Trezor has disclosed a hardware-level vulnerability in the TROPIC01 security chip used in its flagship hardware wallet, Safe 7; however, user funds remain protected under current attack conditions.
The vulnerability was discovered by an independent audit.
This issue was identified by Donjon, Ledger’s independent security research team, during an audit. The auditors successfully extracted one of the three “secrets” protecting the user’s PIN by using a laser fault injection attack on the TROPIC01 chip.
This means that Safe 7’s original three-layer physical isolation will be reduced to two layers. However, Trezor states that merely compromising this chip alone is still insufficient to directly obtain the PIN or take control of the assets in the wallet.
The attack requires disassembly and laboratory equipment.
Trezor states that such attacks require a high threshold, including physical access to the device, hardware disassembly, and the use of specialized laboratory equipment. Therefore, the company believes the chip still provides effective protection, as executing such attacks is costly and time-consuming.
The blockchain security firm Cyvers also told foreign media that such attacks are “highly unrealistic.” The company believes that, from a practical risk perspective, ordinary users are still more commonly threatened by phishing attacks, mnemonic phrase leaks, malicious dApps, and blindly signing transactions without understanding their content.
Cannot be fixed via firmware update
Since the issue stems from a hardware-level defect, Trezor stated that it cannot be fixed via a firmware update. The report noted that Trezor has not yet responded as to whether it will accept refund requests from users.
Trezor also emphasized that users' private keys are not stored on the TROPIC01 chip, which is one reason the company believes funds remain secure. The company also stated that this vulnerability cannot be exploited to implant persistent malicious firmware on the device.
Additional information: This disclosure indicates that the associated risks are primarily concentrated in scenarios where devices are physically accessed and subjected to laboratory-level attacks, with no evidence of large-scale remote theft risks.