E184 | Trezor CTO | How Quantum Computers Could Empty Every Bitcoin Wallet

A quantum computer capable of threatening Bitcoin is, in Trezor CTO Tomas Sushanka’s assessment, very unlikely in the near term, but not impossible. He estimates current machines have around 1,000 qubits, far below what would be needed to pose a relevant threat, and does not expect such a computer within five years or probably even ten. The probability is low, but the possible consequences are severe enough that Bitcoin’s developers and users should take the risk seriously.

Trezor calls its Safe 7 “quantum ready” because its firmware can be updated using post-quantum cryptography; that does not make either the device or Bitcoin quantum-proof. A sufficiently capable quantum computer could target the Bitcoin blockchain itself, so addressing the threat would require broader changes. Blockstream Research has examined potential Bitcoin changes, but agreeing on quantum-safe algorithms could take substantial discussion: Bitcoin’s decentralized decision-making is slow by design, and Sushanka says the issue has drawn serious attention in the community only over the past year or two.

Self-custody means controlling access to Bitcoin directly, rather than relying on an exchange where the customer is represented by a database entry and the custodian controls the assets. It can provide independent access, security, and privacy, but also makes users responsible for backups and protection. Sushanka says self-custody has become much easier than it was ten years ago, yet many people defer security because the threat feels abstract or the process seems difficult. He believes roughly half of people could manage it, far more than the 2–3% he cited as doing so, but still nowhere near all crypto users.

The Safe 7 stores private keys across three chips: a generic chip, an Infineon secure element, and a Tropic Square chip. Sushanka says compromising one or two chips would not be enough to obtain the keys; a PIN adds another barrier, and a passphrase can add further protection. An attacker would need to steal and disassemble the device, and he says there are no known remote exploits for its chips. The discontinued Trezor Model 1 has no secure element, and its chip’s information may be retrieved with proper equipment. Trezor has pledged another ten years of security updates, and Sushanka says significant holders may want to consider upgrading, while stressing that the Model 1 is not useless.

Scams increasingly combine phishing emails, WhatsApp messages, and calls, with AI-generated voices or deepfake video making contact seem legitimate. Sushanka describes fake Celsius recovery messages seeking seed phrases and a deepfake Microsoft Teams call in which attackers tried to persuade two employees to install an extension; both refused after treating an unexpected installation request as a red flag. He also warns that attackers can build trust before making a malicious request, citing the Drift hack, where attackers reportedly deposited one or two million dollars on a protocol to establish credibility. AI can aid attackers with deepfakes and malicious code, but can also help developers find security bugs.

Seed words generate private keys and should be kept offline, but entering them into a phishing site impersonating a wallet provider can expose funds. A backup is no help if its owner cannot find it or does not know whether it is correct, so Sushanka urges users to locate and verify their backups rather than postponing the task. Inheritance is another unresolved challenge: recovery should not depend on services such as Google or Apple. An ETF may simplify access for heirs, and may suit some investors for tax reasons, but requires reliance on a broker. Sushanka favors gradual adoption—starting with a trusted person or a small amount in a hardware wallet, then building confidence—while emphasizing that the right balance between self-custody and outside providers depends on the user.