Deciphering TVL: Why do the same data have completely different meanings across various DeFi projects?

In August 2021, a wave of “false prosperity” regarding the Solana ecosystem swept through the entire crypto world. Former Solana developer Ian Macalinao created multiple overlapping DeFi protocols on Solana under 11 different identities, artificially inflating the blockchain’s TVL data. Subsequently, DeFi data provider Defi Llama made a bold decision to modify the way it calculates total value locked (TVL) on public chains, removing duplicate data caused by protocol stacking. This event also exposed a fundamental issue: many people do not truly understand what TVL actually means, let alone why it can have completely different implications across various projects.

What Does TVL Really Mean? Why Is It Misunderstood?

Total Value Locked (TVL) is the most frequently cited metric in the DeFi ecosystem. On the surface, the definition of TVL is simple—it’s the total amount of assets deposited in a specific protocol. Based on this number, investors can quickly assess a project’s scale and determine whether its valuation is reasonable. Dividing the total market cap by TVL can give an initial estimate of overvaluation or undervaluation.

But this is where the trap lies. TVL means much more than a static asset quantity; it also involves multiple dimensions such as the source, nature, and flow of assets.

First, TVL is a highly misleading static indicator. Current data cannot guarantee future stability, especially in the highly volatile crypto markets. Short-term incentives launched by project teams to boost TVL, sharp price fluctuations, and shifts in user sentiment—all these factors can cause TVL to change dramatically in a short period. Seeing a high TVL today, it might plummet tomorrow due to the end of incentives or market downturns.

Second, the meaning of TVL differs fundamentally at the application layer and the public chain layer. At the application level, TVL represents the current managed funds of a project and has some comparative value across projects. But at the public chain level, stacking of different protocols can lead to inflated TVL—funds may be counted multiple times, causing the overall TVL data of the chain to balloon. This is also why the recent change by Defi Llama caused many chains’ TVL to drop significantly.

TVL in DEXes and Lending Protocols: A True Reflection of Liquidity

To truly understand what TVL means, one must analyze different types of DeFi protocols.

In decentralized exchanges (DEX), the most direct meaning of TVL is the amount of funds in liquidity pools. Take Uniswap as an example: it does not offer mining incentives, and users do not need to stake tokens to trade. Therefore, Uniswap’s TVL equals its liquidity size, which is the purest and most genuine TVL data.

However, not all DEXes are so “honest.” Protocols like Curve and Sushi allow users to stake governance tokens (CRV and SUSHI) to earn a share of trading fees. These staked governance tokens should theoretically be included in TVL, but Defi Llama smartly separates this data into a “Staking” category to avoid confusion. If you want to understand the true liquidity of a DEX, removing the Staking portion from TVL provides the most accurate picture.

The situation is even more complex for lending protocols. Compound’s TVL refers to the “deposit minus borrow” amount—i.e., the total deposits minus total loans—representing the actual available liquidity in the protocol. This figure, along with total deposits and total loans, is also worth noting because it directly reflects the health of the protocol.

Aave extends this concept by allowing users to stake AAVE tokens and LP tokens to earn inflation rewards. These staked funds are also listed separately under “Staking.” MakerDAO is different—users borrow DAI stablecoins issued by the protocol itself, which does not involve fund flow issues. Therefore, Maker’s TVL directly equals the total deposits.

When excluding the Staking portion, individual DEXes and lending protocols generally do not have issues with double counting of TVL. This is because the internal fund flows within a single application are clear and unidirectional. Double counting only occurs when different protocols are combined—for example, Aave using Uniswap LP tokens as collateral, or providing liquidity with Aave-derived tokens (aTokens) in DEXes—leading to the same funds being counted multiple times at the chain level.

Overcounted TVL: The Traps of Yield Aggregators and Liquidity Staking

The real culprits of double counting are “layered applications” built on other DeFi projects.

Yield aggregator protocols are among the most common sources of double counting. Projects like Yearn and Convex Finance operate by depositing user funds into underlying protocols for yield farming. When these funds are counted twice—once in the underlying protocol (e.g., Curve) and once in the aggregator (e.g., Convex)—the chain’s TVL becomes significantly inflated.

For example, Convex Finance is built on Curve, helping users earn higher yields by holding and staking CRV tokens. CRV holders can exchange CRV for CVXCRV, stake it, and share in Convex’s rewards—indirectly locking more tokens in Curve. As of that time, Convex’s TVL reached $4.47 billion, ranking 6th, but the actual source of these funds was in Curve.

Ian Macalinao’s operations on Solana were similar. He created Sunny, a yield aggregator based on the Saber DEX, when Solana’s total TVL was only about $10.5 billion, with Saber and Sunny accounting for $7.5 billion—highlighting the severe double counting issue.

Liquidity staking protocols are also major sources of double counting. These applications issue derivatives that allow users to earn PoS mining rewards while maintaining token liquidity. Take Lido as an example: users deposit ETH, and Lido issues an equivalent stETH derivative. Lido’s TVL is about $7.75 billion, with approximately $7.61 billion of ETH locked, but stETH derivatives are widely used across DeFi.

According to blockchain data, about 21.6% of stETH is used as collateral in Aave, and 14.7% is in Curve’s ETH/stETH liquidity pool. This means the same ETH is counted multiple times—appearing in Lido’s TVL, as well as in Aave and Curve’s TVL.

After the reform, Defi Llama no longer includes funds from liquidity staking protocols in the chain’s TVL; stETH is only counted when used in protocols on other chains. Although this change may omit some data (such as stETH stored on centralized exchanges), it makes the chain’s TVL data more accurate overall.

The Dilemma of Service Layer Protocols’ TVL

Another category includes middleware service protocols that manage funds stored in underlying protocols. Instadapp is a typical example.

Instadapp is essentially a “DeFi gateway” that helps users seamlessly switch between protocols like Aave, Compound, Maker, Uniswap, and Liquity. It simplifies complex DeFi operations, offering flash loans, leverage adjustments, and auto-refinancing. The funds managed through Instadapp are actually stored in these underlying protocols, with Instadapp serving as a front end.

At its peak, Instadapp’s TVL was about $13.5 billion, later dropping to around $2.6 billion. Since all funds managed by Instadapp are stored in other protocols, its TVL should not be counted again in the chain’s total; otherwise, it results in double counting.

The Importance of Correctly Understanding TVL

TVL data can be easily misinterpreted, but it is not without value. The key is to understand what TVL truly represents in different scenarios.

At the application level, TVL indicates the current managed fund size and can be used for cross-project comparisons. At the chain level, there was previously a lot of double counting caused by protocol stacking, leading to severely distorted chain TVL figures. Defi Llama’s reform significantly reduced the chain’s TVL, which may seem less “impressive,” but as the bubble bursts, more accurate data is more valuable for assessing the overall health of the DeFi ecosystem.

Distinguishing “false TVL” from “real TVL” and understanding what TVL means in different contexts can help avoid being deceived by flashy numbers and enable smarter investment decisions.

Statement: The content of this article is for reference only and does not constitute investment advice. Investors should make their own decisions and are not responsible for investment losses.