Scroll has become a busier network with far less money in it, and that split tells a sharper story than either trend alone. People kept arriving and transacting through the autumn, while the stock of value held in its finance protocols shrank to a fraction of its size. Use did not revive the money; the money left while use grew.
Two independent sources frame the question. DefiLlama, an open aggregator of decentralised-finance data, published its Scroll reading on 12 November 2026, putting total value locked at $6.5 million, down 96.7% from $196.3 million a year before. growthepie, an independent analytics service for Ethereum and its layer-2 networks, measured activity from its own on-chain data over the seven days ending a day earlier. The question is whether rising use can rebuild that lost stock, or whether Scroll has settled into a different kind of network, one that people use without parking large sums in it.
The April break that reset Scroll's scale
The monthly opening readings show a network that held a large stock through the early months of 2026 and then lost it in one step. On 1 April 2026 the figure stood at $228.1 million. A month later it stood at $19.0 million. No gradual drift explains a move of that size in a month, and the network never returned to the earlier range.
That step set the scale for everything after it. By the start of November the opening figure had fallen further, to $7.4 million. A stock once measured in the low hundreds of millions was now measured in single-digit millions. Activity could rise from a smaller base and fees could move either way, but the money held in place had already gone.
Ether moved, but not enough to explain the fall
Prices matter here because total value locked is a dollar figure, so it rises and falls with the market value of the assets held. The same aggregator published ether at $2,055 in its latest reading, against $1,884 on 14 August 2026, up 9.1%. The same source put total value locked then at $9.8 million, down 34.0% since then.
Those two movements point in opposite directions, and that is the point. A rising ether price should have lifted the dollar value of ether-denominated holdings, not cut it. The gap leaves only a few explanations the sources allow: money left, holdings changed, or the list of protocols counted changed.
Total value locked changes in two ways: when the quantity held changes and when the price of what is held changes. When prices fall and nothing leaves, the dollar figure still falls. If prices rise and money leaves faster, the dollar figure falls despite the price rise. The August comparison shows the second case.
More activity, and less money paid to use it
Meanwhile, use rose on growthepie's count. Seven-day averages showed daily transactions at 14,360, up 34.3% from 10,693 thirty days earlier. Active addresses reached 1,901, up 38.6% from 1,371 over the same period. The rise was stronger for addresses than for transactions, and it was stronger still against the August week, when addresses stood at 817, up 132.6% since then.
Fees tell a different part of the same story. Users paid $184 a day in fees on average, down 20.0% from $229 a month earlier, even as transactions climbed. They had been $156 in August, up 17.4% in the latest count. Use grew in both windows, but the money paid to use the network did not grow with it in the recent window.
Counts and value measure different things, and the split between them is the finding. Transactions count actions, addresses count points of activity, locked value counts money held in finance protocols, and fees count what users paid. When the first two rise while the third falls and the fourth also falls, the natural reading is light, cheap use: actions that do not require large sums to sit in protocols.
Stablecoins fell with locked value, not with ether
Stablecoins provide a useful cross-check because they are built to keep a steady dollar price, so their total does not move with ether. growthepie put that total at $8.1 million in its latest reading, down 11.1% from $9.1 million on 12 October 2026. It had been $10.8 million in August, down 25.4% since August. The fall is smaller than the fall in locked value over the same broad period, but the direction is the same.
Both dollar-based stocks of value on Scroll moved down together, while activity moved up. That alignment strengthens the main finding, but it does not name a cause.
DefiLlama counts only the protocols in its list, and its dollar figure moves with token prices as well as with money moving in or out, so a fall can reflect coverage, prices, withdrawals, or a mix of all three. growthepie counts activity and stablecoin supply on the network itself, and one address need not mean one person, so its rise cannot be read as a headcount. Neither source explains the step between April and May, and neither names a protocol, a withdrawal, or a decision behind it.
A competing reading fits much of the same evidence: the new activity may be light use that was never going to lock large sums, with more points of activity rather than a revival of finance on Scroll. That reading sharpens rather than overturns the main finding. Scroll in late 2026 is a network people use, not a network where people leave large sums parked. Whether use of this kind can rebuild a stock of locked value, or whether the two will stay apart, will be tested by the next monthly readings: if locked value rises while activity holds, the split is closing; if activity rises again while locked value falls, it is widening. Not investment advice.
Source: DefiLlama, growthepie, Scroll fundamentals · Date: