Unveiling the DeFi TVL Trap: Why Do the Numbers "Lie"?

When a Solana developer uses 11 different identities to layer protocols within the DeFi ecosystem and artificially inflate locked-in value, a widely trusted indicator for years suddenly becomes invalid. The DeFi TVL (Total Value Locked) metric is exposing its biggest weakness—being easily manipulated and misread. In August 2024, the DeFi data tracking platform DeFi Llama changed the calculation rules for chain TVL, defaulting to eliminate double counting between protocols. As a result, the TVL data for multiple chains instantly “shrank.” What exactly happened behind the scenes?

Why DeFi TVL Is Easily Manipulated

In the DeFi world, TVL (Total Value Locked) is the most straightforward indicator of a project’s size. The more funds locked, the more it seems that the project is popular and the ecosystem is healthy. But this logic is often “broken” in practice.

The root of the problem lies in the nature of DeFi TVL—it is a static snapshot, not a true reflection of liquidity. When the same funds move between multiple DeFi protocols, each movement is counted repeatedly in TVL. Imagine depositing $100 into lending platform Aave, which issues you a derivative token aToken, then using that aToken to provide liquidity on a DEX. Now, the same $100 is counted in the TVL of two protocols. Amplify this logic, and when complex structures like yield aggregators, liquidity staking, and cross-protocol combinations appear, the same funds can be counted in 5, 10, or even more protocols’ TVL.

Coupled with factors like price volatility, project incentives, and short-term capital inflows, DeFi TVL can fluctuate far more than expected. Last year, some projects on Solana artificially inflated the chain’s TVL multiple times in a short period through such methods.

Real Liquidity vs. Inflated Data

Not all DeFi protocols have problematic TVL. To understand the real state of DeFi, we must examine different types of protocols separately.

In decentralized exchanges (DEXs), if there are no staking incentives, TVL equals actual liquidity. For example, Uniswap’s TVL directly reflects the amount of funds provided by liquidity providers. However, new-generation DEXs like Curve and Sushi introduced governance token staking mechanisms, allowing users to stake tokens for fee sharing. These staked tokens are often listed separately as “Staking” in DeFi data platforms. In theory, they should be included in TVL, but doing so can artificially inflate liquidity figures.

In lending protocols, the meaning of DeFi TVL varies. Compound’s TVL is the “interest spread,” representing the total deposits minus total loans, indicating the actual available liquidity. MakerDAO’s TVL equals the total deposited amount because the borrowed DAI is a newly issued stablecoin that doesn’t affect the deposited funds. Aave adds token staking to this, further complicating the interpretation of TVL.

The True Meaning of TVL in Different DeFi Protocols

The multi-layered structure of the DeFi ecosystem means that TVL has multiple interpretations. To accurately evaluate a project, one must understand what each protocol’s TVL actually represents.

Yield aggregators are a common source of TVL inflation. Platforms like Yearn Finance and Convex Finance don’t generate liquidity themselves; they simply transfer user funds to underlying protocols (like Curve) for yield farming. When Solana’s TVL was only $10.5 billion, related DEXs and yield aggregators accounted for $7.5 billion, a significant overstatement. Convex Finance’s TVL once reached $4.47 billion, but these funds are essentially liquidity within Curve.

Liquidity staking protocols can cause further distortion. Take Lido as an example: users deposit ETH and receive stETH derivatives, which are then used in Aave as collateral or in Curve pools. Approximately 21.6% of stETH is in Aave, and 14.7% in the ETH/stETH pool on Curve. This means the same assets are counted in Lido, Aave, and Curve’s TVL, creating triple counting. Currently, DeFi Llama has adjusted its calculation rules to exclude staked assets from chain TVL unless they are on-chain protocol assets.

Service applications also contribute to hidden TVL inflation. Instadapp, as a DeFi middleware, offers cross-protocol asset management and flash loans, but the funds are stored in underlying protocols like Aave and Compound. Instadapp’s TVL once reached $13.5 billion, but this is essentially a secondary calculation of the underlying protocol funds.

Beware of DeFi Applications That Cause Double Counting of TVL

Identifying applications that cause TVL inflation is the first step toward rationally assessing the DeFi ecosystem.

Any application built on other DeFi protocols can cause double counting. Yield projects deposit funds into underlying protocols for mining; liquidity staking tokens move across multiple protocols; service applications act as intermediaries. When these applications’ TVL is added to the total chain TVL, it results in multiple layers of double counting.

It’s important to note that this isn’t necessarily a flaw of these applications—they provide real services. The problem arises when using a single TVL metric to compare vastly different DeFi projects. For example, $100 in TVL for a truly liquid DEX versus $100 in a yield aggregator means very different things in practice.

Rationally Viewing DeFi TVL: The Stories Behind the Data

DeFi TVL isn’t valueless; the key is understanding its true meaning in different contexts. At the application level, TVL can be used for horizontal comparison—comparing two DEXs’ TVL reflects liquidity differences; changes in lending protocol TVL can indicate shifts in user confidence. But at the chain level, simply comparing absolute on-chain TVL values can be dangerous.

The recent change in calculation rules by DeFi Llama, which significantly lowered chain TVL figures, is an important correction during the “bubble burst” process. Removing inflated data allows us to see the real scale of funds and project quality in the DeFi ecosystem. Protocols that maintain stable TVL are the core applications worth paying attention to.

Understanding the limitations of DeFi TVL doesn’t negate its value; instead, it encourages more rational evaluation of projects. Data speaks, but only if we ask the right questions and understand the underlying logic.