Hardware Security Keys: What They Are and Who Needs One

A small USB or NFC device that makes phishing impossible for the accounts that support it. How it works, how to set it up with a spare, and whether you need one.

What it is

A hardware security key is a small device, usually USB-A, USB-C or NFC, that stores cryptographic keys in a chip designed to never reveal them. When a website asks you to log in, the browser passes a challenge to the key, you touch it to prove a human is present, and it returns a signature. The site verifies the signature with the public key it stored when you registered. Your password, if the site still has one, becomes the less important factor.

The standards are FIDO2 (the current one, made of WebAuthn in the browser and CTAP2 between browser and key) and its predecessor U2F. Passkeys use the same cryptography; a security key is simply a passkey that lives in a dedicated device instead of your phone. See what are passkeys for that side of the story.

Why it beats every other second factor

The signature includes the domain of the site as the browser sees it. On a phishing page at a look-alike domain, the key either has no credential for that domain or produces a signature the real site rejects. There is no code to read out, no prompt to approve by mistake, nothing an attacker can relay. Google made keys mandatory for its staff in 2017 and reported no successful phishing of employee accounts since. That is the whole pitch, and it holds up.

Compare with the alternatives in our 2FA guide: SMS can be intercepted or SIM-swapped, authenticator codes can be phished in real time, push prompts can be approved by fatigue. A key cannot be fooled by a website, only by physical theft plus knowledge of your PIN.

Modes you will see

  • FIDO2 / WebAuthn: the passwordless and second-factor mode used by Google, Microsoft, Apple, GitHub, most password managers and a growing list of banks.
  • U2F: older second-factor-only mode, still accepted by many sites; any FIDO2 key supports it.
  • OTP: some keys can also emit one-time codes (Yubico OTP, or stored TOTP secrets), useful for services that only support codes. The TOTP generator shows how those codes are computed.
  • PIV and OpenPGP: smart-card modes for SSH, code signing and disk encryption. For administrators, this is where a key pays for itself twice.

Setting one up properly

  1. Buy two. Losing your only key with no backup can mean losing accounts. Register both on every service; keep the spare at home or in a safe.
  2. Set a PIN on the key (required for FIDO2 passwordless). It protects against use by someone who steals the device.
  3. Start with email and your password manager, then financial and work accounts.
  4. Remove weaker factors once the key works: delete SMS as a recovery option where the site allows it.
  5. Store the account’s backup codes in your password manager anyway. They are the last resort.

Who needs one

System administrators, developers with production access, finance staff, journalists, activists, executives and anyone with a public profile that attracts targeted attacks: yes, and your organisation should pay for them. Everyone else gets most of the benefit from platform passkeys, which are free and already in your phone. A key is still a good idea for the two or three accounts whose loss would be catastrophic, and at the price of a takeaway meal it is cheap insurance.

What it does not do

A key protects the login. It does not protect a session that is already open, does not stop malware on your machine from acting as you, and does not help on sites that do not support it. Keep the operating system updated, keep passwords unique with the generator and a manager, and treat the key as the strongest link in a chain that still has other links.

New tools and guides, about once a month.

No tracking pixels, no upsells. Unsubscribe with one click.

Newsletter