Hardware wallets have long been promoted as the gold standard for self-custody, isolating private keys offline from internet-connected threats. In 2026, a cascade of supply-chain compromises, third-party data breaches, and firmware failures has eroded that premise.
From tampered Ledger devices reported in Singapore a month before the story broke, to weak seed generation on Coldcards that enabled massive drains, and customer data leaks at Trezor, Keystone, and others that fuel targeted phishing, the cumulative evidence indicates these devices are no longer a reliably viable solution for secure long-term storage.
Singapore police documented at least two cases of tampered Ledger wallets purchased from local retailers since September 10, with combined losses of about S$3.5 million (roughly $2.7 million).
Preliminary findings pointed to malicious hardware components capable of capturing recovery phrases or private keys.
This local reporting preceded Ledger’s public alert on October 9 by nearly a month.
That alert concerned devices sold through the authorized Southeast Asian reseller CryptoBilis, where an unauthorized implant—hidden under the screen—read the 24-word seed phrase during setup and transmitted it over LTE via an eSIM.
On-chain researchers estimated drains in the $86–100 million range across hundreds of wallets; Ledger confirmed an implant in at least one affected unit while stating its own systems and direct sales channels were untouched.
The company later asked retailers in Singapore and several other Asian markets to suspend sales.
The Singapore timeline raises questions about earlier warning opportunities and the opacity of reseller distribution.
Even official “Genuine Check” verification can pass on tampered units because the Secure Element itself remains unmodified.
Similar implant reports have surfaced involving other sellers, underscoring that the attack vector is the supply chain after devices leave the factory.
Parallel problems have hit customer data held by third parties, exposing owners to phishing and physical risks without ever touching keys.
Ledger’s January 2026 incident involved its e-commerce partner Global-e, leaking order details such as names, emails, shipping addresses, and phone numbers.
Trezor suffered a larger exposure through shipping provider ShipMonk: an initial disclosure of roughly 14,000 recent customers expanded in September to include approximately 67,000 older U.S. records (totaling around 80,700) that should have been deleted under a 90-day policy.
The data included full names, addresses, phones, and order numbers.
Additional Trezor exposures via email and support vendors enabled phishing campaigns.
Keystone’s October website breach, traced to a third-party monitoring tool, exposed about 25,964 buyer records plus checkouts; attackers also stole the permanent WebAuth key burned into devices at manufacture for authenticity verification.
Keystone is destroying 18,835 unsold units as a result.
SafePal reported roughly 39,800 customer records compromised via an order-tracking plugin.
Collectively these leaks create high-quality target lists for social-engineering attacks that impersonate support, law enforcement, or the vendor itself—attacks that have already been observed.
Hardware wallets depend on physical delivery, yet the supporting logistics and commerce systems repeatedly fail to protect the associated personal data.
Coldcard’s entropy failure represents a different but equally damaging category of risk. A March 2021 firmware change (affecting Mk2/Mk3 builds 4.0.1–4.1.9 and earlier firmware on later models) caused seed generation to fall back to a weak software pseudo-random number generator instead of the intended hardware RNG.
Effective entropy collapsed to roughly 40 bits on older models and about 72 bits on Mk4/Mk5/Q devices, far below the 128-bit target for a 12-word BIP-39 seed. Devices behaved normally and displayed standard recovery phrases, yet the resulting keys were brute-forceable offline.
Beginning July 30, 2026, attackers swept funds; estimates of losses range from roughly $88 million to over $130 million in Bitcoin across thousands of addresses.
Updating firmware cannot strengthen seeds already generated under the bug.
Users who added substantial private dice-roll entropy or a strong passphrase were largely spared, but the default generation path was compromised for years.
Other vendors have faced related issues.
Ledger’s Donjon team has publicly demonstrated physical attacks against certain Trezor models involving glitching or supply-chain countermeasures.
BitBox has patched firmware vulnerabilities and dealt with phishing tied to shared newsletter providers.
Keystone’s stolen authenticity key raises the prospect of counterfeit devices that pass online checks.
These incidents share common roots: complex supply chains, opaque reseller networks, reliance on third-party processors for order and shipping data, and the inherent difficulty of generating and verifying true hardware entropy under real-world manufacturing and update processes.
The 2026 record shows that hardware wallets introduce new points of failure—tampering before first use, persistent weak seeds, and customer databases that become attack surface—while the core promise of isolated key security remains contingent on unbroken trust in manufacturers, resellers, and every intermediary.
Direct purchase, independent entropy sources, multi-signature setups, and rigorous verification help mitigate specific risks, yet the pattern of large-scale losses and exposures indicates these devices no longer provide a sufficiently reliable or viable standalone solution for high-value self-custody. Users must treat them as one imperfect component among others rather than a complete answer.
