem matters to fluid dynamics researchers and maybe to aerospace engineers. It does not matter to OpenAI’s balance sheet, which is what actually matters to OpenAI’s investors right now. You can solve every unsolved equation in mathematics, but if your operating costs exceed your revenue by an order of magnitude, you have still just solved a math problem. You have not solved the math problem.
This is not to diminish the technical achievement. It is to point out that the timing is suspicious enough to make a satirist blush. When your business model is still theoretical and your burn rate is very real, you announce breakthroughs in pure mathematics—not because they move your business forward, but because they move the conversation away from your financial statements.
The real 90-year-old problem OpenAI needs to solve: how to make money.