Hardware wallet security: Ledger and Trezor models compared

But the disclosure forced the entire hardware wallet industry to confront an uncomfortable truth: every chip, even the auditable ones, has attack surfaces. If you're self-custodying meaningful capital on-chain, this is the hardware wallet comparison you need to read before you buy.

I've personally held, tested, and stress-tested both ecosystems over multiple cycles. The Ledger vs. Trezor debate is no longer a clean ideological split between "closed source = bad" and "open source = good." It's a layered engineering conversation about chipsets, certifications, recoverability, and the trade-offs you're actually signing up for. Let's break it down.

The Core Security Split: Closed-Source Silicon vs. Open-Source Transparency

Ledger's entire lineup — Nano S Plus, Nano X, Flex, and Stax — runs a proprietary operating system called BOLOS on top of certified Secure Element chips (the ST33 series from STMicroelectronics). Your private keys never leave that chip. The firmware is closed-source, which means you cannot independently audit the code running on the device. You're trusting Ledger's certification claims and their track record.

Trezor took the opposite philosophical stance for over a decade: open-source firmware, open-source hardware schematics, auditable by anyone. The trade-off was that older models (Model One and Model T) shipped without a Secure Element chip at all, leaving them physically extractable through techniques like voltage glitching — the very technique Ledger Donjon famously demonstrated on a Trezor in 2020.

But the philosophy has shifted. Trezor's newer lineup now leans on Infineon's OPTIGA Trust M (V3) chip — a closed-source EAL6+ certified Secure Element — to protect the PIN and seed at rest. The chip itself is closed, but the firmware interacting with it remains fully auditable. The Trezor Safe 7 pushes further with a dual Secure Element architecture: OPTIGA Trust M (V3) plus the TROPIC01 from Tropic Square — billed as the world's first independently auditable, open-source Secure Element. This matters because the supply chain for these chips is concentrated in European vendors, mirroring the same push for sovereign secure hardware we see across strategic tech infrastructure — including Europe's recent moves to fund sovereign intelligence platforms for space operations.

Ledger bets on certified, closed silicon you can't inspect. Trezor bets on open firmware you can. Both now ship Secure Elements. The debate moved from ideology to engineering.

Physical Attack Resistance: From Voltage Glitching to EAL6+ Chips

Hardware wallet security ultimately comes down to what happens when someone gets physical access to your device. The Common Criteria certification levels — EAL5+ and EAL6+ — measure how rigorously a chip has been evaluated against tampering, but certification alone doesn't tell the full story. The real question is how each device responds when an attacker has your hardware in hand and unlimited time.

Ledger's Nano X ships with an EAL5+ Secure Element. The higher-tier devices — Flex and Stax — run EAL6+ chips. Trezor Safe 3 and Safe 5 also use EAL6+ OPTIGA Trust M (V3) chips. At the certification level, the modern flagships are roughly neck-and-neck.

The real divergence is in threat model:

• Voltage glitching — feasible on older Trezor Model One and Model T units, requires physical access, and was demonstrated publicly by Ledger Donjon in 2020. If you hold a Model T, treat it as legacy hardware.
• Side-channel and fault injection attacks — relevant to all modern Secure Elements, including the TROPIC01 in the Safe 7, and the ST33 series in Ledger's lineup. These are lab-grade attacks requiring specialized equipment, but they're not theoretical — they're demonstrated.
• Supply chain risk — both vendors source chips from a small number of European and Asian foundries. Closed-source firmware means you can't verify what was loaded onto the chip at the factory. Open-source firmware means you can check the code, but not the silicon gates beneath it.
• Brute-force and PIN extraction — this is where device policies diverge sharply. Three incorrect PIN entries on any current Ledger device trigger a factory reset, wiping the seed entirely. Trezor takes a different approach: repeated incorrect PIN entries trigger exponential backoff delays — the waiting time between attempts grows longer with each failure — eventually rendering the device effectively locked without immediately wiping the seed. The specific lockout thresholds vary by model and firmware version, and the Secure Element in newer Safe-series devices adds its own tamper-response layer. The practical takeaway: both ecosystems make brute-force PIN attacks impractical, but they achieve it through different mechanisms.

The table above reflects an important nuance that casual hardware wallet reviews tend to flatten: Ledger and Trezor reach roughly equivalent levels of PIN protection, but through architecturally different decisions. Ledger wipes fast and hard. Trezor delays and locks out. If your threat model involves a thief who grabs the device and immediately starts guessing, both approaches stop the attack. If your threat model involves forensic equipment that bypasses the PIN entirely, the Secure Element certification level is what matters — and that's where the EAL6+ ratings come in.

The TROPIC01 Disclosure: What Actually Happened in June 2026

This is the one everyone in the space is still talking about. In June 2026, Ledger Donjon disclosed a vulnerability in the TROPIC01 Secure Element used in the Trezor Safe 7. The attack vector was laser fault injection, which can in theory bypass signature verification on the chip.

Two things matter here:

First, no user funds were lost. Trezor confirmed that the Safe 7's multi-layered security architecture — combining the TROPIC01 with the OPTIGA Trust M, plus the PIN rate-limiter and the open-source firmware checks — prevented exploitation in a real-world setting. The vulnerability required highly specialized lab equipment, physical device access, and precise targeting. This is not a "plug in a USB and drain the wallet" scenario.

Second, the disclosure proved a thesis. Tropic Square built the TROPIC01 specifically to be independently auditable. The vulnerability being found and reported — rather than exploited in the wild — is the open-source security model working as intended. Bugs get found. The question is whether the response is fast and transparent. In this case, it was — Trezor acknowledged the finding promptly and began coordinating on a mitigation path.

An open-source chip got hacked by a motivated lab. That's the model working — not failing. You want your security research to happen in the light, not in the dark.

Here's the operational reality for Safe 7 owners: check with Trezor directly for the latest firmware status and whether a verified mitigation addressing the TROPIC01 vulnerability has been released. Don't assume — confirm. Until you've verified that your firmware version addresses the disclosed attack vector, treat the device as potentially exposed to this specific lab-grade scenario. For most users with reasonable physical security practices, the risk remains theoretical. For high-value targets, firmware verification becomes non-negotiable. This is the operational hygiene that separates people who actually custody their own keys from people who just hold them.

Recovery Mechanisms: Ledger Recover vs. Shamir Backup

This is where ideology hits hardest. In 2023, Ledger introduced Recover — an optional, paid service that splits your seed phrase into three encrypted fragments, distributes them to custodians, and reconstructs the seed after identity verification. The rollout was rocky. Critics argued that any firmware capable of exporting seed fragments — even encrypted ones — creates a structural attack surface for future updates.

Ledger's position: Recover is opt-in, runs on certified hardware, and the fragments cannot be reassembled without your identity. Users who never enroll in Recover retain full self-custody. The counter-argument from the security community is that the capability to export seed material — even if gated behind opt-in and encryption — is a systemic property of the firmware, not just a feature toggle. If you don't trust that the export path is truly sealed when you haven't enrolled, Recover is a reason to look elsewhere.

Trezor's answer is Shamir Backup (SLIP-39), a standard that lets you split your seed into multiple shares directly on the device — for example, 3-of-5 shares required to reconstruct. No third party is involved. You distribute the shares to lawyers, family members, safety deposit boxes, or geographically separated locations. There is no cloud custodian, no KYC step, no subscription.

The elegance of Shamir is that it solves the single-point-of-failure problem — your seed phrase written on one piece of paper in one location — without introducing any external trust assumptions. The trade-off is operational complexity: you need to manage multiple physical shares, ensure they're stored securely and separately, and test your recovery process before you need it for real. If any three of your five shares are destroyed or inaccessible, you lose access. It's a different failure mode than "I lost my paper," but it's still a failure mode that requires planning.

For high-net-worth self-custody, Shamir is the more trust-minimized option. For users who fear losing their seed and want a recovery path that doesn't require trusting their own backup discipline, Recover is a real product for a real problem. Just read the documentation before you opt in — and understand that you're trading some sovereignty for recoverability convenience.

Ecosystem Versatility: Asset Support, Connectivity, and Price

Beyond security philosophy, you need a device that fits how you actually transact day-to-day. The best secure offline crypto storage is useless if it creates friction that pushes you back toward hot wallets.

The 5,500+ vs. ~1,400 asset gap matters if you rotate into long-tail altcoins frequently. If you mostly hold BTC, ETH, and a handful of majors, Trezor's coverage is more than enough. If you actively trade emerging tokens across EVM chains, Cosmos, Solana, and the next wave of L2s, Ledger's broader support saves you from constantly bridging between hot and cold storage. This is where Ledger's ecosystem genuinely earns its premium — it's not just a bigger number on a spec sheet, it's fewer times you leave your keys on an exchange or a browser extension.

Bluetooth is a convenience feature and a small attack surface expansion. Wired-only is the purist's choice. The Trezor Safe 7 finally added Bluetooth — a concession to mobile-first users who refused to carry a USB-C cable everywhere. If your threat model prioritizes reducing every possible wireless attack vector, stick with wired devices and disable Bluetooth at the OS level.

Pricing tells its own story about target audiences. Trezor's Safe 3 at $59 is the cheapest entry point into a modern, Secure Element-equipped hardware wallet — it's a serious device at an impulse-buy price. The Safe 5 at $129 hits the sweet spot for most self-custody users: EAL6+ chip, touchscreen, open-source firmware. The Safe 7 at $249 competes directly with the Ledger Flex, trading asset breadth for the dual-SE architecture. Ledger Stax at $399 is the premium tier with a curved E-Ink touchscreen that's more about daily carry aesthetics than security differentiation.

One detail worth noting: Trezor's open-source approach means third-party developers can build on top of the platform — community firmware, alternative wallet interfaces, custom integrations. Ledger's closed ecosystem is more polished out of the box but more constrained. If you're the kind of user who runs Electrum or Sparrow instead of the manufacturer's desktop app, Trezor's interoperability story is stronger.

What I'd Actually Buy

If your priority is maximum trust minimization, open-source firmware you can audit, and a recoverability scheme that keeps your seed off third-party servers: Trezor Safe 5. It's the best balance of Secure Element certification, open-source transparency, and price. Pair it with Shamir Backup and you've built a self-custody stack that doesn't depend on any single vendor's continued goodwill.

If your priority is asset breadth, mobile Bluetooth convenience, and a polished hardware experience: Ledger Flex. You get EAL6+ certification, the largest altcoin support in the industry, and a touchscreen UI that actually works well. Just skip Recover unless you've specifically evaluated the trade-off and accepted the implications.

If you're a hardware security maximalist with capital to deploy: Trezor Safe 7 — but verify the firmware status with Trezor before loading significant funds. The TROPIC01 vulnerability was a wake-up call, not a deal-breaker. You want auditable silicon. You just want confirmation that the mitigation has shipped and been independently validated before you trust it with your full stack.

Three steps before you move funds:

1. Buy directly from the manufacturer. No Amazon, no eBay, no third-party resellers. Tampered hardware is a real, documented attack vector — and the supply chain for high-end electronics is not as clean as you'd like to believe.

2. Initialize the device yourself, generate the seed offline, and write it down on metal — not paper. Paper burns, ink fades, and a coffee spill shouldn't be the reason you lose access to your savings.

3. Test your recovery process with a small balance before you trust the device with your full stack. Seed recovery is a skill, not a checkbox. Practice it when the stakes are low.

Hardware wallets are not a set-and-forget product. They're operational infrastructure. Choose the model that matches your threat model, keep your firmware updated, and remember: the chip is only as good as the verification you've actually done — and the patch you can confirm is installed.