Understanding what TVL is: Deciphering the bubbles and realities in the DeFi ecosystem

In August 2024, a wave of “TVL Faking” controversy shook the DeFi community. Ian Macalinao, a former Solana developer, operating as 11 independent developers, stacked multiple protocols on Solana to artificially inflate on-chain TVL, thoroughly exposing the complex issues behind this seemingly simple metric. DeFi Llama subsequently revised its method of calculating blockchain TVL, defaulting to exclude double counting, sparking a industry-wide rethinking of “what does TVL really represent.”

What is TVL: An Easily Misread Data Point

Total Value Locked (TVL) is widely used in DeFi to measure the amount of funds locked within a project or blockchain. On the surface, TVL is straightforward—larger TVL indicates a bigger management of funds, making it an intuitive way to assess project size. Investors often use the “Market Cap / TVL” ratio to quickly determine if a project is undervalued or overvalued.

But the problem lies in the fact that this seemingly objective number can be “adjusted” in multiple ways. What TVL actually is depends on how you define it.

The Three Traps of TVL: Static Data, Scenario Differences, Double Counting

First, TVL is a static indicator. The current amount of locked funds cannot reflect future trends, especially in the highly volatile crypto markets, where short-term incentives and token price swings can cause significant fluctuations in TVL. A project with high TVL might quickly see its funds dry up after reward periods end.

Second, the meaning of TVL varies greatly across different application types. This means you cannot directly compare the TVL of a DEX with that of a lending protocol. For example, Uniswap’s TVL represents the liquidity pool size, while Compound’s TVL reflects the “deposit minus borrow” amount—i.e., the remaining available liquidity within the protocol.

The most critical trap occurs at the blockchain layer: layered stacking of protocols can lead to severe overestimation of the entire chain’s TVL. This is the core issue behind Ian Macalinao’s event.

Focusing on Specific Applications: What Does Different Protocols’ TVL Really Mean

TVL in DEXs: A True Reflection of Liquidity

In decentralized exchanges (DEXs), TVL more closely reflects its true meaning. Take Uniswap as an example: since it does not require liquidity mining or staking LP tokens or governance tokens, its TVL directly represents the total funds provided as liquidity. The more funds, the lower the slippage, and the better the trading experience.

However, when DEXs introduce governance token incentives, the situation becomes more complex. Users of Curve and Sushi can stake governance tokens to earn a share of trading fees. DeFi Llama separates this staked token amount as “Staking,” rather than including it in TVL. The logic is: although these tokens are locked in protocols, they are not part of the liquidity but are certificates of earning protocol rewards.

TVL in Lending Protocols: Varies by Protocol

The calculation methods for TVL in lending protocols differ. Compound’s TVL is based on “deposit minus borrow,” representing the remaining available liquidity within the protocol.

Aave further complicates this by allowing users to stake AAVE tokens and LP tokens to earn inflation rewards. These staked funds are also separated out and listed under “Staking,” not in the basic TVL.

MakerDAO’s model is different. The funds borrowed from Maker are in DAI stablecoins issued by the protocol, which do not affect the actual deposited assets. Therefore, Maker’s TVL directly represents the total deposited funds, i.e., the total collateral.

The Black Hole of Double Counting: Why Some Projects’ TVL Is Artificially High

Yield Aggregator Projects: Recursive Capital Flows

Yield-generating projects exemplify the “matryoshka doll” nature of DeFi—users deposit funds into yield aggregators, which then transfer these funds into other underlying protocols for mining. This design causes the same funds to be counted multiple times at the chain level.

Yearn and Convex Finance are typical examples. Take Convex: it is essentially an application built on Curve. Users can earn higher Curve mining yields through Convex because it holds large amounts of CRV tokens and stakes them, boosting mining efficiency for users on Curve.

When Solana’s total TVL was only about $10.5 billion, Saber (a DEX on Solana) and Sunny (a yield aggregator built on Saber) accounted for $7.5 billion—meaning a 71% double counting rate. This was the technical basis for Ian Macalinao’s ability to “fictitiously” create the appearance of Solana’s prosperity.

Liquidity Staking: Derivatives’ Reuse

Liquidity staking protocols (like Lido) are another area prone to double counting. Lido allows users to stake ETH to earn PoS rewards and issues stETH derivatives to provide liquidity.

Lido’s TVL is about $7.75 billion, with $7.61 billion of ETH locked on Ethereum. But that’s not the whole story—stETH is not static. About 21.6% of stETH is deposited into Aave as collateral, and about 14.7% is providing liquidity in Curve’s ETH/stETH pool. These stETH tokens are already included in Aave and Curve’s respective TVL statistics.

DeFi Llama’s current approach is: when calculating blockchain TVL, it no longer counts the principal in liquidity staking protocols; only when stETH is used in other on-chain protocols will it be included. But this introduces new issues—some stETH is held in centralized exchanges or traditional lending institutions, theoretically backed by ETH staked on-chain, yet not reflected in the chain’s TVL.

Service Tools: Middle Layer Traffic Traps

Some protocols offer one-stop services, managing funds across multiple underlying protocols. These funds also exist within other protocols, leading to inflated chain TVL.

Instadapp is a typical example of such “middleware.” It provides asset management functions for major DeFi protocols like Aave, Compound, Maker, Uniswap, Liquity, simplifying complex DeFi operations. Users can perform cross-protocol leverage adjustments, debt swaps, and auto-refinancing via Instadapp’s interface.

Instadapp’s TVL once reached $13.5 billion, but now is about $2.6 billion. Since the funds managed by Instadapp are fully stored in other protocols, not counting them in the chain’s TVL is correct—this is a classic example of double counting.

After the Bubble Burst, Can TVL Data Be Trusted?

DeFi Llama’s adjustments, though causing a sharp decline in blockchain TVL, represent a necessary “purification” of data. When the bubble bursts, what we see is a more authentic state of the ecosystem.

TVL has different values at different levels:

• At the application level, it can be used for horizontal comparison of current project sizes
• At the blockchain level, only after removing double counting can TVL accurately reflect the true scale of the ecosystem

Understanding what TVL is essentially means understanding the maturity process of DeFi. Early data “flooding” reflected the industry’s immaturity, while now, standardized calculation of TVL indicates industry maturation. Future innovations may further change how TVL is defined—such as introducing time-weighted averages (TWAP) or user activity weights—but regardless of evolution, the core goal should be to more accurately reflect the real state of the DeFi ecosystem.

(This article is a rewrite and analysis based on PANews reports)

Statement: The content of this article solely represents the personal views of the author and does not reflect the stance of any organization. All content and opinions are for reference only and do not constitute investment advice. Readers should conduct their own research and bear the consequences of any investment decisions made based on this article.