Quantum computers dont exist yet in a form that can break todays encryption. But the data you are encrypting right now? Attackers can harvest it, store it, and decrypt it later when the hardware catches up. That is the core argument behind a push to start post-quantum migration with credentials, not general infrastructure.
The Global Risk Institutes 2025 Quantum Threat Timeline report shows 51-70% of surveyed security specialists believe a cryptographically relevant quantum computer will arrive within 15 years. Peter Shor proved back in 1994 that a powerful quantum computer could factor large numbers and crack discrete logarithms — breaking the public-key cryptography that systems use to establish trust. Symmetric encryption like AES-256 and modern hashing remain safe. The problem is specifically the key-exchange step.
What makes this relevant today is the harvest-now, decrypt-later tactic. Any encrypted traffic intercepted now should be treated as already exposed. Government deadlines are already moving: NSA requires new national security systems to start supporting quantum-resistant algorithms by January 2027, with full quantum resistance by 2035. NIST draft IR 8547 deprecates RSA-2048 and ECC P-256 after 2030 and disallows them after 2035.
Not all encrypted data carries the same risk. Session tokens expire in months. Credentials can persist for years — especially non-human identities like service accounts and API keys that nobody rotates. Those long-lived secrets are the highest-value targets for harvest-and-hold attacks.
A credentials-first migration starts with four steps. First, inventory your cryptographic dependencies — find your password managers, secrets managers, and PAM platforms. This phase usually exposes forgotten service accounts and hardcoded secrets. Second, prioritize by confidentiality lifetime and exposure reach, not system size. A small, long-lived secret that brokers access to critical systems outweighs a vast but short-lived dataset. Third, adopt hybrid cryptography combining classical and quantum-resistant algorithms so you are not betting on a single new approach. Fourth, build for crypto-agility — make algorithm swaps configuration changes, not re-engineering projects, by centralizing credential cryptography.
A full enterprise transition could take 5 to 15 years. The discovery phase alone can eat 1 to 2 years. Starting now is not premature. It is late.
