Resolution criteria
This market resolves to YES if, before December 31 at 11:59 PM UTC, a confirmed security advisory, peer-reviewed publication, or public proof-of-concept demonstrates that an AI system or AI-derived mathematical algorithm has broken a core cryptographic primitive or consensus mechanism of a major public blockchain (e.g., Bitcoin or Ethereum).
To resolve to YES, the event must satisfy at least one of the following criteria:
Private Key Recovery: A practical method derived via AI that allows recovering a private key from an exposed public key for standard signature schemes (e.g., ECDSA or Schnorr) using available classical hardware.
Hash Function Break: A viable preimage or collision attack against hash functions used by major chains (e.g., SHA-256 or Keccak-256) enabled by AI mathematical breakthroughs.
Consensus Break: A functional protocol-level attack generated or executed via AI that breaks double-spend protection or block validation on a top-20 cryptocurrency network without acquiring majority hash power or stake.
Theoretical warnings, social media posts without a published exploit, smart contract logic bugs, or vulnerabilities unrelated to protocol-level cryptography or consensus will resolve to NO.
Background
In October 2026, Ethereum Foundation researcher Justin Drake urged the crypto industry to prepare for "bunker mode" following OpenAI's release of hundreds of new AI-generated mathematical proofs. Drake posited that AI-driven mathematical discovery could accelerate cryptanalysis of Elliptic Curve Digital Signature Algorithm (ECDSA) faster than quantum computing. While Ethereum co-founder Vitalik Buterin acknowledged the potential long-term impact of AI on mathematical research, cryptographers from Coinbase and a16z noted that no practical exploit or attack vector against current elliptic curve cryptography has been demonstrated.
This description was generated by AI. Review and verify everything here yourself. You can edit, replace, or delete any part of this description, including the resolution criteria. You do not need to trust the AI output.
1,000