Moving Q-Day Left: How AI Swarms Halved the Quantum Cost to Break ECDSA

Assessments of “Q-Day”—the arrival of a cryptanalytically relevant quantum computer—routinely focus on hardware roadmaps such as physical qubit counts or gate fidelities. However, this perspective overlooks an equally critical vector: algorithmic compression. When the logical circuit cost to execute Shor’s algorithm drops, the physical hardware threshold required to compromise public-key cryptography falls in tandem.

This session presents a case study on how an ambitious responsible-disclosure initiative by Google Quantum AI—proving their breakthrough via an open zero-knowledge verification suite—unexpectedly sparked the largest optimization of Shor’s reversible elliptic-curve kernel to date: the community-driven, open-source ECDSA.fail challenge. By embedding LLMs into automated simulation and formal verification loops, agentic pipelines uncovered non-trivial optimizations that cut Google’s classified benchmark score by over 60%.

Attendees will explore how agentic AI is transforming quantum circuit compilation, examine how to differentiate genuine algorithmic breakthroughs from test-harness reward hacking, and see firsthand how this heralds a new era of AI driven hardware-software-algorithm co-design of Quantum technologies.