Hardware wallet implants found in Ledger and Trezor devices sold through Mediamarkt and Southeast Asian resellers -- $93.4M stolen. How the attack works, why Secure Elements do not protect against it, and what it means for crypto custody.
Your Hardware Wallet May Already Be Compromised — and the Industry Doesn't Want You to Know Why
For years, the advice has been simple: if you hold serious crypto, buy a hardware wallet. Not software. Not an exchange. A dedicated device with a Secure Element, air-gapped from your computer, designed so that your private keys never leave the chip. The security industry built an entire culture around this prescription. Ledger, Trezor, Coldcard — these names became synonymous with safety. The community repeated the catechism: not your keys, not your coins. The device was the last line of defense.
That line has been crossed. And the breach wasn't a software exploit, a phishing attack, or a firmware vulnerability that a patch could fix. It was a screwdriver.
Over the past week, the crypto world absorbed one of the most unsettling security revelations in its relatively short history: hardware implants — actual physical hardware, planted inside hardware wallet enclosures — have been stealing cryptocurrency from real users. Estimated losses now stand at $93.4 million across approximately 471 addresses, according to onchain analysis by Yfarmx and Bitquery. The attacker's primary method: capture the 24-word recovery seed phrase at the moment of device initialization, then transmit it over a live cellular network to a server controlled by the thieves. No cryptographic key needed. No signature forged. Just a second microcontroller and an eSIM, tucked inside the casing.
The targets included Ledger, the market leader, whose devices were found compromised through authorized resellers in Southeast Asia and, more troublingly, through Mediamarkt Saturn — Europe's largest consumer electronics retailer — in Germany. Rival manufacturer Trezor, long positioned as the privacy-conscious alternative, acknowledged the same day that it, too, had discovered implants in devices sold by third-party resellers, bringing the total number of affected hardware wallet brands to two.
The story raises a question that goes far beyond this specific incident: if the Secure Element can't be trusted because it isn't the only thing inside the box, what is the actual attack surface for a hardware wallet? And given that the attack works regardless of whether the device's primary chip is secure, does the entire category of "hardware wallet" need to be reconsidered?
How the Attack Works — and Why It Defeats the Security Model Entirely
To understand what happened, you need to understand what a hardware wallet actually is — and what most people incorrectly assume it to be. A modern hardware wallet like the Ledger Nano X contains two separate microcontrollers. The first is the Secure Element — a dedicated, tamper-resistant chip certified to standards like Common Criteria or SESIP — whose sole job is to generate cryptographic keys, sign transactions, and protect the seed phrase. The second is a general-purpose microcontroller that handles everything else: the display, the buttons, the USB or Bluetooth communication with the host computer, and the firmware that manages the user interface.
The security model rests entirely on the assumption that only the Secure Element ever touches the seed phrase. The second microcontroller — often running an ARM Cortex-M chip with no special tamper resistance — is considered untrusted. It talks to the Secure Element through a defined protocol, but it never learns the seed. The two chips are physically and logically separated. If the second chip is compromised, the seed remains safe. This is the foundational assumption of the entire hardware wallet industry.
The implants found in both Ledger and Trezor devices attack exactly this assumption — not by breaking the Secure Element, but by completely bypassing it. In the Ledger implants, researchers at Ledger's own security division, known as Donjon, found a second microcontroller planted inside the device enclosure, connected to the screen bus. During device initialization — the exact moment when the Ledger generates a fresh 24-word recovery phrase and displays it to the user — the implant's microcontroller watches the data traveling to the display, decodes each word as it appears, stores it in onboard memory, and then transmits the complete seed phrase to the attackers over LTE using a provisioned eSIM embedded in the implant board. The entire operation is invisible to the device's operating system and to any integrity checks the device performs.
Critically, the implant does not modify the Secure Element or the Ledger operating system. It does not alter transactions or signatures. It sits passively until initialization, then activates. Because the Secure Element is never touched, Ledger's own verification checks — the mechanisms the device uses to confirm it hasn't been tampered with — pass without flagging anything unusual. The attack works precisely because the implant is a second computer, inserted alongside the legitimate one, that watches everything the user does with their own hardware.
The Trezor implants operated somewhat differently, reflecting the architectural differences between the two products. Because Trezor's code is fully open source — a point Trezor has long cited as a transparency advantage — the attackers were able to obtain the firmware, modify it to include pre-generated seed phrases known to the attackers, and install it on a replacement microcontroller. The modified device would then "generate" a seed phrase for the user, but the phrase was one the attackers had already recorded. Because the malicious firmware ran its own version of the device's code, it needed to pass Trezor's software integrity checks; it did this by first running a legitimate firmware integrity check against a copy of the real firmware stored on an external microSD card, before allowing the malicious behavior to take over. The open-source architecture that Trezor advertised as a trust mechanism became, in this context, a roadmap for the attackers.
In both cases — Ledger and Trezor — the implants defeated the security model by targeting the moment of initialization rather than the moment of signing. The industry spent a decade hardening the signing layer: non-reentrancy guards, hardware-bound keys, Secure Element certification. The seed phrase generation step was treated as safe because it happened inside the certified chip. That assumption was wrong — not because the chip was broken, but because the chip wasn't the only thing in the room.
Joe Grand Found This in March. Ledger Did Nothing.
The most damaging chapter of this story may not be the theft itself, but what happened before it. Joe Grand — a renowned hardware hacker who has reverse-engineered everything from arcade cabinets to medical devices — began investigating Ledger wallets in August 2025 after a user on Reddit reported finding unexpected circuitry inside a Nano X purchased from an online marketplace in Thailand. The user sent the device to Grand, who confirmed the implant: an extra board with an LTE module, an eSIM, and an antenna, hiding inside the device casing.
Grand purchased eight more wallets from various online sellers across Southeast Asia — all from what he described as sketchy, unverifiable sources on Lazada, the Alibaba-owned e-commerce platform — and found that four of them were compromised with similar implants. He notified Ledger via email on March 4, 2026, and had a Zoom call with their team. He subsequently gave two conference presentations — at Tropicon in April and Hardwear.IO in May — and posted his slides publicly online.
Ledger's response was, by Grand's account, underwhelming. The company told him it would investigate to determine who was behind the implants but did not engage substantively with his findings. Grand said he received no meaningful follow-up for months, and when he eventually asked, a Ledger representative told him the company had found no evidence that someone had used the implants to steal cryptocurrency. There was no public disclosure. No customer warning. No advisory. The world's leading hardware wallet manufacturer had been told by a prominent security researcher that its devices were being counterfeited with active surveillance hardware — and chose to treat it as an isolated curiosity.
The situation escalated after Mark Karpeles — the former CEO of Mt. Gox, the Tokyo exchange that collapsed in 2014 with the loss of 850,000 bitcoin — encountered Grand's research online in September 2026 and decided to test the thesis himself. Karpeles purchased two Ledger wallets and found more sophisticated iterations of the implant than the ones Grand had documented. He published his findings online. The story spread rapidly through the crypto community. Within days, a coordinated heist drained approximately $93.4 million in cryptocurrency from wallets connected to devices sold by the Southeast Asian reseller Cryptobilis.
The attack may have been precipitated by the disclosure itself. Blockchain intelligence suggests the attackers had been sitting on the compromised devices — perhaps for months — accumulating wallets with large balances before pulling the trigger. Once news of the implants began spreading widely, the thieves appear to have rushed to extract funds from as many compromised wallets as possible before owners could respond.
Some in the community have praised Karpeles for exposing an operation that might otherwise have gone undetected indefinitely. His Mt. Gox history adds an ironic texture: the man who failed to protect the funds of thousands of Bitcoin users in 2014 may have been instrumental in preventing an even larger theft in 2026.
Authorized Resellers Were Compromised — and That's the Most Alarming Part
The implants Grand found in 2025 were all in wallets purchased from unauthorized, fly-by-night sellers — the scenario every hardware wallet manufacturer warns against. The standard advice — buy only from the manufacturer's website or from authorized resellers — was presented as sufficient protection. In the weeks following the disclosure, that advice held up in the sense that wallets purchased from reputable channels appeared clean.
Then the story changed. On October 9, 2026, Ledger disclosed that funds were being drained from wallets sold through Cryptobilis, an authorized Ledger reseller operating in Malaysia, Indonesia, and the Philippines. Onchain analysis by Bitquery traced losses to approximately 311 unique addresses, totaling roughly $92.9 million. Yfarmx's analysis put the figure at $93.4 million across 471 addresses. Both analyses agreed that the drain was executed in a single coordinated operation, with funds moved across seven blockchains — Tron, Bitcoin, Ethereum, BNB Chain, Polygon, Base, and Arbitrum — before being funneled through cross-chain bridges, THORChain, and Tornado Cash in what Chainalysis described as a "sophisticated cross-chain laundering operation."
More troubling was a separate report from a German user who purchased a Ledger Nano X from a MediaMarkt store in August 2026. The customer purchased a second wallet from a different MediaMarkt location on the same trip and found an implant in that one as well. MediaMarkt Saturn, Europe's largest consumer electronics retailer, operates in ten European countries with approximately 1,000 stores. It is listed as an authorized Ledger reseller in Austria, Germany, Spain, and Turkey. The company has since suspended sales of Ledger hardware wallets pending investigation.
Ledger's subsequent statement suggested a possible explanation — and a troubling one: the implants may have entered the supply chain not through direct counterfeiting of Ledger's own manufacturing, but through authorized resellers who obtained devices from illegitimate sources or who restocked returned products without proper verification. "We are reminding our authorized resellers and distributors to source products only through authorized channels and never to restock returned products," Ledger wrote in a public statement. "This is an ongoing threat that doesn't just impact Ledger, but the whole ecosystem."
In other words, the threat is not that attackers can compromise a factory or forge a device. The threat is that a legitimate reseller — one trusted by thousands of European customers who walked into a MediaMarkt store and bought a sealed box off the shelf — can be a vector for a supply chain attack that no software update can fix.
Trezor, Too: A Tale of Two Different Problems
On October 13, 2026 — the same day the Ledger situation was dominating crypto social media — Trezor issued a statement saying it would be reducing its authorized reseller network "out of an abundance of caution." Chief Commercial Officer Danny Sanders wrote that there was "not a single piece of evidence that Trezor devices have been compromised." This statement was, at best, incomplete.
Joe Grand had notified Trezor in August 2026 — prior to the company's public statement — about finding an implant in a Trezor Model T purchased via Facebook Marketplace in Canada, and a second implant in a Trezor One obtained from a TikTok shop in August. In the Trezor One case, the buyer had noticed approximately $50 transferred from his account before becoming suspicious. Both devices were discontinued models that Trezor still supports with software updates through 2031 and security updates through 2036, but no longer manufactures.
Trezor's initial response to Grand was courteous and detailed — the company's technical team correctly identified the implant as similar to counterfeit devices discovered on Russian marketplaces in 2022–2023, correctly noted the replaced microcontroller and modified firmware, and correctly described the use of an external microSD card to pass integrity checks — but did not publicly disclose the finding. When Grand wrote to Trezor again in early September asking about publishing his findings, the company's representative asked for a heads-up so it could prepare a public comment. Trezor's subsequent public statement claiming no evidence of device compromise was therefore surprising, given that the company had been in direct communication with Grand about exactly that scenario for two months.
The Trezor attack surface is architecturally different from Ledger's, which creates a different risk profile. Because Trezor's firmware is open source, the barrier to entry for an attacker is lower — they don't need to reverse-engineer a proprietary system, they just need to download the code. But the counterpoint is that open source also means anyone can audit the code, and Trezor users who know what they're doing can verify the binary running on their device. For most users, however — the majority who simply plug in their Trezor and follow the setup wizard — the open-source architecture provides no meaningful protection against a pre-planted malicious microcontroller, because the verification step requires technical knowledge and deliberate action that the standard user experience does not encourage.
The $93.4 Million Question: Can It Be Recovered?
Blockchain intelligence firms have been tracking the movement of stolen funds since the October 9 drain. Chainalysis confirmed it was "investigating reports of funds stolen from holders of Ledger products" and disclosed that its analysts had identified a sophisticated cross-chain laundering operation. Bitquery traced funds across seven separate blockchains and identified a pattern consistent with professional money laundering: rapid fragmentation, cross-chain bridging to obscure the trail, and mixing through protocols designed to break the onchain link between sender and receiver.
Recoverability, at this stage, appears unlikely. THORChain — the decentralized cross-chain liquidity protocol — has previously frozen stolen funds in some circumstances, but its ability to act is constrained by the protocol's design and the speed at which stolen assets can be moved. Tornado Cash, used in the laundering chain, is designed specifically to break onchain traceability. Tether and Circle froze approximately $318,000 of the stolen assets — less than 0.1% of the total — because the stolen assets were primarily in XRP, Ethereum, and Bitcoin, not the stablecoins that issuers can freeze.
Ledger has activated its User Protection Fund, which currently holds approximately $464 million, and has indicated it intends to compensate affected users. The company's response to the incident — compensation via a fund that was announced in 2023 following a previous data breach — has been more substantive than the typical industry response to a security incident. But the fund was designed for data breach scenarios, not physical device compromise, and questions remain about the scope of coverage and the speed of disbursement.
The Deeper Problem: Hardware Wallets Solved the Wrong Threat Model
Security researchers and cryptographers have long debated the threat model for personal cryptocurrency custody. The classic formulation — "protect your private keys from remote attackers" — led directly to the hardware wallet category: devices designed to generate and store keys offline, signing transactions without ever exposing the key to a networked computer. The threat model assumed the attacker was remote: a hacker, malware, a phishing site, a compromised exchange.
The attack demonstrated over the past two weeks targets a completely different threat actor: a physical adversary who can modify the device before or during the supply chain process, before the user ever receives it. This is not a new threat category. Supply chain attacks on hardware are as old as hardware itself. But the hardware wallet industry, in its rush to solve the digital threat model, appears to have underweighted the physical one.
The uncomfortable implication is that a hardware wallet's security is only as strong as the physical integrity of the entire distribution chain between the factory floor and the customer's hands — including every customs checkpoint, warehouse, regional distributor, and retail shelf. That chain is long, involves dozens of entities across multiple countries, and is staffed largely by people who have no way to detect a microcontroller swap or an antenna embedded in a plastic housing.
The Secure Element, the certification standards, the air-gapped design — all of these remain meaningful protections against digital attacks. But against a physical supply chain adversary capable of opening a sealed box, replacing a chip, resealing the enclosure with indistinguishable packaging, and selling it through an authorized reseller, the Secure Element is irrelevant. The seed phrase is compromised before the device is ever initialized.
What Users Can Do — and What the Industry Needs to Change
For users who already own hardware wallets purchased from third-party sellers, the situation is genuinely difficult. Ledger's advisory is clear: customers who purchased from Cryptobilis or who have devices purchased from any non-direct channel should not initialize those devices, should not enter existing seed phrases into them, and should transfer funds to a device purchased directly from the manufacturer or from a verified authorized reseller with a known, clean logistics chain. This advice, while correct, creates significant friction and cost for users who may have held assets on the same wallet for years.
For the industry, the incident demands structural changes that go beyond firmware updates or reseller audits. Several proposals have circulated among security researchers: hardware attestation at the manufacturing level, with cryptographic signatures applied at the factory and verifiable by the user before initialization; the relocation of seed phrase generation to a dedicated, physically isolated chip with no external communication capability and no external bus access; the elimination of general-purpose microcontrollers in the display and button path, so that no data leaves the Secure Element even in display form; and the establishment of an industry-wide incident response protocol for supply chain compromise, including mandatory disclosure timelines and compensation mechanisms.
None of these solutions are simple, and all of them impose costs — in manufacturing complexity, in device cost, in user experience — that the hardware wallet industry has historically been reluctant to absorb. The alternative is to accept that a hardware wallet is a useful but fundamentally limited component in a custody strategy: valuable protection against remote digital threats, but not against a physically determined adversary with access to the supply chain.
The irony is that the industry's response to this incident will likely be a combination of firmware updates, reseller audits, and an expansion of user protection funds — all of which address the aftermath of a physical attack without materially hardening the device against the physical attack surface that made it possible. The threat model has changed. The industry knows it. Whether it acts on that knowledge before the next $93 million disappears is the question that matters now.
Frequently Asked Questions
Did the Secure Element in Ledger devices get hacked? No. The Secure Element — the certified chip responsible for key generation and transaction signing — was not compromised in any of the documented cases. The implants targeted the secondary, non-secure microcontroller and the device's display bus, capturing seed phrases during the initialization process before they ever reached the Secure Element. This is precisely why Ledger's own integrity checks did not detect the compromise.
How were the attackers able to plant hardware inside sealed devices? The implants appear to have entered the supply chain through reseller channels, not through Ledger's direct manufacturing. In the Southeast Asian cases, devices were obtained from authorized regional distributors and modified before being resold. In the European cases, the mechanism remains under investigation, but Ledger's public statement suggested that returned devices may have been restocked without verification, or that devices were sourced from illegitimate channels despite the reseller's authorized status.
Are Trezor devices more secure than Ledger devices because they are open source? The answer is more complicated than it appears. Trezor's open-source firmware means the code is publicly available and auditable, which is a genuine security advantage in theory. In practice, the implant attacks documented in both Trezor One and Trezor Model T devices exploited the open-source architecture to create modified firmware that could pass Trezor's own integrity checks. The open versus closed source debate in hardware security may need to be reconsidered in light of these findings.
Can devices purchased directly from the manufacturer's website be trusted? Direct purchase from the manufacturer's website remains the lowest-risk channel available, because it minimizes the number of entities that handle the device between the factory and the customer. However, the broad scope of the compromise — spanning multiple brands, multiple countries, and both unauthorized and authorized channels — suggests that no channel can be treated as categorically safe. Security-conscious users should verify device integrity through available methods, including checking for unexpected components and, where possible, confirming the device's provenance with the manufacturer before initialization.
Has the stolen cryptocurrency been recovered? As of October 13, 2026, recovery of the stolen funds appears unlikely. Tether and Circle froze approximately $318,000 of the stolen assets, representing less than 0.1% of the estimated total. The remainder was moved across seven blockchains and mixed through cross-chain bridges and privacy protocols designed to break onchain traceability. Chainalysis is actively tracking the funds, but the sophistication of the laundering operation suggests professional actors who are unlikely to leave the assets in a traceable state.
What is Ledger's User Protection Fund and will affected customers be compensated? Ledger's User Protection Fund currently holds approximately $464 million and was established in 2023 following an earlier data breach. The company has indicated it intends to use the fund to compensate affected users from the Cryptobilis-related drain. The scope of coverage and the timeline for disbursement have not been fully specified as of this writing.
How does this compare to previous major crypto thefts? The $93.4 million drain is significant but not unprecedented in crypto's history. The Bybit hack of February 2025 resulted in approximately $1.5 billion in losses. The Ronin bridge hack of March 2022 drained approximately $625 million. What distinguishes the Ledger incident is not its scale but its mechanism: it is the first major documented case of physical supply chain compromise — implants planted in hardware before it reached users — resulting in direct theft of user funds at scale. The combination of physical access, cellular transmission capability, and authorized retail distribution represents a new threat vector that the industry has not previously had to confront.