MPC is the most interesting thing happening in the wallet space, and the hardest to explain without resorting to jargon. Let's try.
The problem MPC solves
Traditional crypto wallets have a single private key. Whoever has the key has the funds. Lose the key, lose the funds. The whole industry is built around how to back up and protect this one number.
MPC, or multi-party computation, says: **what if no single device ever holds the full key?**
The idea
In an MPC wallet, your "key" is split into multiple parts, called **shares**, distributed across devices. To sign a transaction, all the devices (or a threshold of them) must cooperate. At no point during this cooperation is the full key ever assembled in one place — even briefly, even in memory.
The math behind this is gnarly (it uses something called Shamir's Secret Sharing and a lot of clever cryptography), but the practical effect is:
- You can lose your phone. As long as you have the other share(s), you can recover.
- The wallet provider's servers never have the full key. If they get hacked, the attackers get shares, not a key.
- There is no "seed phrase to lose." Recovery works differently.
What this means in practice
Zengo is the most prominent example. When you set up a Zengo account: - Your device generates a share. - Zengo's servers generate another share. - To sign a transaction, both shares must cooperate. - To recover, you prove your identity via your email, your device biometric, and a "recovery file" you stored.
There is no 12-word seed to write down. There is no "if you lose it, you lose everything" moment. This is genuinely the most user-friendly wallet design that has ever shipped at scale.
The trade-offs
MPC is not a free lunch. The trade-offs:
**1. The wallet provider is now a part of your security model.** If Zengo's servers go down, or get compromised, you have a problem. They don't have the full key, but they do have one of the shares. With a hardware wallet, you can operate in a basement with no internet; with an MPC wallet, you can't.
**2. The implementation is closed.** The math is published; the actual code is not. You're trusting the implementation, not just the math. The math says "no single shareholder can sign." The implementation might have a bug.
**3. The "no seed phrase" claim is partially marketing.** You still have something to back up — the recovery file. Lose your device, lose your recovery file, lose your email access, and you're stuck. It's a different failure mode, not a non-failure mode.
**4. It's relatively new.** The MPC ecosystem is 5 years old, not 15. We don't yet have the long tail of incidents and recoveries that we have with seed-based wallets.
When MPC makes sense
- For crypto-curious users who are terrified of seed phrases
- For families where one person holds the device and another can help with recovery
- For small-to-medium amounts where convenience matters more than the maximum possible security
When MPC doesn't make sense
- For long-term storage of large amounts (a hardware wallet is still stronger here)
- For users who want to operate independently of any company
- For users who want a fully open-source stack with reproducible builds
The bigger picture
MPC is one answer to the "how do you keep a key safe" question. Hardware wallets are another. Multi-sig is another. Shamir backup is another. None of them are "the answer." They are different trade-offs, suited to different threat models and different users.
What matters is not which one you pick, but that you pick one **deliberately**, with eyes open about what it does and doesn't protect against.