The nine ways DeFi takes your money

Smart contract hacks get the headlines. Most losses come from risks you agreed to without measuring.

DeFi in 2026 is no longer a niche. Stablecoins alone have passed 312 billion dollars, and lending, liquid staking and tokenised real-world assets are ordinary infrastructure. That growth brought ordinary users into a system whose risks are real and rarely explained in plain terms.

1. Impermanent loss

Providing liquidity to a pool means holding two assets in a fixed ratio. When their prices diverge, your position ends up worth less than simply holding the two tokens would have been. If ETH doubles while you provide ETH/USDC liquidity, your pool position is worth less than holding — and the trading fees only partly offset it.

It is called impermanent because it reverses if prices return to the original ratio. In a trending market they usually do not. Our impermanent loss panel compares the pool outcome against simply holding, with your fee income included, so you can see which side actually wins.

2. Liquidation and the health factor

Borrowing in DeFi means over-collateralising: a typical loan-to-value ratio is 66–75%, so you deposit $150 to borrow $100. If the collateral falls or the debt grows and your health factor drops below the liquidation threshold — roughly a 120% collateral-to-loan ratio — anyone may liquidate your position to repay the debt, with a penalty on top.

3. Stablecoin and LST depegs

A stablecoin can drift from its target, sometimes gradually and sometimes very fast. Even temporary deviations create real losses if you have to exit, rebalance or repay during them. Curve's 3pool lost over 10% of its value during the USDC depeg of March 2023 before recovering.

The same risk applies to anything meant to track a reference value, including liquid staking and restaking tokens. The depeg watch panel monitors nine major stablecoins live and flags any drifting away from a dollar.

4. Oracle failure

Smart contracts cannot read the outside world; oracles feed prices in. If a protocol relies on bad or delayed data, it triggers incorrect liquidations or misprices assets. Single-source oracles are vulnerable to manipulation — oracle attacks caused 403.2 million dollars of losses in 2022 alone.

5. Counterparty and platform failure

On 18 April 2026, hackers stole an estimated 290 million dollars from major DeFi lending platforms. Aave then experienced mass withdrawals to the point where some lenders, including stablecoin lenders, could not withdraw their funds at all.

That second sentence matters more than the first. Loss of access during stress is one of the most damaging outcomes in this entire category — worse than a paper loss, because it removes your ability to act.

6. Bridge risk

Moving assets across chains adds smart contract and operational risk on top of everything else. Wrapped assets are claims, not the asset itself.

7. Emissions risk

A large share of advertised yield is often token emissions rather than real revenue. Emissions can be cut by a governance vote at any time, and the yield you joined for disappears. Our yield reality panel lets you model what happens if emissions halve.

8. Gas economics on small accounts

A position needing six transactions at a few dollars each can consume the entire annual yield of a small deposit. The yield reality panel calculates how many days you must stay just to repay gas — for many small positions, the answer is longer than the position will exist.

9. Governance and admin keys

If a small group controls upgrade keys or emergency functions, you are trusting that control structure as much as the code. Poor governance design can allow concentrated voting power to push harmful changes through.

A checklist before depositing anything

1. Name the risk you are being paid for. If the yield is high, something is carrying it.
2. Check who can pause withdrawals. Access matters more than APY.
3. Keep part of your stablecoins outside yield products, liquid and reachable.
4. Avoid long lockups and diversify across venues rather than chasing the top rate.
5. Model the downside first — health factor, impermanent loss and gas — before the upside.

None of this argues against using DeFi. It argues for entering with the numbers in front of you rather than the advertised percentage.