Key Takeaways
- What it is: The value of assets held in or assigned to protocol smart contracts under a defined methodology.
- How it is built: Adapters query contract balances, positions, or protocol state, map tokens to prices, and apply inclusion and double-counting rules.
- Core expression: TVL = sum of qualifying token quantities × reference prices, with protocol-specific inclusion rules.
- Best use: TVL describes asset value deposited under the methodology; it does not equal revenue, liquidity, solvency, unique capital, or protocol value.
- Main limitation: Token prices move, recursive deposits double count, native tokens inflate TVL, borrowed assets can appear multiple times, and adapter errors occur.
- Practical rule: Do not treat a headline TVL figure as real deposited capital or protocol value until the adapter's inclusion rules, recursive-deposit handling, and native-token weighting are checked.
Who This Guide Is For
Total value locked is the most quoted DeFi statistic and contains two words that are both misleading. Very little of it is locked, and the value is a price multiplied by a quantity where only the quantity is a protocol fact.
Read the formula section if you want to understand why TVL rises and falls without a single deposit or withdrawal. Read the pitfalls section before comparing two protocols or two chains by TVL, because double counting and native-token composition make many such comparisons meaningless.
Educational content. Not individualized financial advice.
What Does Total Value Locked Measure?
TVL measures the market value of assets currently held in a protocol's contracts. It is a quantity of tokens multiplied by their prices, summed across every asset the protocol holds.
The word locked implies commitment that mostly does not exist. In a lending market, depositors can typically withdraw on demand subject to available liquidity. In an automated market maker, liquidity providers can exit at will. Assets genuinely locked for a fixed term, such as some vesting or bonding arrangements, are the exception rather than the rule. TVL is a snapshot of what happens to be deposited right now, not a measure of committed capital.
Plain-language definition
TVL is the current market value of everything sitting in a protocol's smart contracts.
Technical definition
For a protocol with contract set K holding token balances b(k, j) of asset j, TVL is the sum over all k and j of b(k, j) × P(j), where P(j) is the market price of asset j. Token balances are read directly from chain state and are unambiguous. Prices are external, and the contract set is an editorial judgment about which addresses constitute the protocol.
How Is Total Value Locked Constructed?
Only one of the three inputs is a protocol fact. That asymmetry explains most of the disagreement between TVL sources.
- Define the protocol boundary. Which contract addresses count? A protocol with a treasury, a staking contract, a governance escrow, and its core pools can be measured four different ways depending on which are included, and there is no protocol-level answer to this question.
- Read balances from chain state at a stated block. This step is exact.
- Attach prices from an external feed. This step imports every problem of the price source, including thin liquidity on smaller assets.
- Decide on deduplication where deposits are represented by receipt tokens that are themselves deposited elsewhere.
The double-counting problem
Deposit an asset into a lending market and receive a receipt token representing the claim. Deposit that receipt token into a second protocol as collateral. The underlying asset now appears in the first protocol's TVL, and the receipt token appears in the second protocol's TVL. Aggregate across a chain and the same asset has been counted twice while backing one economic position.
Chains with deep composability accumulate long chains of this. Aggregators handle it inconsistently, some deduplicating and some not, and the treatment is usually disclosed only in methodology documentation rather than on the chart. Ecosystem-level TVL figures should be assumed to contain double counting unless the deduplication policy is stated.
Native token composition
A protocol whose TVL consists largely of its own governance token is measuring a reflexive quantity. The token price depends partly on the protocol's perceived success, which is measured partly by TVL, which is computed from that price. Separating native from external assets is a required disclosure, not an advanced refinement.
Formula and Measurement Logic
TVL = Σ (token balance in protocol contracts × token price)
Decompose the change to see what is actually moving. If b is quantity and P is price, then a change in TVL splits into a quantity effect and a price effect. A protocol can lose half its deposits while TVL rises, if prices doubled over the same interval. It can also lose half its TVL with every depositor sitting still.
Quantity-held series = Σ token balances, denominated in native units
Charting balances in native units alongside dollar TVL is the single most useful correction available on this metric, and almost no dashboard does it by default. It answers the question people think TVL is answering: are deposits arriving or leaving?
| Decision | Options | Effect |
|---|---|---|
| Contract boundary | Core pools only, or including treasury and staking | Can change the figure by a large multiple |
| Receipt tokens | Deduplicated or counted | Ecosystem totals inflate without deduplication |
| Native token | Included or reported separately | Determines how reflexive the figure is |
| Borrowed assets | Netted against deposits or not | Lending TVL can double count the same capital |
| Price source | Which feed, and handling of illiquid assets | Thin markets can be marked at unreachable prices |
| Bridged representations | Counted on both chains or once | Cross-chain totals inflate |
How Should Total Value Locked Be Interpreted?
Ask what changed before asking what it means. Every TVL move decomposes into price and quantity, and only the quantity component says anything about user behavior.
| Quantity | Price | Reading |
|---|---|---|
| Rising | Rising | Genuine inflows amplified by the market |
| Flat | Rising | Nothing happened; the chart is a price chart |
| Falling | Rising | Depositors leaving into strength, masked by prices |
| Falling | Falling | Withdrawals compounding the market move |
The row that matters most is the third, because it is the one TVL conceals. Deposits leaving during a rising market is a meaningful signal about a protocol and is invisible on a dollar-denominated chart.
TVL is also not a revenue or a health measure. A protocol earns from activity, not from balances sitting still. A large lending market with low utilization generates little interest income while showing enormous TVL. Protocol fees and revenue answer the economic question that TVL is usually recruited to answer, and they are separate series.
Incentive-driven deposits are the other structural caveat. Capital attracted by token emissions typically leaves when emissions stop, so TVL accumulated during an incentive program measures the program's budget more than the protocol's appeal.
Step-by-Step Workflow
- State the contract boundary: which addresses are counted as the protocol.
- Pull token balances in native units and chart them beside dollar TVL.
- Decompose any move into its price and quantity components before interpreting it.
- Separate native-token TVL from external-asset TVL.
- Check the deduplication policy for receipt tokens before using any ecosystem aggregate.
- Check whether an incentive program is running or has just ended.
- Pull protocol fees and revenue to answer economic questions, rather than inferring them from TVL.
Worked Hypothetical Scenario
A lending protocol reports the following across one quarter.
| Asset | Start balance | End balance | Start price | End price |
|---|---|---|---|---|
| ETH | 120,000 | 78,000 | $2,400 | $4,100 |
| Stablecoin | 310,000,000 | 240,000,000 | $1.00 | $1.00 |
| Native governance token | 40,000,000 | 44,000,000 | $1.80 | $3.10 |
Start TVL: (120,000 × 2,400) + 310,000,000 + (40,000,000 × 1.80) = $288m + $310m + $72m = $670m.
End TVL: (78,000 × 4,100) + 240,000,000 + (44,000,000 × 3.10) = $320m + $240m + $136m = $696m.
TVL rose 4 percent, and every dashboard will report growth. Now read the quantities. ETH deposits fell 35 percent. Stablecoin deposits fell 23 percent. The only balance that grew was the protocol's own governance token, and $136m of the $696m total is now that token, up from $72m of $670m.
External capital left in size. The dollar figure rose because ETH appreciated 71 percent and the native token 72 percent. A reader looking at TVL alone would conclude the protocol gained ground during the quarter. The native-unit balances say the opposite, and the rising share of reflexive native-token value makes the headline figure less reliable at the end of the quarter than at the start.
What Can Make the Interpretation Wrong?
- Price moves read as flows. TVL is quantity times price and most large moves are the price term. Native-unit balances are the corrective.
- Double counting via receipt tokens. One economic position can appear in several protocols' TVL. Ecosystem aggregates inflate unless deduplication is stated.
- Native token composition. A protocol whose TVL is mostly its own token is measuring a reflexive loop where the metric and the price each depend on the other.
- Locked is a misnomer. Most TVL is withdrawable on demand, so it measures current placement rather than commitment.
- Contract boundary disagreement. Including or excluding a treasury or staking contract can change the figure by a large multiple, and there is no protocol-level answer.
- Mercenary capital. Deposits attracted by emissions leave when emissions stop, so incentive-era TVL measures the budget.
- TVL is not revenue. A large idle balance earns nothing. Fees and revenue are separate series and answer the economic question directly.
- Illiquid marks. Thinly traded assets can be valued at prices no seller could actually realize at size.
Cross-Network and Provider Comparison
Comparing TVL across chains compounds every problem above with two more.
Bridged assets are represented on the destination chain by a wrapped token while the original sits in a bridge contract on the source chain. Depending on how each chain's TVL is computed, the same underlying asset can be counted on both sides. Ecosystem-level cross-chain totals are particularly exposed to this, and the correction requires reconciling bridge escrow balances against wrapped supply, which few published figures attempt.
Chain composition also differs in ways that make equal TVL figures non-equivalent. A chain whose TVL is predominantly stablecoins has a figure that barely moves with the market. A chain whose TVL is predominantly volatile native assets has a figure that is largely a leveraged price chart. Both can print the same number and behave completely differently over a quarter.
Fee levels affect what kind of activity a chain hosts, so a cheap chain accumulates many small positions and an expensive one fewer large ones. That changes the concentration of TVL substantially without changing the total, and concentration is what determines how quickly the figure can unwind.
Advanced Analytical Methods
Native-unit balance series
Charting deposits in token units rather than currency removes the price term entirely and answers the flow question directly. This is the most valuable transformation available on TVL and should be the default rather than an advanced technique.
Excluding native tokens
Reporting external-asset TVL separately from native-token TVL breaks the reflexive loop and produces a figure that reflects outside capital's assessment of the protocol.
Net deposits after borrowing
In lending markets, gross deposits count capital that has been borrowed and redeposited. Netting borrowed amounts against deposits gives a figure closer to the genuine outside capital in the system.
Utilization and revenue alongside TVL
TVL with low utilization is idle capital. Pairing the balance figure with utilization and with protocol fees separates capital present from capital productive, which is what most TVL comparisons are really trying to establish.
Concentration measures
The share of TVL held by the largest few depositors determines how fast the figure can unwind. A protocol with high TVL concentrated among a handful of addresses carries a very different risk profile from one with the same total spread widely.
Practical Checklist
- I stated which contract addresses count as the protocol.
- I charted native-unit balances alongside dollar TVL.
- I decomposed the move into price and quantity components.
- I separated native-token TVL from external-asset TVL.
- I checked the receipt-token deduplication policy for any aggregate.
- I checked whether an incentive program was running or ending.
- I used fees and revenue for economic questions rather than TVL.
- I checked depositor concentration.
- I did not describe withdrawable deposits as locked capital.
Frequently Asked Questions
Is TVL actually locked?
Rarely. In most lending markets depositors can withdraw on demand subject to available liquidity, and liquidity providers in automated market makers can exit at will. Assets genuinely locked for a fixed term are the exception. TVL is a snapshot of what is currently deposited, not a measure of committed capital.
Why does TVL change when nobody deposits or withdraws?
Because TVL is token quantity multiplied by token price, and price is the more volatile term. A protocol can lose a third of its deposits and still report rising TVL if prices appreciated over the same interval. Charting balances in native units removes the price term and answers the flow question directly.
What is double counting in TVL?
Depositing an asset into one protocol produces a receipt token representing the claim. Depositing that receipt token into a second protocol makes the same underlying asset appear in both protocols TVL. Ecosystem aggregates should be assumed to contain this unless the deduplication policy is stated.
Why does native token composition matter?
A protocol whose TVL consists largely of its own governance token is measuring a reflexive quantity: the token price depends partly on perceived success, which is measured partly by TVL, which is computed from that price. Reporting external-asset TVL separately breaks the loop.
Is TVL a measure of protocol revenue?
No. Protocols earn from activity, not from balances sitting still. A large lending market with low utilization generates little interest income while showing enormous TVL. Protocol fees and revenue are separate series and answer the economic question directly.
How do incentive programs affect TVL?
Capital attracted by token emissions typically leaves when emissions stop, so TVL accumulated during an incentive program measures the program budget more than the protocol appeal. Checking whether a program is running or has just ended should precede any interpretation of a TVL trend.
Can TVL be compared across chains?
Only with care. Bridged assets can be counted on both the source and destination chain, and composition differs sharply: a chain whose TVL is mostly stablecoins barely moves with the market, while one dominated by volatile native assets is largely a leveraged price chart. Equal figures can behave completely differently.
What is the difference between total value locked and assets under management?
Assets under management describes assets a manager holds on behalf of clients under a mandate, with a legal relationship and a reporting framework behind the figure. Total value locked describes tokens sitting in contracts, with no manager, no mandate and no obligation to anyone. Deposits can usually be withdrawn at will and the protocol has no discretion over them. Presenting the two as comparable implies an equivalence in oversight and stability that the underlying arrangements do not share.
Does total value locked capture leverage inside a protocol?
It captures the effects of leverage without identifying it, which inflates the figure. A user can deposit collateral, borrow against it, deposit the borrowed asset, and repeat, so one unit of original capital registers several times in the protocol's deposit totals. The recursive position is real and the deposits exist, but the underlying capital does not. Protocols and dashboards that report deposits net of borrowings give a figure closer to the capital actually committed.
References
These sources should be reviewed during editorial verification. They support data structures and methods, not the hypothetical conclusion. Provider formulas, chain rules, and APIs can change; confirm current documentation before publication.
- Bitcoin Developer Guide: Block Chain: Bitcoin ledger, blocks, proof of work, and transaction history.
- Bitcoin Developer Guide: P2P Network: Full nodes, propagation, and verification.
- Coin Metrics: Hash Rate: Hash-rate methodology.
- Ethereum.org: Proof of Stake: Validator roles, staking, and consensus.
- Ethereum.org: Proof-of-Stake Rewards and Penalties: Validator duties, rewards, and penalties.
- Ethereum.org: Maximal Extractable Value: MEV definitions and transaction ordering.
- DefiLlama: Methodology: TVL, fees, revenue, and volume methods.
- DefiLlama: Data Definitions: TVL, fees, revenue, and holder revenue distinctions.