Ethereum didn’t have a flashy year. It had a grown-up one.
TVL finished around $68.8B, down a touch over 7% YoY. In past cycles, that single number would’ve been enough to spark the usual “Ethereum is dying” takes. This time it doesn’t land, because it misses what actually changed.
The balance sheet cooled, but the engine got hotter. The network generated north of $4B in fees in 2025, up more than 45% from 2024. That divergence is the story: maturation driven by efficiency.
App Dominance
A handful of protocols now anchor Ethereum’s balance sheet.
Aave remains the gravitational centre of DeFi, commanding more than 26% of total network TVL. It is followed closely by Lido, which controls just over 20%, and EigenLayer, which has quietly grown into a ~10% share through restaking. Together, these three protocols alone account for more than half of all capital deployed on Ethereum.
Beyond them, depth is still meaningful but clearly tiered. Binance’s staked ETH product, EtherFi, Ethena, Sky, Spark, and Uniswap form the second ring of liquidity — relevant, active, but no longer shaping the system’s centre of mass.
This is what consolidation looks like in practice. Capital is choosing familiarity, trust, and embedded distribution over experimentation.
Fees
Despite the contraction in headline TVL, Ethereum generated over $4 billion in fees this year — a 45% increase year-on-year. That divergence is the defining story of 2025.
Ethereum is holding less capital, or at least less dollar-denominated value, while monetising activity far more efficiently. The network is doing more with less, a mark of a system shifting from speculative expansion toward sustained usage.
Three protocols dominate this fee engine. Lido accounts for roughly 17% of all Ethereum fees. Aave follows with around 13%, and Sky contributes close to 10%. Together, they generate nearly 40% of the network’s total economic throughput.
However, even within that dominance, change is visible. While Lido and Uniswap still sit at the top of cumulative fee rankings, their year-on-year fee growth has slowed or declined. Market share is not collapsing, but it is being chipped away by faster-moving competitors.
Revenue
When fees are stripped back to actual protocol revenue, the hierarchy shifts. Total yearly application revenue across Ethereum settled around $786 million, slightly down year-on-year. That gap between fees paid and revenue captured reflects a growing share flowing to validators or being burned, rather than accruing to protocol treasuries.
Within that narrower slice, Sky has emerged as the standout monetisation engine. It captured roughly 22% of all protocol revenue this year, far ahead of both Aave and Lido, which each sit around 11%.
Sky’s $137 million in revenue dwarfs Aave’s $92 million and Lido’s $86 million. The spread between first and second place is not subtle. It reflects a business model optimised for extracting value from activity, not merely hosting it.
DEX Dominance
Uniswap remains the sovereign of swaps. It closes the year with roughly $2.8 billion in TVL, over $2 trillion in cumulative volume, and nearly $600 million in fees. Its V3 architecture continues to define capital efficiency for AMMs. Yet dominance is eroding.
Uniswap still controls just over half of DEX TVL, but Curve has staged one of the quietest recoveries of the year, growing its share from 25% to over 36%. Stablecoin-focused liquidity is back in demand, and Curve is once again relevant.
More disruptive, however, is the rise of Fluid. Despite lacking traditional TVL tracking in many aggregations, Fluid has captured meaningful volume and fee share. It now ranks second in both DEX volume and fee generation, pulling in over $140 million in fees on more than $300 billion in volume.
Uniswap’s share of DEX volume fell from nearly 70% to below 50% this year. Its fee dominance dropped from over 90% to under 70%. Curve, meanwhile, surged from a negligible fee share to over 20%.
This is not a collapse. It’s fragmentation. Execution quality, routing efficiency, and specialization are starting to matter more than brand alone.
On a market-cap-to-TVL basis, Curve and Balancer sit at the bottom of the spectrum. Both protocols secure deep, sticky liquidity relative to their market caps, implying that the market is sceptical about how defensible their fee streams are over the long run.
Curve’s resurgence in fee share this year proves it remains economically important, and how quickly DEX liquidity can migrate when incentives or execution quality shift.
Uniswap occupies a different valuation regime. Despite losing volume and fee dominance over the year, it continues to trade at one of the more attractive price-to-fees ratios in the sector.
That disconnect matters. It suggests the market is willing to underwrite Uniswap’s cash-generation ability even as competitive pressure increases, effectively treating it as infrastructure rather than a growth play. In other words, Uniswap is being priced less like a speculative AMM and more like a mature exchange primitive.
Lending Dominance
If DEXs are fragmenting, lending is doing the opposite. Aave is the undisputed hegemon of on-chain credit. It holds roughly eight times the TVL of its closest competitor and generates nearly seven times the fees. In a sector where trust, liquidity depth, and risk management are everything, Aave has no peer.
Its dominance actually expanded this year, growing from around 61% to over 67% of lending TVL. Spark contracted. Morpho stagnated. Maple, however, surged — climbing from under 1% to over 7% market share, driven by institutional adoption and private credit demand.
Fee share tells the same story with a twist. Aave still captures nearly half of all lending fees, but Maple now commands over 21%, directly carving into Aave’s territory. The market is not rejecting the leader; it is selectively building alongside it.
Aave still trades on a conservative P/F relative to its role as Ethereum’s credit backbone. Euler is the only outlier with a lower P/F, despite operating at a fraction of Aave’s scale and liquidity depth.
LST DOMINANCE
Lido remains the giant. With over $25 billion in TVL, it holds more than double the capital of Binance’s staked ETH product. It also retains the LST triple crown: the highest TVL, the highest fees, and the highest revenue. But its moat is narrowing.
Over the past year, Lido ceded nearly 10 percentage points of market share. Binance Staked ETH gained over 12 points, climbing to roughly 23.5% of the market. Centralised exchange distribution is back, and it is working.
Lido’s fee dominance remains volatile, oscillating between 60% and 75% in line with ETH price and activity. Binance’s fee capture, by contrast, shows a steady, linear climb. The difference is not product quality, but distribution.
Despite their strategic importance, LST tokens still trade like the market doesn’t treat them as claim-on-cashflow assets. Market cap to TVL ratios are compressed across the board, and even Lido’s relative premium looks modest versus other DeFi verticals, largely because most LST tokens are structurally governance and utility instruments, not revenue-sharing equities.
In plain terms, the staking layer can print fees, but tokenholders often don’t directly capture them. Until that changes, through explicit value accrual, clearer buyback mechanics, or credible governance control over monetisation, the market will keep valuing LST tokens more like protocol “access keys” than ownership in the economic engine.
There is also a more structural reason for that discount. LST fees and revenue are almost entirely a function of the underlying asset they track. When ETH goes up, staking revenue goes up. When ETH activity slows, it comes back down. That makes LST economics highly correlated to the base asset itself, not to any unique protocol alpha.
For a user, that creates a simple question. Why take exposure to a riskier governance or utility token, with no direct claim on cash flows, when the same upside can be captured more cleanly by holding ETH itself? Until LST tokens can clearly decouple their value from the price action of the underlying asset, or offer differentiated upside beyond passive correlation, that trade-off will continue to cap how the market values them.
THE KINGMAKERS
Kingmakers do not just hold assets; they facilitate the primary economic activity of their respective sectors.
Aave commands a massive 58.65% share of all DeFi borrows ($21.26B), serving as the ecosystem’s primary source of credit.
Lido maintains 46.69% dominance in Liquid Staking TVL ($25.74B), acting as the liquidity foundation for staked ETH.
Uniswap captures 20.75% of total DEX volume ($2.53B daily), remaining the highest-velocity trading venue in DeFi.
The data confirms that Aave and Lido have achieved “escape velocity” in their sectors, controlling nearly half or more of their respective markets. However, Uniswap’s volume-to-TVL efficiency, Maple’s institutional borrow capture, and Curve’s renewed grip on stablecoin liquidity all point to the same shift: market share is increasingly being won through specialised utility, not just passive asset accumulation.
What 2025 Really Says About Ethereum
Ethereum didn’t get louder this year. It got tighter. TVL cooled, fees surged, and power consolidated. The network is moving away from speculative sprawl and toward something harder to fake: economic density.
Aave, Lido, and Uniswap still sit at the core. They’re the default venues for credit, staking, and price discovery. But the edges are where the signal is shifting; Curve is reclaiming stablecoin gravity, Fluid is pulling flow with a different architecture, and Maple is proving there’s real demand for on-chain credit that doesn’t look like DeFi 2021.
Ethereum in 2025 isn’t a story about attracting capital. It’s a story about earning on the capital that’s already here.
This analysis was built using DefiLlama Pro, with a custom dashboard supporting the data.