Science
Cryptographers surpass Google's secret quantum codebreaking claims, report says
Scientific American reports that Google's attempt to use zero-knowledge proofs to withhold an encryption-breaking quantum algorithm prompted researchers to independently replicate and improve the code.
The short version
- Google claimed in March 2026 that quantum computers could break standard elliptic curve encryption by 2030, using a zero-knowledge proof to keep the underlying algorithm secret.[Scientific American]
- The secrecy motivated outside researchers to independently replicate the work, with cryptologist André Schrottenloher and an online AI-assisted group each exceeding Google's benchmarks.[Scientific American]
- Following an initial flaw in Google's proof and subsequent replication by peers, Google co-author Craig Gidney advocated for open scientific publication over concealment.[Scientific American]
- It remains uncertain whether corporate researchers will continue experimenting with zero-knowledge proofs to guard proprietary findings or return to open disclosure.[Scientific American]
Key facts
- In March 2026, Google announced that breaking certain elliptic curve cryptography would require only 1,175 logical qubits and over two million Toffoli gates, withholding the exact algorithm via a zero-knowledge proof.[Scientific American]
- An error identified by computer scientist Keegan Ryan's team initially made Google's zero-knowledge proof invalid by allowing false claims, which Google corrected in mid-April 2026.[Scientific American]
- Cryptologist André Schrottenloher reproduced and improved Google's quantum algorithm after 63 days, publicly releasing a more efficient version.[Scientific American]
- An online developer community independently used AI-based agent systems to match and beat Google's findings within three days, nearly halving the needed Toffoli gates.[Scientific American]
- Google co-author Craig Gidney conceded that hiding scientific work behind zero-knowledge proofs was ineffective and endorsed open publication.[Scientific American]
What remains uncertain
- Whether corporate researchers with proprietary secrets will adopt zero-knowledge proofs as a standard publication method or abandon the practice remains unresolved.[Scientific American]
Sources
Outlet counts describe coverage, not independent confirmation. Reports may share a wire service or original source.
- Why mathematicians are hiding research results in zero-knowledge proofsScientific American metered