The company thesis

The technical goal is a quantum computation that recovers an elliptic-curve private key from its public key at a size used by deployed systems, executed reliably enough to matter. Published resource estimates for that computation have fallen as arithmetic circuits, compilation, and error-correction schemes have improved, and hardware demonstrations have moved fault tolerance from theory toward engineering. Neither trend shows that deployed cryptography is broken today.

Three kinds of work must come together. Circuit improvements reduce the logical resources a computation needs. AI research infrastructure widens the set of candidate improvements that can be proposed, tested, and recorded. Hardware integration determines whether a logical circuit can actually execute on a specific machine, with its architecture, error characteristics, and interfaces. A better circuit alone is not a computation, and a machine alone is not an attack.

The standard defence against a future quantum attack is migration: owners move assets to signature schemes designed to resist it. Owners who no longer possess the credentials needed to authorize transactions may be unable to complete an ordinary migration. What happens to those assets also depends on the rules adopted by the network. Potomac’s recovery-and-return program is intended to address this problem through technical research and a governance framework developed before operations begin.

Networks are debating how to respond. Draft Bitcoin proposals such as BIP 360 and BIP 361 and Ethereum research on emergency recovery describe possible migration and rescue mechanisms. They are proposals, not adopted network policy, and a company cannot promise that historically spendable assets will remain spendable by the same means. Lost keys are a problem that protocol changes and recovery governance both affect; neither makes inactive assets ownerless.

Quantum hardware companies develop devices and their native capabilities. Potomac develops what runs on them for this problem: cryptanalytic algorithms, AI research systems, and the software that integrates an application across compilation, fault tolerance, and execution. We intend to be a customer and application partner for those companies, buying compute access and running joint experiments rather than building machines.

The defensible capability is execution: turning a mathematical improvement into a workload that runs reliably on a real system, and being able to show, with reproducible evidence, what it cost and how well it worked. That capability compounds across hardware generations and providers.