Not all encryption is equally exposed. A large quantum computer would break the math behind how computers agree on keys and prove identity, while the ciphers that scramble your stored data mostly hold. That difference decides where you spend money first.
Shor’s algorithm solves the math these rely on.
Grover’s algorithm reduces their safety margin; it doesn’t break them.
No known quantum attack meaningfully weakens these.
Estimated size of the quantum computer needed to break RSA-2048. Better algorithms and error correction keep cutting the number. Real machines are still far smaller than any of these.
The 2026 figure is from a preprint that had not been peer reviewed at the time of writing.
Anything that must stay private past the dashed line is a harvest-now, decrypt-later target if it travels over today’s encryption. Typical ranges; your retention rules may differ.
Fix key exchange first for anything carrying long-lived data, because that’s what harvest-now attacks target. Signatures and certificates matter too, but an attacker can’t go back in time to forge yesterday’s login.