What is total value locked (TVL), and what does it miss?

Open almost any decentralised finance (DeFi) dashboard, and the first number on the page is total value locked (TVL). Protocols quote it in announcements and rankings sort by it. It reads like a measure of size, or even of trust. It is narrower than that: a snapshot of what sits in a set of smart contracts, priced in dollars at one moment.

What is total value locked?

Total value locked is the dollar value of the crypto assets deposited in a protocol's smart contracts at a given moment. A tracker reads the token balances those contracts hold from the blockchain, multiplies each by a current market price and adds them up. The same sum across every protocol on a chain gives that chain's TVL.

The word "locked" overstates it. In most lending markets, exchange pools and vaults, depositors can withdraw whenever they choose. TVL counts the money sitting in a protocol right now. It says nothing about how long that money will stay.

Why does TVL change when nobody deposits anything?

TVL moves with prices as much as with deposits. If a pool holds 1,000 ether and the price of ether falls by a fifth, the pool's TVL falls by a fifth even though no one withdrew a single coin.

This makes a rising TVL hard to read on its own. It may reflect new money arriving, or it may reflect the same deposits being repriced in a rally. The two can look identical on a headline chart. One way to tell them apart is to compare the number of tokens deposited rather than their dollar value.

How can the same dollar be counted twice?

The same underlying asset can appear in more than one protocol's TVL. Suppose someone deposits 10 ether in a lending protocol, borrows stablecoins against it and deposits those stablecoins in an exchange pool. Only 10 ether of outside capital entered the system. Both protocols still report a deposit, and a total that adds them together counts the position twice.

Liquid staking adds another layer. A staking service takes ether and issues a receipt token in return, and that receipt can then be deposited elsewhere as collateral. Counted without care, the staked ether and the receipt token standing in for it both show up.

Trackers try to correct for this. DefiLlama is one of the most widely used. Its methodology states that borrowed coins are not counted, and that nothing is counted twice within a single protocol. It also leaves liquid staking out of chain totals by default, and offers a setting that strips out cross-protocol double counting. When it stopped counting some of these positions twice in 2022, several chains' reported TVL fell by more than US$1bn.

Academic work suggests the gap can be large. A 2024 paper titled Piercing the Veil of TVL estimated that at the peak of DeFi activity on 2 December 2021, headline TVL was roughly double the value that could be redeemed out of the system. The difference came to about US$140bn.

What does TVL not tell you?

TVL does not measure revenue, usage or safety. A protocol can hold large deposits that sit idle and earn it little in fees, while a smaller one handles far more activity. For where returns in DeFi come from, see our note on DeFi yield (opens in a new tab).

TVL is also not a security rating. More money in a protocol does not make it safer. A bigger pot gives attackers more to steal and more reason to try. The quality of its code, its admin keys and its outside dependencies say more about risk than the size of the pot. Our note on what a smart contract audit proves (opens in a new tab) covers one of those checks.

TVL can also leave faster than it arrived. Deposits drawn in by incentive rewards may move on once the rewards end. After an exploit, a protocol's TVL can fall within hours as the drained funds and the withdrawals that follow both register at once.

What questions does a TVL figure raise?

A TVL figure works best as the start of three questions. Is the change coming from new deposits or from prices? How much of the total is the same asset counted in several places? Which tracker produced the number, and using which method?

That last question matters more than it looks. A 2025 Bank for International Settlements working paper noted that the way TVL is calculated is not standardised across providers. It proposed a verifiable version built only from on-chain data, so that anyone could reproduce the figure.

Where Northtape fits

Northtape publishes no TVL figures and ranks no protocol by its deposits. It is a news desk. Stories about exploited smart contracts, and drained pools are detected by the Risk Radar protocol lens.

FAQs

Does a higher TVL mean a protocol is safer? No. TVL measures deposits, not security, and a larger pool can make a protocol a bigger target for attackers.

Is the money in TVL actually locked? Usually not. In most lending markets and exchange pools depositors can withdraw at any time, so the name describes what is deposited rather than a lock-up.

Why do different sites show different TVL for the same protocol? Each tracker decides which contracts and tokens to count, how to price them, and how to treat borrowed, or re-deposited assets. Those choices can produce different totals.

Does TVL include borrowed funds? On DefiLlama it does not, so that borrowing and re-depositing in a loop cannot inflate the total. Other trackers may treat borrowing differently, which their methodology pages set out.

None of this is investment advice. It describes how TVL is measured, not whether any protocol, token or deposit is a good or safe place for funds.