Coinbase CEO Identifies Quantum Computing as Top Long-Term Threat to Crypto

Coinbase CEO Brian Armstrong informed investors that quantum computing, not artificial intelligence, represents the most significant long-term risk to cryptocurrency security. He made this statement during the exchange’s third-quarter earnings call as the company begins upgrading its custody systems to withstand future quantum attacks. His perspective aligns with NIST‘s now-fixed transition plan dates and shrinking estimates of the hardware required to compromise current cryptography.
The Threat and Coinbase’s Response
Armstrong distinguished between AI potentially discovering new ways to defeat existing cryptography and quantum computers performing calculations that fundamentally undermine it. He noted that four decades of attempts to break elliptic-curve cryptography using improved algorithms have “basically made zero progress,” according to the call transcript reported by Yahoo! Finance. Quantum technology, he stated, is “a different story,” calling it “a matter of when, not if.”
Coinbase has initiated efforts to protect its custody infrastructure, which holds customer assets, Armstrong confirmed. He also urged the underlying blockchains for these assets to adopt their own quantum-resistant security measures. In January, the company established an advisory board focused on quantum computing and blockchain to guide the industry. Armstrong expressed hope that the Bitcoin community would agree on a path forward within the next few quarters, a point he reiterated on the call.
The Timeline and Data Behind the “When”
The phrase “when, not if” suggests a looming schedule. NIST’s transition plan will deprecate RSA, ECDSA, and related algorithms after 2030 and prohibit their use after 2035, with earlier deadlines for high-value federal systems. Though these deadlines primarily apply to federal agencies, they will inevitably impact contractors and suppliers.
Key figures from the provided sources:
- NIST IR 8547 targets 2030 for key-establishment, 2031 for signatures, and full migration by 2035 for federal systems.
- Alice & Bob projects, theoretically, that breaking RSA-2048 would require 22,593 cat qubits and 3.14 days, and ECDLP-256 would need 29,588 qubits and 61.8 minutes, according to the company. These are estimates, not demonstrations.
Separately, Google researchers modeled a future machine capable of recovering a private key in approximately nine minutes after obtaining the public key, assuming some pre-computation. This describes a modeled capability, not an actual attack on Bitcoin.
Integrating post-quantum signatures into blockchains is a complex task. StarkWare has addressed this by launching a public Starknet Post-Quantum Hub to track its migration with verifiable evidence.
For additional information on Post-Quantum Cryptography deadlines and exchange responses, refer to QNews coverage of the migration timeline.