Onchain analyst Willy Woo posted AI-generated estimates on Sunday, Oct. 11, 2026, putting the odds that Bitcoin's digital signatures could be broken within two years at 1.3%. Bitcoin.com News then ran the question past roughly four AI models. One of those models put the risk to Bitcoin's security at 76% by 2046.
That's a wide spread — near-term, the models see almost nothing. Two decades out, at least one sees a coin flip.
Where the 1.3% figure comes from
The number is Woo's, and it's an AI-generated estimate, not a cryptographic audit. He published it on Sunday and attached it to the question of whether Bitcoin's signature scheme — the elliptic-curve math that proves you own the coins you're spending — could be cracked inside a two-year window. The models' answer: not likely.
That's the reassuring half of the story. A 1.3% probability is the kind of number that gets read as noise. It's not zero, but it's the sort of risk that doesn't move a market.
Why 2046 looks different
The longer horizon is where the models stop agreeing with each other. Bitcoin.com News consulted approximately four AI models, and at least one of them returned a 76% probability that Bitcoin's security would be at risk by 2046. Twenty years is enough time for quantum hardware to matter, and the models clearly treat it that way.
The gap between 1.3% and 76% isn't a rounding error. It's two different questions being answered. Near-term, the constraint is hardware — how many stable qubits you can actually assemble and run. Long-term, the constraint is math, and the math doesn't get safer with age.
What the numbers don't say
None of this is a protocol finding. These are language models estimating risk, the same way they'd estimate the odds of a hurricane or an election. They don't hold keys, they don't run Shor's algorithm, and they don't know what hardware exists in a lab right now that hasn't been published.
What they do is produce a consistent shape: negligible near term, contested over decades. That's roughly how the cryptographic community has talked about the problem too, though not with these specific percentages attached.
The upgrade question nobody's answered
Bitcoin's signature scheme is baked into the protocol. Replacing it means a soft fork, wallet migrations, and convincing a decentralized network to move — a process that historically takes years of argument before it takes months of code. The models were asked about risk. They weren't asked about remediation timelines, and those are different clocks.
Woo's Sunday post didn't propose a fix. The Bitcoin.com News follow-up didn't either. So the practical question is still open: at what probability does the network start treating post-quantum signatures as urgent rather than theoretical? A 1.3% two-year risk clearly doesn't clear that bar. A 76% twenty-year risk might — but nobody has said what the threshold is.
For now, the estimate lives on a social post and a set of model outputs. The next thing to watch is whether anyone with a say in Bitcoin's development puts a real number on the table, or whether this stays an AI parlor trick.
