Ava Labs founder and CEO Emin Gün Sirer has a blunt message for anyone losing sleep over the prospect of artificial intelligence cracking the mathematics that secures their crypto wallets: the algorithm is probably not where the danger lies.

Speaking on October 9, 2026, Sirer argued that AI is far more likely to break into crypto systems through buggy software than by breaking ECDSA, the digital signature scheme that underpins most cryptocurrency transactions. He also disclosed that Avalanche, the blockchain built by Ava Labs, is evaluating backup cryptography options in case the signature scheme is ever compromised.

The Case, in One Sentence

"We will see AI exploiting system-level bugs long before ECDSA goes away," Sirer said.

The remark redirects attention away from cryptography itself and toward the layers built on top of it: blockchain software, libraries, and wallet applications. In his assessment, those everyday components form the more urgent attack surface.

ECDSA, or Elliptic Curve Digital Signature Algorithm, is the mechanism that proves ownership of coins when a transaction is signed, and it secures wallets across much of the crypto industry. Avalanche, which launched in 2020, relies primarily on ECDSA for transaction authorization, specifically a variant built on secp256k1—an elliptic curve configuration used throughout the sector and the same curve popularized by Bitcoin.

Sirer's comments followed a wave of AI-driven mathematical advances that reignited debate across the crypto community about whether cryptographic signatures remain safe. Research from Ethereum Foundation researcher Justin Drake has raised concerns that AI could accelerate breakthroughs affecting ECDSA signatures.

A Backup Nobody Hopes to Use

While downplaying the near-term risk to ECDSA, Sirer is not assuming the scheme will hold forever. Ava Labs is weighing a proposal from Haseeb Qureshi, a managing partner at the crypto venture firm Dragonfly Capital, for what has been described as a "cryptographic recovery mode."

Qureshi's "backup crypto" idea outlines an alternative cryptosystem that could be integrated into Avalanche to protect users if ECDSA ever fails. Separately, Sirer said an internal team at Ava Labs is actively exploring hash-based approaches as a hedge against vulnerabilities that AI might exploit. Hash-based schemes draw their security from hash functions rather than elliptic curve mathematics, so a break in one family would not automatically undermine the other. Hash-based signatures are also among the approaches studied in the field of post-quantum cryptography, where the U.S. National Institute of Standards and Technology has already published standards for hash-based signature schemes.

Sirer noted that there have been no confirmed exploits, and no immediate changes are coming to the Avalanche protocol. He offered no specific timeline or implementation details for any potential upgrades.

From Quantum Fears to AI Fears

Long-running debates about the durability of blockchain cryptography have historically centered on quantum computing, which—if built at sufficient scale—could run algorithms such as Shor's to recover private keys from public keys and break schemes like ECDSA. Sirer noted that discussions around Avalanche's security have traditionally revolved around long-term quantum resistance strategies. His latest framing takes a middle course: the cryptographic question deserves preparation, but the nearer-term battleground is everyday code.

What It Means for Avalanche and the Wider Industry

If AI-enabled attackers arrive at the software layer first, then the security of wallets, libraries, and node software deserves at least as much attention as the cryptography itself. Because ECDSA—and secp256k1 in particular—sits beneath a large share of the crypto ecosystem, any credible threat to it would not remain an Avalanche-only problem for long.

Swapping or supplementing the cryptography under a live blockchain is the kind of upgrade that requires coordination across developers, validators, wallet providers, and users.

For now, the signals to watch are fairly concrete: whether Ava Labs publishes details on Qureshi's recovery mode concept, what the internal hash-based research produces, and whether other networks reliant on ECDSA begin openly studying similar fallbacks.

Source: CryptoBriefing