Live
Moonwell Brief · Joel ObafemiIssue Q-№001 · March 2026

Moonwell Brief

The Contraction Quarter

Moonwell entered 2026 at $172.9 million supplied and exited Q1 at $102.9 million, a 40 percent decline driven by a February oracle incident and deleveraging that left the December cohort holding under half its dollars and $6.4 million of bad debt stuck on the books.

BaseOptimismMoonbeam
14 min readCite this issue

§ 01The quarter in one paragraph

1 2026 was a contraction. Moonwell opened the year at $172.9 million supplied and closed March at $102.9 million, a 40 percent decline. The break came in February, when a cbETH oracle incident and broad deleveraging drove 3,627 liquidations in one month and cut combined supply by more than a third. March stabilized the bleed, but the damage was structural: of the suppliers present at the December close, only 76 percent of wallets and 48.5 percent of their dollars remained at March 31, and bad debt rose across the quarter from $4.3 million to $6.4 million, lodged in illiquid markets. The quarter also retired Moonriver, Moonwell's 2021 origin, and set up Ethereum, approved at quarter's edge.

§ 02KPI dashboard

Three-month trajectory, Q1 2026 (constant-price)
MetricJan 31Feb 28Mar 31Quarter
Combined supplied$172.9M$110.7M$102.9M−40.5%
Core lending$122.1M$77.8M$74.0M−39.4%
Vaults TVL$50.8M$32.9M$29.0M−42.9%
Active suppliers19,53617,99417,975−1,561
Bad debt (shortfall)$4.3M$6.3M$6.4M+$2.1M
Protocol revenue$103K$275K$53K$432K total
Take rate (annualized)0.70%3.24%0.60%1.70% qtr

The decline was front-loaded into February. January and March bracketed it: a $122 million Core base at the start, $74 million at the end, with nearly the entire step-down in the middle month.

§ 03Multi-month chart pack

Supply trajectory by component, Q1 2026
Core lendingVaults
$0$55.6M$111.2M$166.9M$222.5MDecJanFebMar
Core lending and vaults stack to combined supply, which fell from $205.9M at the December baseline to $102.9M by March. Core found a floor near $74M; the vaults kept bleeding. Hover any month for the split.
Source: GraphQL indexer + on-chain capture (oracle-priced, constant-price)
Wallets hold while capital thins: retention, Q1 2026
Wallet retentionDollar retention (NRR)
0%24.03%48.06%72.09%96.12%DecJanFebMar
Wallet retention never left the high 80s even as dollar retention collapsed to 63.9 percent in February. The wallets stayed; the dollars deleveraged, then began returning in March.
Source: GraphQL indexer (constant-price cohort, wallets ≥ $10 supplied)

Month-over-month dollar retention tells the recovery story the levels obscure: 84.8 percent in January, 63.9 percent in February at the depth, then 87.5 percent in March. The cohort kept leaving across the quarter, but by March those who remained had largely stopped.

Cohort retention triangle, Q1 2026
Colour byeach cell: W = wallets · $ = dollars
CohortM0M1M2M3
Dec 2025
20,028 wallets
100%
W 100%
84.8%
W 85.6%
56.5%
W 76.9%
48.5%
W 73%
Jan 2026
19,536 wallets
100%
W 100%
63.9%
W 83.8%
56.2%
W 79.1%
Feb 2026
17,994 wallets
100%
W 100%
87.4%
W 88.3%
Mar 2026
17,975 wallets
100%
W 100%
Each row is a monthly cohort (the active deposit base of wallets ≥ $10 at that month-end); columns are months since. Each cell shows dollar retention (NRR, large) and wallet retention (W). The December base entered the quarter and kept only 48.5 percent of its dollars by March, the deleveraging written as a decay curve. Read down a column to compare cohorts at the same age; the damage is concentrated in the February step, a period effect every cohort shares, not a quality difference between them.
Source: Envio GraphQL indexer: per-wallet SupplyDelta + RateSample, oracle-priced, constant-price, $10 floor

§ 04Governance review

Q1 was the busiest governance quarter of the period: 22 proposals, 16 executed, 5 cancelled, 1 queued. The agenda tracked the quarter's arc, from legacy wind-down in January to incident response in February to recovery in March.

Material executed proposals, Q1 2026
ProposalTitleMonth
MIP-R38Moonriver Complete DeprecationJan
MIP-B56Transfer Vault Risk Curatorship to Anthias LabsJan
MIP-X43Activate OEV Wrappers for Non-Composite MarketsFeb
MIP-B57Revert cbETH Market Oracle on BaseFeb
MIP-B58MFAM Onboarding and cbETH Incident RestitutionMar (queued)
MIP-B59Add VVV Market to Moonwell on BaseMar

The material parameter changes were the cbETH oracle revert (MIP-B57) and its restitution follow-up (MIP-B58), the vault curatorship transfer to Anthias Labs (MIP-B56), additional oracle-value wrappers (MIP-X43), and the only new listing, VVV on Base (MIP-B59). The quarter's sharpest dissent was a cancelled Moonriver admin-migration proposal that drew 94 million WELL against.

§ 05Per-chain quarterly review

Base was the quarter, and the contraction was Base's: Base Core fell from roughly $116 million to $68.6 million, and all five vaults fell with it. Optimism was the quiet exception, ending at $3.9 million Core with utilization climbing to 48 percent, the only chain to add supply in the final month. Moonbeam held near $1.5 million throughout. Moonriver was deprecated in January and left the active set.

§ 06Per-market quarterly deep dives

The large markets diverged sharply. USDC fell from $34.0 million to $18.7 million even as it stayed the most-borrowed market. cbBTC fell hardest among the majors, from $22.7 million to $8.2 million. The standout the other way was MORPHO, which grew every month, from $6.6 million to $8.7 million. cbETH was the quarter's problem market: its oracle was reverted in February after producing distorted rates, and it ended the quarter carrying $1.7 million of the protocol's bad debt.

§ 07Financials and risk

The quarter generated $1.62 million of gross fees and $431,600 of protocol revenue, a 1.70 percent annualized take rate. Both are dominated by February's liquidation wave: $796,000 of the gross fees were liquidation bonuses, and most of the revenue was the 3 percent protocol seize share of liquidated collateral, not recurring interest, so the underlying monetization is far thinner than the headline take rate suggests. Cumulatively through March, Moonwell has produced roughly $4.0 million of gross fees and $539,000 of revenue since inception.

The risk story is the one that should worry a delegate most. Bad debt did not fall as the protocol stabilized; it rose, from $4.3 million in January to $6.4 million by March, and stayed concentrated in three illiquid markets, cbETH ($1.7 million), cbXRP ($1.7 million), and VIRTUAL ($1.4 million), where underwater positions cannot be liquidated profitably. The February liquidation wave cleared the performing collateral and left the bad debt behind.

Standing bad debt by market, Q1 2026
cbXRPVIRTUALcbETHOther markets
$0$1.73M$3.46M$5.18M$6.91MDecJanFebMar
The total rose from $4.6M to $6.4M. cbXRP and VIRTUAL carried over from the October incident; cbETH appeared in February after its oracle was mispriced, and the composition shifted toward it.
Source: On-chain comptroller.getAccountLiquidity at each month-end block
Liquidations by collateral asset, Q1 2026
wstETHcbBTCAEROWETHcbXRPOther
$0$1.92M$3.84M$5.76M$7.68MJanFebMar
Collateral seized in liquidations, by asset. February's deleveraging ($7.1M seized across 3,627 events) concentrated in wstETH and cbBTC; by March it had collapsed to roughly $42K. Hover any month for the per-asset split.
Source: GraphQL indexer (Liquidate events, collateral seized, all chains)
Annualized take rate by month, Q1 2026
0%0.87%1.75%2.62%3.5%JanFebMar
The annualized take rate spiked to a one-off 3.24 percent in February on the liquidation seize share, then normalized. Stripping the liquidation revenue, recurring monetization is far thinner than the quarter's 1.70 percent headline.
Source: daily_financials (protocol revenue ÷ fees, annualized)

Across the quarter, Anthias Labs, Moonwell's Base risk partner, worked the bad debt from both ends. End-of-December recommendations raised reserve factors on the incident-affected markets (AERO, VIRTUAL, and cbXRP to 35 percent) to rebuild reserves, the January risk assessment flagged several markets whose caps no longer functioned as effective controls, and after February's cbETH oracle incident the March recommendations added a cbETH interest-rate package, lifting its reserve factor to 20 percent and steepening the curve, explicitly to channel interest toward repaying that shortfall.

§ 08Cross-deployment and token

Across its own deployments, Q1 left Moonwell more concentrated on Base than it began: with Moonriver retired, more than 94 percent of combined supply sat on Base at quarter-end. Ranked against the other Base lending venues at quarter-end, Moonwell was the third-largest by core lending supply, behind Morpho Blue and Aave V3 and ahead of Compound V3, Fluid, Seamless, and Euler. Its $68.6 million of Base core lending kept it an order of magnitude below the two leaders but firmly in the next tier, holding its niche through a quarter that began reshaping the Base lending market around it. The Moonwell vaults are Morpho Blue deposits, counted under Morpho's total, so they are excluded from the ranking to avoid double-counting.

Base lending venues by supplied USD, March 31, 2026
$0$905.6M$1.81B$2.72B$3.62BMorpho BlueAave V3MoonwellCompound V3FluidSeamlessEuler V2
Moonwell (highlighted) ranks third on Base by supplied USD. The linear scale shows the gap honestly: Morpho Blue and Aave V3 are an order of magnitude larger, but Moonwell sits clear of the next tier.
Source: DefiLlama Base-chain TVL + borrowed; Moonwell row is its own oracle-priced Base core
Base lending venues by supplied USD, March 31, 2026
RankVenueSupplied
1Morpho Blue$3.35 billion
2Aave V3$1.18 billion
3Moonwell$68.6 million
4Compound V3$41.0 million
5Fluid$39.8 million
6Seamless$5.5 million
7Euler V2$4.6 million

WELL fell 21 percent across the quarter, from $0.005374 to $0.004228, a market cap of $19.1 million at the close. Staked stkWELL rose from 23 to 26.2 percent of circulating supply, the governance base deepening even as token price and TVL fell.

§ 09Open questions the data raised

Three questions carry into Q2. First, the bad debt: $6.4 million stuck in illiquid markets that the deleveraging could not clear, so whether it ever resolves or must be written off. Second, the vaults: down 43 percent across the quarter and falling every month even after Core stabilized, so what stops the leak. Third, monetization: a 1.70 percent take rate that is mostly liquidation-driven, so whether the protocol can earn a durable spread from lending itself.

§ 10Looking ahead

Q2 opens with Core lending stabilized, the vaults still leaking, the bad debt unresolved, and Ethereum freshly launched. The questions for the next quarterly are whether Moonwell returns to net growth or settles into a smaller, more concentrated steady state, whether the new Ethereum deployment attracts durable liquidity, and whether the cbETH restitution and the MIP-X56 OEV fee-split change close the loops Q1 left open.

Methodology

This quarterly composes the three monthly snapshots of Q1 2026. Levels are constant-price, read from the Moonwell oracle at each month-end block; Core lending is on-chain, vaults, risk, and wallet-level metrics from the protocol indexer and on-chain comptroller reads. Financials are the dashboard ledger; comparators are DefiLlama Base-scoped (net Base TVL plus Base borrowed), with Moonwell shown as its own oracle-priced Base core lending and the vaults excluded as Morpho Blue deposits. The quarter-level retention figure (76 percent wallet, 48.5 percent dollar) measures the December cohort against March 31; the month-over-month figures in §03 are separate. Liquidations, governance, OEV, and fees are quarter-window totals.
Found any inaccurate data?DM on X|Email