DEXs are often framed as inefficient businesses because protocol revenue looks low relative to gross fees. That framing is wrong.

Over the last 30 days, DEXs generated $220.7M in fees but only $85.5M in protocol revenue.

On the surface, that 39% fee-to-revenue conversion looks weak compared to perps. In reality, it reflects intentional economic design, not leakage.

The missing dollars did not disappear. They were paid out to liquidity providers.

DEXs are built as two-sided markets, not vertically integrated exchanges. Liquidity is externalized. Capital comes from LPs, and LPs must be compensated for inventory risk, impermanent loss, and opportunity cost. The fee split is the incentive that keeps liquidity deep and spreads tight.

This is why DEXs behave like toll roads rather than operating companies. Every swap, arbitrage, rebalance, and bridge pays a fee, but the protocol only takes a cut for maintaining the routing layer. The rest is the cost of attracting and retaining liquidity.

Compare that to other DeFi sectors:

- Perps internalize execution and keep almost all fees.
- Lending passes most fees to depositors as yield.
- DEXs pass most fees to LPs to buy liquidity.

Each model optimizes for a different constraint.

DEXs optimize for volume stability and composability, not maximum protocol margin. That trade-off is deliberate. If the protocol tried to capture more fees, liquidity would thin, spreads would widen, and volumes would fall. Revenue would not rise; the market would move elsewhere.

So the revenue gap is not inefficiency.
It is the price of liquidity.

When you look at DEXs this way, the correct question is not “Why don’t they earn more?” but “How much volume can they sustain at scale while keeping LPs solvent?”

That is the real business DEXs are running.