State of Crypto Risk: Q1 2026

February 2026

We've now rated 197 protocols across 14 sectors. Here's what the data actually shows — not vibes, not narratives, just what falls out of systematic risk scoring.

The safe stuff got safer

Lending is boring, and that's the point. Aave V3 and Compound both score in the A range. Their liquidation engines have survived every drawdown since 2022 without protocol-level bad debt. The mechanisms are well-understood, the documentation is thorough, and the oracle setups have redundancy baked in. If you're looking for where to park capital with minimal economic risk, overcollateralized lending on battle-tested protocols is still the answer.

Lido is in a similar spot. Liquid staking was risky when it was new — concentrated validator sets, untested withdrawal mechanics, governance capture vectors. Three years later, the category has genuinely matured. Lido's validator set is diversified enough that single-operator risk is no longer a top concern, and the withdrawal queue has been stress-tested in real market conditions.

Synthetic dollars are the new black box

Ethena is the case study here. It scores well on documentation (the team publishes their funding rate data publicly) and scale (large TVL implies deep liquidity), but the interaction severity score is elevated. The issue isn't any single mechanism — it's that basis trades, custody arrangements, and redemption flows interact in ways that only become visible during funding rate inversions. We've seen this movie before with algorithmic stables, and while Ethena's design is fundamentally different, the "everything works until it doesn't" pattern is the same.

More broadly, every synthetic dollar protocol in our coverage universe has a higher average interaction severity score (18.3/25) than any other sector. That's not a knock on the teams — it's a property of the design space. Layering derivatives exposure under a stable peg creates irreducible interaction risk.

Restaking is accumulating tail risk

EigenCloud and the restaking sector sit in a weird spot. The core staking mechanism is straightforward — deposit ETH, earn yield from AVS operators. But the economic dependencies between restaked protocols are creating a web of correlated slashing exposure that nobody has a complete map of.

Here's the concrete concern: when AVS operators run multiple services, a bug in one service's slashing conditions can cascade into slashing across all services that operator validates. The protocols have slashing caps, but those caps assume uncorrelated failures. The whole point of restaking is to correlate the validator set — that's the feature. It's also the risk.

We scored restaking protocols an average of 14.1/20 on mechanism novelty. For context, battle-tested lending scores 4-6.

Oracle surface is actually improving

Good news: across our full coverage, the average oracle surface score has dropped from ~9/15 to ~7/15 over the past quarter. More protocols are implementing multi-oracle fallback (Chainlink + Pyth + TWAP), and circuit breakers are becoming standard. The days of "one Chainlink feed goes stale and the protocol eats $50M in bad debt" are fading. Not gone — we still see single-feed dependencies in smaller protocols — but the trend is clear.

What we're watching next

Cross-chain composition risk. Protocols deploying on 5+ chains inherit the bridge security of each one. Most don't document which bridges they depend on, and none have modeled the compound failure probability. This is the most underpriced risk in crypto right now.

Governance redesigns. Several major protocols are restructuring their token economics under regulatory pressure. New governance mechanisms mean new attack surfaces. We'll be rescanning MakerDAO (now Sky), Uniswap, and Aave governance changes as they ship.

The HR20. We're now publishing the Hindenrank 20 — risk ratings for the 20 largest protocols by TVL. Think of it as a crypto blue-chip risk index. Most of the HR20 grades well, but a few names might surprise you.

All ratings use Hindenrank's six-dimension risk rubric. Lower score = lower risk. See methodology for details, or browse the full leaderboard.