---

name: bitcoin-l2-mezo description: | Mezo: Thesis's Bitcoin-first Cosmos SDK / CometBFT chain with EVM (Evmos fork), BTC for gas, tBTC bridging and a Proof-of-Authority validator set. Mainnet 28 May 2025. MUSD is its Bitcoin-backed CDP stablecoin, a Threshold USD / Liquity fork with a 110% minimum collateral ratio. Chain TVL ~$71.7M on 15 September 2026. USE WHEN: building on Mezo, integrating MUSD, evaluating a major downstream consumer of tBTC. allowed-tools: Read, Grep, Glob

Mezo

Built by Thesis (the Keep / Threshold / tBTC team). Mainnet launched 28 May 2025. The pitch is a "full-stack Bitcoin economy": BTC comes in over tBTC, pays for gas, and backs a native stablecoin.

Per the mezo-org/mezod README, the chain is a Cosmos SDK chain with EVM compatibility running on CometBFT, forked from the LGPL version of Evmos and heavily modified. Stated design goals: user ownership, reliable bridging with tBTC, BTC for gas, a dual staking model, EVM compatibility.

Modules in x/: bridge, evm, feemarket, poa.

Validator set is Proof-of-Authority

This is the single most important thing to know and it is easy to miss. From the x/poa module spec:

• The initial validator set is fixed in genesis.
• New validators submit candidate applications; the module owner approves them.
• The owner can kick a validator out.
• All validators have equal voting power.
• Ownership transfers in a two-step propose/accept flow.
• The module also holds an Emergency Team role controlling chain "lockdown levels".

So "dual staking" on Mezo is about rewards and economic weight, not about permissionless validator entry. Treat Mezo as a permissioned chain until the poa module is replaced.

Bridge

AssetsLocked events observed on the origin side are written into pseudo-transactions: at most one per block, always at index 0, carrying at most 10 events, and rendered by JSON-RPC as if they called the AssetsBridge precompile's bridge method so explorers can show them. Pseudo-transactions are only persisted if a node sets enable-indexer = true in config/app.toml — otherwise they are invisible to JSON-RPC and to explorers. Budget for that when building bridge monitoring.

MUSD

Bitcoin-backed CDP stablecoin. Per mezo-org/musd/docs, MUSD is based on Threshold USD, itself a fork of Liquity — so the trove model, Stability Pool and redemption mechanics are Liquity's.

Parameter	Value
Minimum collateral ratio	110%
Price floor mechanism	redeem MUSD for $1 of BTC (TroveManager.redeemCollateral)
Price ceiling	~$1.10, set by the 110% MCR
Borrowing rate	0.1%, governable
Redemption rate	0.75%, governable
Interest	simple, fixed, non-compounding

Two Liquity divergences worth knowing:

• Interest is per-trove and sticky. A global governable rate applies to new troves; changing it does not touch existing ones. Borrowers may refinance to the current global rate at will (paying a refinancing fee that behaves like the borrowing rate).
• No Stability Pool incentive token. Liquity v1 paid LQTY to Stability Pool depositors; MUSD pays nothing, and the Stability Pool is seeded with a protocol bootstrap loan that can only leave via liquidations. Interest and fees flow to the MUSD Savings Rate vault instead.

Contracts follow Liquity naming: BorrowerOperations.openTrove, ActivePool, StabilityPool, DefaultPool, TroveManager.liquidate / .redeemCollateral.

Scale

From Mezo's own H1 2026 review (fetched 15 September 2026):

Metric	Value	Date the source gives
Total TVL	$110.5M	"as of July 31"
Trading venues	Kraken, Coinbase Exchange, Bitget, KuCoin, Bullish, Aerodrome, "among others"	"as of July 31"
MUSD in circulation	28.0M (27.0M of it on Mezo)	table footnote only
Lifetime MUSD transfer volume	$789M	table footnote only
BTC locked as veBTC	912.6 BTC	table footnote only
Outstanding loans	$13.1M against 826 BTC collateral	table footnote only
System collateralisation	370%	table footnote only

Note the source's self-contradiction: only Total TVL and trading venues are labelled "as of July 31", while the table's own footnote reads "All figures pulled onchain, as of June 31, 2026" — a date that does not exist. Do not restate either as a single clean cutoff; the rest of the table is somewhere around end-June to end-July 2026.

Bullish (NYSE: BLSH) is named as the anchor customer, deploying 1000 BTC through Mezo Prime.

DefiLlama put the Mezo chain at ~$71.7M TVL on 15 September 2026 — a different metric and a later date than the $110.5M above, so do not present them as the same series. Re-check: https://defillama.com/chain/Mezo

The MEZO token and veMEZO were live by H1 2026 (blog posts on MEZO emissions and veMEZO run from April 2026; the H1 review places the token in Q1 2026). Sources disagree on the exact TGE date — do not state one without checking Mezo's own announcement.

Use cases

• Borrow against BTC without selling it, at a fixed simple interest rate.
• A major tBTC destination — DefiLlama had tBTC at ~$336.0M and the Mezo chain at ~$71.7M on 15 September 2026, so tBTC integration work often lands here.
• EVM dApps paying gas in BTC.

Limitations

• PoA validator set — see above. Owner-gated entry, owner-driven removal, plus an Emergency Team with lockdown powers.
• Bridge trust inherits tBTC's threshold-signer assumptions on top of Mezo's own.
• Liquity-derived liquidation risk — 110% MCR is thin. A fast BTC drawdown liquidates troves with little buffer, and the Stability Pool has no yield incentive attracting depositors.
• Bridge events are invisible by default unless the custom indexer is enabled.
• Mezo published an incident report for 2 July 2026 (docs/reports/2026-07-02-leveldb-corruption.md): the testnet archive node crash-looped on LevelDB corruption in the early morning UTC of 2 July 2026, and mainnet validator operators hit the same corruption in the following days. Read it before assuming a clean operational record.

See also

• threshold-tbtc/SKILL.md
• rootstock-rsk/SKILL.md
• bob/SKILL.md
• core-dao/SKILL.md