Key Takeaways
- Ledger CTO Charles Guillemet rejects claims that AI could cheaply break Bitcoin’s ECDSA cryptography within months.
- A broader breakthrough against public-key cryptography could threaten banking, secure communications and critical infrastructure.
- Breaking ECDSA would not automatically defeat every encryption system; the impact depends on the attack’s scope.
Ledger Chief Technology Officer Charles Guillemet has pushed back against predictions that artificial intelligence could soon break Bitcoin’s cryptography cheaply, arguing that such a breakthrough would threaten far more than cryptocurrency wallets.
In a post on X, Guillemet challenged claims that AI could defeat the Elliptic Curve Digital Signature Algorithm, or ECDSA, within months.
He warned that a broader collapse of public-key cryptography would undermine banking, secure communications and critical infrastructure.
“If Asymmetric cryptography falls, Bitcoin will be the least of our problems,” he wrote.
His argument separates the potential consequences of a cryptographic breakthrough from evidence that anyone can achieve one today.
The Threat Extends Beyond Bitcoin Wallets
Bitcoin uses digital signatures to establish that someone controlling a private key has authorized a transaction. If attackers could derive that key from its public counterpart, they could potentially forge valid spending instructions.
The consequences would extend beyond blockchains if the underlying breakthrough also defeated cryptographic systems that protect other digital services.
Public-key cryptography helps computers establish trusted connections, authenticate identities and exchange secrets. TLS, the protocol behind secure web connections, combines authentication, key establishment and symmetric encryption to protect communications against interception and tampering.
A practical attack against widely used elliptic-curve systems could therefore threaten affected websites and services, alongside cryptocurrency holdings.
However, breaking ECDSA would not automatically break every encryption algorithm or every TLS connection. Different systems use different mathematical assumptions, and an attack’s scope would depend on the weakness researchers discovered.
Bitcoin also uses more than one signature scheme. Taproot uses Schnorr signatures rather than ECDSA, although both operate over the secp256k1 elliptic curve. An attack on the underlying discrete-logarithm problem could threaten both; an ECDSA-specific flaw would require separate analysis.
Mathematical Progress Does Not Establish an Imminent Attack
Guillemet’s central objection concerns the leap from impressive mathematics to predictions of immediate cryptographic failure.
He cited a recent integer-multiplication result claiming an improvement beyond the n log n barrier, with an exponent improvement of 2^-182. He described it as interesting mathematics, but argued that it does not change researchers’ understanding of how to break ECDSA.
Faster integer multiplication and recovering private keys from elliptic-curve public keys address different computational problems. Progress on one does not, by itself, demonstrate a practical solution to the other.
Assessing an alleged breakthrough would require an attack method, realistic resource estimates and independent scrutiny, not simply evidence that AI can produce mathematical results.
AI-assisted cryptanalysis also differs from quantum computing. The former would involve discovering or improving attacks using conventional computing; the latter introduces a different computational model capable, in principle, of defeating certain existing public-key systems.
Guillemet urged readers to monitor research and take cryptographic risks seriously while rejecting claims that theoretical progress already establishes an approaching catastrophe.
His skepticism concerns the claimed timeline and practical capability. It does not establish that today’s algorithms will remain secure indefinitely.
Security Planning Already Has a Starting Point
Organizations do not need proof of an imminent attack to begin preparing for future cryptographic weaknesses.
The US National Institute of Standards and Technology finalized its first three principal post-quantum standards in August 2024. They cover key establishment and digital signatures, providing alternatives designed to resist attacks from both conventional and quantum computers.
The standards include ML-KEM for establishing shared secrets, plus ML-DSA and SLH-DSA for signatures.
NIST encouraged administrators to begin integrating them because migration takes time. That process involves software, hardware, protocols and dependencies across organizations.
Those standards do not automatically upgrade Bitcoin. Changing its spending rules requires separate protocol work and coordination.
Preparation can proceed without treating speculative timelines as established facts. The practical challenge involves strengthening digital trust across financial and communications systems, while demanding evidence for claims that existing protections have already failed.
