Most of the largest DeFi protocols can meet their obligations. Very few are growing, and one category sits almost entirely below investment grade. This report sets out what bond-style grades say about 51 protocols, what changed between our July and November screens, and what our own ratings got wrong.
Each protocol gets three separate grades on the AAA-to-D scale used in bond markets. Solvency is the credit grade: can the protocol survive stress and pay its users back. Legitimacy grades whether it is what it claims to be. Growth is an outlook on adoption. The three are never averaged.
50 of the 51 are screen grades: the methodology applied in reduced form, with 11 inputs per protocol, so the whole table stays comparable. They are provisional. One, Aave, is a full rating, refreshed November 13, 2026.
TVL, 12-month TVL change and 90-day drawdown come from DefiLlama, parent-level listing, as of November 12, 2026. For 11 protocols every other input is now listed with its source or marked as an analyst estimate (checked November 13, 2026). For the rest, the remaining inputs are analyst estimates from July 27, 2026. No protocol pays for a rating.
The distribution is centred on A and BBB. The highest grade awarded is AA and the lowest is B.
| Solvency Grade | Protocols | Names |
|---|---|---|
| AA | 1 | Uniswap |
| A | 13 | Lido, Sky (MakerDAO), Morpho, ether.fi, Ethena, Hyperliquid, JustLend, Spark, Jupiter, Maple, PancakeSwap, Venus, Aave |
| BBB | 17 | Pendle, Jito, Curve, Kamino, Raydium, Rocket Pool, Marinade, Compound, Convex, Sanctum, Aerodrome, mETH Protocol, Frax, GMX, Liquity, StakeWise, Ondo |
| BB | 13 | Lombard, Kelp DAO, Drift, Renzo, Fluid, Balancer, Suilend, dYdX, Euler, Benqi, Yearn, Moonwell, SushiSwap |
| B | 7 | EigenLayer, Babylon, Symbiotic, Solv, Puffer, Bedrock, NAVI |
Blue bars are investment grade, gold bars are below it. Legitimacy is stronger still: 43 of 51 are investment grade on that axis.
Restaking protocols share the same problem in the data: deposits have left. Every protocol in the table below holds less TVL than a year ago, and most have lost more than half.
| Protocol | Category | Solvency | TVL | 12-Month TVL Change | Worst 90-Day Drawdown |
|---|---|---|---|---|---|
| ether.fi | LRT | A 70 | $4.3B | -49% | 24% |
| Kelp DAO | LRT | BB 52 | $0.75B | -40% | 36% |
| Renzo | LRT | BB 45 | $0.09B | -92% | 30% |
| Bedrock | LRT | B 41.5 | $0.3B | -52% | 31% |
| Symbiotic | Restaking | B 39.8 | $0.27B | -69% | 44% |
| Puffer | LRT | B 38.3 | $0.02B | -85% | 67% |
| EigenLayer | Restaking | B 37.5 | $5.2B | -63% | 28% |
Kelp DAO's risk score is under review: it does not yet reflect the April 2026 bridge exploit described below. A change there can only lower its grade.
46 of 50 screened protocols have a lower TVL than twelve months ago, and 26 have lost half or more. Growth grades reflect that: 47 of 51 are below investment grade on Growth.
| Growth Grade | Protocols | Names |
|---|---|---|
| AAA | 1 | Hyperliquid |
| A | 2 | Morpho, Ondo |
| BBB | 1 | JustLend |
| BB | 15 | Lido, Sky (MakerDAO), EigenLayer, ether.fi, Ethena, Pendle, Babylon, Uniswap, Spark, Jupiter, Maple, PancakeSwap, Lombard, Sanctum, Aave |
| B | 14 | Jito, Curve, Kamino, Raydium, Marinade, Compound, Kelp DAO, Aerodrome, Solv, Drift, mETH Protocol, Fluid, GMX, Euler |
| CCC | 18 | Rocket Pool, Venus, Symbiotic, Convex, Renzo, Balancer, Frax, Suilend, Puffer, Bedrock, Liquity, dYdX, Benqi, NAVI, Yearn, StakeWise, Moonwell, SushiSwap |
A weak Growth grade is not a statement about safety. 27 protocols are investment grade on Solvency while below it on Growth. A lender and a growth investor are asking different questions of the same protocol, which is why the grades are published separately.
| Category | Rated | Investment Grade On Solvency | Strongest | Weakest |
|---|---|---|---|---|
| Lending | 12 | 7 of 12 | Spark A | NAVI B |
| Liquid Staking | 7 | 7 of 7 | Lido A | Jito BBB |
| DEX | 7 | 5 of 7 | Uniswap AA | Balancer BB |
| LRT | 5 | 1 of 5 | ether.fi A | Puffer B |
| Derivatives | 4 | 2 of 4 | Hyperliquid A | Drift BB |
| Yield | 3 | 2 of 3 | Pendle BBB | Yearn BB |
| CDP | 2 | 2 of 2 | Sky (MakerDAO) A | Liquity BBB |
| Restaking | 2 | 0 of 2 | Symbiotic B | EigenLayer B |
Categories with a single rated protocol: Ethena (Yield / Synthetic Dollar, A); Babylon (BTC Staking, B); Jupiter (DEX Aggregator, A); Maple (Credit, A); Lombard (BTC LST, BB); Solv (BTC Yield, B); Fluid (Lending / DEX, BB); Frax (CDP / LSD, BBB); Ondo (RWA, BBB).
15 Solvency grades fell and 6 rose. Two things drove the downgrades. TVL kept falling, and Methodology v1.1 now scores TVL stability on the worse of the 90-day drawdown and the 12-month decline. In every downgrade below, the 12-month decline is the larger of the two.
| Downgraded | Category | Solvency | 12-Month TVL Change | 90-Day Drawdown | Drawdown Scored |
|---|---|---|---|---|---|
| Pendle | Yield | A → BBB | -82% | 25% | 82% |
| Drift | Derivatives | BBB → BB | -79% | 33% | 79% |
| Euler | Lending | BBB → BB | -76% | 36% | 76% |
| NAVI | Lending | BB → B | -75% | 43% | 75% |
| Yearn | Yield | A → BB | -72% | 33% | 72% |
| Solv | BTC Yield | BB → B | -66% | 18% | 66% |
| EigenLayer | Restaking | BB → B | -63% | 28% | 63% |
| Fluid | Lending / DEX | A → BB | -63% | 21% | 63% |
| dYdX | Derivatives | BBB → BB | -62% | 34% | 62% |
| Benqi | Lending | BBB → BB | -55% | 28% | 55% |
| Convex | Yield | A → BBB | -54% | 23% | 54% |
| Rocket Pool | Liquid Staking | A → BBB | -53% | 32% | 53% |
| Liquity | CDP | A → BBB | -45% | 27% | 45% |
| Compound | Lending | A → BBB | -42% | 24% | 42% |
| mETH Protocol | Liquid Staking | A → BBB | -42% | 24% | 42% |
| Upgraded | Category | Solvency | 12-Month TVL Change | 90-Day Drawdown | Drawdown Scored |
|---|---|---|---|---|---|
| Morpho | Lending | BBB → A | +55% | 8% | 8% |
| ether.fi | LRT | BBB → A | -49% | 24% | 49% |
| Ethena | Yield / Synthetic Dollar | BBB → A | -39% | 9% | 39% |
| Jupiter | DEX Aggregator | BBB → A | -30% | 25% | 30% |
| Sanctum | Liquid Staking | BB → BBB | -13% | 27% | 27% |
| Renzo | LRT | B → BB | -92% | 30% | 92% |
Each protocol page shows its July 27, 2026 and November 12, 2026 inputs side by side. Some upgrades also reflect inputs that were replaced with sourced figures on November 13, 2026; the page for each says which.
A ratings publisher should show its errors. These are ours so far.
How to cite.
Found an error? Tell us. Corrections are published with a date.
