On November 20, 2026, a mathematician named Levent Alpöge, who works at the AI company Anthropic, announced something that made headlines worldwide. He says that, working with the AI assistant Claude, he has found a proof of the Riemann Hypothesis, one of the most famous unsolved problems in all of mathematics. The claim has not been checked by independent experts yet. But even as a claim, it is a remarkable moment, so let’s break down what it means.
What is the Riemann Hypothesis?
Prime numbers are the building blocks of all whole numbers: 2, 3, 5, 7, 11, 13… They have no obvious pattern. They seem to pop up at random along the number line.
In 1859, the German mathematician Bernhard Riemann noticed that the pattern of the primes is secretly linked to a special mathematical function, now called the zeta function. This function has “zeros,” which are special inputs that make the answer equal to zero. Riemann guessed that all the interesting zeros line up perfectly on one single straight line. Computers have checked trillions of these zeros, and every single one sits on this line. But in mathematics, checking trillions of cases is not a proof. Someone has to show it works for all of them, forever.
Why does it matter?
• It reveals how primes behave. If the hypothesis is true, we know exactly how evenly the primes are spread out. They are random, but in a very controlled way.
• Hundreds of other results depend on it. Many mathematical papers begin with “assuming the Riemann Hypothesis is true…” A proof would turn all of those maybes into certainties.
• It connects to physics, the most exciting fact for me! The zeros behave strangely like the energy levels of quantum systems, which hints at a deeper hidden structure that we scientists would love to understand.
• There is a $1 million prize. It is one of the seven Millennium Prize Problems. Still, the money is the least interesting part.
What exactly is being claimed?
Alpöge says he spent roughly six weeks working with Claude and now believes the result is a full proof. He also says the method may extend to the Generalized Riemann Hypothesis, a broader version of the problem.
However, proofs this big take months to verify, and even brilliant ones sometimes hide small gaps. Alpöge says he has formalized the argument in Lean, a programming language mathematicians use to have a computer check every logical step, and he has released the Lean files. There is no human-readable paper yet, so experts now need to check two things: that the files state the Riemann Hypothesis correctly, and that they really prove it. Until that is independently confirmed, the right attitude is excited but patient.
This is also separate from an earlier result. In August, Anthropic reported that Claude had improved a long-standing bound on how many zeros are known to sit on the line. That was real progress, but it was not a proof of the hypothesis.
Will this break internet security?
No. This is the most common worry, and the answer is a calm no. Online banking, messaging and shopping rely on a lock called RSA. It works because multiplying two huge primes is easy, but taking the result and figuring out the original two primes is extremely hard.
The Riemann Hypothesis is about the overall pattern of the primes, such as how many there are up to a given size. It says nothing about how to find the two specific primes hiding inside one particular number. It is like knowing the average number of houses per street in a city: it does not help you find one person’s front door. So even if the proof is correct, your passwords, bank and messages stay exactly as safe as they were yesterday.
What happens next?
We must watch for three things: independent experts checking the Lean files, a readable paper explaining the argument, and any gaps people find. If it holds up, it will be a landmark in the history of mathematics, and one of the first times an AI helped crack a problem of this size. For now, let's enjoy the excitement, and let the mathematicians do their careful work.