Ethereum is on course for its strongest third quarter on record. At the time of writing, CoinGecko's hourly prices put ETH roughly 71% above its 1 July level; the exact figure moves with the market, but the claim that the gain exceeds 67% is already true.
That is excellent news for holders. It is not, however, an answer to the most interesting question around Ethereum.
DefiLlama currently counts about $53.6 billion locked in DeFi on Ethereum itself, before adding the activity and assets living on its Layer 2 networks. Ethereum is useful. It is liquid, widely integrated and difficult to replace. But a useful network and a valuable token are not the same proposition. The missing step is value capture: which parts of all that activity create durable demand for ETH, reduce its supply or pay its holders?
The distinction matters because Ethereum's engineers have become very good at making the network cheaper.
The old explanation was beautifully simple
EIP-1559 gave ETH a value-capture mechanism anyone could explain in one sentence: users pay the network's base fee in ETH, and the protocol destroys that ETH.
When blockspace was expensive, the burn could exceed the new ETH issued to validators. More use meant more fee pressure; more fee pressure meant more ETH removed from supply. During those periods, Ethereum the network and ETH the asset appeared to be connected by a short, visible pipe.
That pipe still exists. Execution base fees are still burned. Blob fees paid by rollups are also burned. But the volume flowing through it has changed.
Since Dencun introduced blobs through EIP-4844, rollups can publish data to Ethereum in a separate, much cheaper fee market. That was the point. Ethereum chose to sell more blockspace at a radically lower unit cost so that users could transact on L2s without paying mainnet prices. As of 23 September, growthepie's 30-day average placed Ethereum blob data availability at about three cents per megabyte.
This is a product success and, in the short term, a revenue compression.
Scaling breaks the one-transaction, one-fee intuition
A rollup can combine thousands of user transactions, collect fees from those users, compress the data and make one comparatively small payment to Ethereum. The difference between what the rollup collects and what it pays the L1 can accrue to the rollup operator, its sequencer, its token or its users through lower prices.
Only the portion paid to Ethereum enters ETH's direct fee economy.
This does not mean L2s pay Ethereum nothing. They pay for data availability, settlement and proof verification, and blob prices can rise when demand exceeds the protocol's target. It means the relationship is no longer linear. Ten times as much L2 activity does not automatically mean ten times as much ETH burned.
Nor does $53.6 billion of TVL equal $53.6 billion of demand for ETH. TVL includes stablecoins, tokenized bitcoin, governance tokens and assets counted inside lending and liquidity protocols. It measures assets deployed in applications, not revenue earned by Ethereum and not ETH purchased by those applications.
That is why “Ethereum is winning” can be true while “ETH must rise proportionally” remains unproven.
ETH still captures value in three ways
The sceptical case becomes too strong if it looks only at fees. ETH has three distinct economic roles.
First, ETH pays for blockspace. Execution and blob base fees are burned. This is the cleanest route from usage to token scarcity, but scaling has deliberately made each unit of usage cheaper. The bull case requires total demand for Ethereum blockspace to grow faster than its capacity becomes cheaper.
Second, ETH is the security collateral. Validators must stake ETH, and an attack on consensus requires acquiring and risking a large share of the stake. Ethereum.org currently shows about 43.4 million ETH staked, around 35% of supply. That is real demand and real economic utility. It also pays a yield, although part of that yield comes from new issuance rather than external revenue: stakers receive it, while non-stakers absorb the dilution.
Third, ETH can be money inside its own economy. It is used as collateral in DeFi, paired in liquidity markets, held in treasuries and bridged into rollups. This monetary premium could ultimately matter more than fee burn. But it is behavioural, not automatic. Applications can denominate balances in stablecoins, rollups can promote their own tokens, and users can interact through sponsored transactions without consciously holding ETH.
The strongest ETH thesis is therefore not “every Ethereum transaction buys ETH.” It is that the ecosystem chooses ETH as its reserve asset and that the network's security becomes valuable enough to support demand for scarce, slashable collateral.
The dashboard that would settle the argument
Price and TVL are insufficient. To see whether Ethereum's success is reaching ETH, we would watch:
- execution and blob fees paid to the L1, and how much ETH they burn;
- net supply change after validator issuance, not burn in isolation;
- the share of ETH staked and the yield coming from fees rather than issuance;
- ETH's share of collateral, liquidity and treasury balances across L1 and L2;
- what rollups pay Ethereum relative to the fees they collect from users;
- how many rollups use Ethereum for data availability instead of a cheaper external provider.
Those measures can move in opposite directions. A cheaper Ethereum can gain users and lose near-term fee revenue. More staking can strengthen security and increase issuance. Higher L2 margins can make rollups excellent businesses without making ETH an equally excellent asset.
That is not a contradiction. It is what happens when a protocol develops an economy above it.
Our answer, for now
Ethereum has built something unusually useful. The Q3 rally suggests the market is again willing to pay for that possibility. But infrastructure quality alone does not close the valuation loop.
The direct loop of activity, fees and burn is weaker than it was when mainnet congestion drove deflation. The indirect loop of staking, collateral, settlement and monetary premium is larger, but also harder to measure and easier for competing assets to intercept.
So the honest conclusion is conditional. Ethereum can succeed while ETH undercaptures that success. ETH wins decisively only if the ecosystem needs the token not just to operate the base layer, but to secure, collateralize and settle the expanding economy built above it.
The next phase of the debate should not ask whether Ethereum is useful. That case is already strong. It should ask where each dollar of economic value ends up.
This is general market commentary, not investment advice or a recommendation to buy or sell any asset.