"TVL meaning" is not as simple as it seems: why do the same metrics vary greatly across different projects?

When it comes to DeFi projects, the most common question investors ask is “How’s the TVL?” But this seemingly simple metric actually hides a complex truth. A recent “fake hype” in the Solana ecosystem has thoroughly exposed that the meaning of TVL is far more complicated than we imagined—some TVL data is genuine and reliable, while others are heavily inflated. How can we truly understand the meaning of this indicator to avoid being fooled by false prosperity?

Solana’s “Bubble” Warning: How One Person Faked $7.5 Billion in TVL

The story begins with Solana. Former Solana developer Ian Macalinao created numerous overlapping protocols on Solana as 11 independent developers, wildly inflating Solana’s TVL data. When Solana’s TVL was only $10.5 billion, projects like Saber and Sunny accounted for $7.5 billion, a staggering 71%.

This case perfectly illustrates another side of TVL—its potential for severe distortion. At the time, DeFi data provider DeFi Llama discovered this issue and on August 6th, changed the way chain TVL was calculated, defaulting to exclude double counting caused by overlapping protocols. This reform caused the on-chain TVL to drop significantly but also made the data more accurate.

What Does TVL Really Represent: Hidden Traps in a Static Metric

Total Value Locked (TVL) is a core metric for evaluating projects in DeFi, seemingly straightforward—larger TVL indicates a bigger management of funds. Some investors even divide market cap by TVL to judge whether a project is over- or undervalued. But this simplistic understanding often leads to major mistakes.

The true meaning of TVL is hard to define in one sentence because it is a highly misleading metric. First, it is static data; current TVL figures cannot guarantee they will be maintained in the future. Especially in a highly volatile crypto market, short-term incentives from project teams and token price fluctuations can cause large swings in TVL. Second, in different application layers, TVL can mean entirely different things—some represent real liquidity, some represent staked yield commitments, and some are results of multi-layered double counting.

In DEXs, TVL Means Liquidity; in Lending Protocols, It Means Something Quite Different

To understand what TVL signifies, one must look at specific projects.

In DEXs (Decentralized Exchanges), the most direct meaning of TVL is liquidity. For example, Uniswap has no liquidity mining or staking of LP tokens or UNI tokens, so its TVL directly reflects tradable liquidity. However, many DEXs now incorporate governance token staking mechanisms; Curve and Sushi allow users to stake governance tokens to earn a share of trading fees, and this staked amount can theoretically be included in TVL. DeFi Llama handles this by listing such staking separately under “Staking,” allowing users to distinguish between real liquidity and staked tokens.

In lending protocols, the meaning of TVL varies even more. Compound’s TVL represents the “interest spread”—total deposits minus total borrows, indicating remaining liquidity in the protocol. Aave’s case is more complex; users can stake AAVE tokens and LP tokens to earn inflation rewards, and these staked tokens are also listed separately under “Staking.” MakerDAO’s logic is entirely different: funds borrowed from Maker are in DAI stablecoins issued by the protocol, not consuming the deposited collateral, so Maker’s TVL directly reflects total deposits.

These details show that the same “TVL” can have completely different meanings across protocols.

Why Does Double Counting Occur: Yield Aggregators, Liquidity Staking, Middleware TVL Traps

If TVL means different things in different projects, the situation at the chain level is even more complex—double counting of TVL occurs frequently.

Some projects are built on top of other DeFi protocols, with user funds fully stored in these underlying protocols. The most typical example is yield aggregators. Yearn and Convex Finance (built on Curve) are such projects—they help users automatically farm yields in underlying protocols. Convex, holding large amounts of CRV tokens and staking them, helps users earn higher yields on Curve. Users can exchange CRV for CVXCRV and stake to share rewards. As of August 10th, Convex’s TVL still reached $4.47 billion, ranking 6th. In these projects, TVL is entirely double-counted—funds are included in both Curve’s and Convex’s TVL.

Liquidity staking protocols are also major sources of double counting. Take Lido, with a TVL of $7.75 billion, of which $7.61 billion is ETH locked on Ethereum. But the derivative stETH issued by Lido is widely used in other DeFi projects—about 21.6% of stETH is used as collateral in Aave, and about 14.7% in Curve’s ETH/stETH liquidity pool. These stETH tokens are counted in Lido’s TVL as well as in Aave’s or Curve’s TVL. DeFi Llama has stopped double counting such funds, but this may lead to underestimating total TVL, as some stETH is held on centralized exchanges, with the corresponding ETH still staked on-chain.

Another category is middleware service providers like Instadapp. It offers seamless cross-protocol DeFi operations, simplifying interactions across Aave, Compound, Maker, Uniswap, Liquity, and others. Instadapp even proposed the concept of a “DeFi smart layer” (DSL) to become a foundational infrastructure for DeFi. The key issue is that the funds managed by Instadapp are stored in other protocols, and its TVL should not be double-counted in chain statistics. At its peak, Instadapp’s TVL was about $13.5 billion; currently around $2.6 billion, ranking 10th, but most of this data was previously included in chain TVL figures.

Breaking the TVL Illusion: What Does DeFi Llama’s Reform Mean

TVL data is prone to serious misinterpretation, but that doesn’t mean it is worthless. The key is understanding what TVL truly represents in different scenarios.

At the application layer, TVL is a snapshot of current project data, useful for cross-comparing similar projects. At the chain level, previous double counting inflated chain TVL figures. DeFi Llama’s recent change—eliminating double counting caused by protocol stacking—significantly lowered chain TVL. This may seem like data being “exposed,” but in fact, the cleaned data is what we truly need—more authentic and valuable in terms of reference.

With this reform, the DeFi ecosystem begins to move away from false prosperity and into a more rational evaluation phase. Understanding the multiple meanings of TVL is the prerequisite for using this metric correctly.