SOL Weekly Data Report — September 21–27, 2026

2026-09-28 · Weekly Data Report · SolDataLab Research Desk

A week where every level metric set a series high and the only thing that fell was the churn. SOL closed at $121.38, up 9.2%, on the highest weekly close of the seven weeks this report tracks. TVL rose 7.4% to $6.64 billion, network fees rose 15.3% to $112.88 million, Jito tips rose 26.4% to $2.87 million, and stablecoin supply printed a record $16.74 billion. DEX volume fell 11.9% to $17.48 billion - with a measurement story attached that this report will not paper over. Every number is machine-fetched and traceable, and every judgment below is stated as a judgment rather than buried in a table.

The week in numbers

One table, one direction, one asterisk. Price, market cap, TVL, fees, tips and every stablecoin line rose; DEX volume fell, and part of that fall is the restated denominator plus a reporting change at one aggregator. The rest of this report separates the movement that is measured from the movement that is measurement - because the two are easy to confuse when they point opposite ways in the same week.

The honest frame for the table: four of these ten rows set series highs at once - close, TVL, fees and stablecoin supply - which on its own describes a week where the chain's fundamentals and its price agreed. The lone declining row is also the row this report trusts least this week, for reasons the DeFi section lays out. A table that sets four highs in one row usually means the underlying week was strong; in this case it also means the source's restatement finally stopped fighting the story.

Price: a Tuesday step, a Thursday test, a Saturday high

The week's shape: a 6.9% step into the September 22 point did the heavy lifting early, a 3.0% pullback into Thursday tested whether the move would hold - it held $4 above the prior week's $111.02 close - and a 4.3% step into the September 26 point took SOL to $122.08, the week's high close. Sunday closed $121.38, down just 0.6% from the high, which is what consolidation after a leg up is supposed to look like. The structure - higher low into the test, higher high into Saturday, quiet fade into Sunday - is what an uptrend that intends to continue looks like on a daily-close series.

Against the majors, SOL gained 5.2 percentage points on Bitcoin (+4.0% for the week) and 7.3 points on Ethereum (+1.9%). Market cap moved to roughly $71.3 billion on 587.7 million circulating SOL, up 9.4% on the rounded figures, leaving SOL 58.8% below its $294.33 all-time high of January 19, 2025 (CoinMarketCap). Two consecutive weeks of 9-12% gains with widening outperformance against both majors is the relative-performance signature of a trend week, not a squeeze - and the all-time-high arithmetic is worth stating plainly because it keeps the move in proportion: a 59%-from-peak asset just had its best two-week stretch in months, and is still less than half of where it was at the top.

The seven-week series (September 28 vintage)

Read on one basis, the week set four series highs at once: close ($121.38), TVL ($6.64B), fees ($112.88M) and stablecoin supply ($16.74B). USDC matched its six-week high of $7.30B. DEX volume's $17.48B sits mid-pack - below the $19.75B and $19.84B prints on this basis, above the $16.51B trough. A table that sets four highs in one row usually means the underlying week was strong; in this case it also means the source's restatement finally stopped fighting the story.

The restatement is the part a casual reader misses. Five of the six prior weeks had their fee prints cut by 4-13% on September 28, and the Sep 14-20 DEX total was lifted from $18.49B to $19.84B. This report uses the September 28 vintage throughout so the seven rows are internally consistent - but the archive articles keep the vintages they were published with, each stating its own basis. The practical lesson: never compare a freshly published week against an old week's printed number without checking whether the source quietly moved the goalposts between them.

DeFi depth and activity

TVL's 7.4% rise lagged price's 9.2%, so TVL expressed in SOL actually fell: deposits grew more slowly than the unit they are denominated in. That is normal in a price week and not a warning sign - it is arithmetic, and this report states it so it is not mistaken for outflows. The cleaner signal is in the fee line: $112.88 million is the highest weekly print of the seven weeks, driven by a steady $15-17M daily band rather than one spike, and Jito tips at $2.87 million are also a series high. Tips measure priority paid for inclusion - adversarial flow - and they grew 26.4% while measured volume fell, which is one more reason to treat the DEX series with care this week.

The DEX decline needs its honest sentence, and it is two sentences. DefiLlama restated the prior week upward to $19.84B, and Jupiter's reported volume fell far more sharply than its fee line during the week, which points to a reporting change at the aggregator level rather than a usage collapse. Against the $18.46B print published in last week's article, this week would read -5.3% rather than -11.9%. Either way, Sunday's $1.08B daily print - roughly half of Saturday's $2.16B - carries the Jupiter gap and should be read as a floor, not a collapse. Fees and tips, which cannot be double-counted the way routed volume can, are the harder currency this week, and they both printed series highs.

Fee-per-volume: why this report trusts fees over DEX

Dividing the week's fees by its DEX volume gives a fee-to-volume ratio near 0.65% ($112.88M against $17.48B), in the same band the series has held since late August. That ratio is the tell: it is stable precisely because so much of the volume is adversarial flow paying for priority, and it is why a volume number that swings 12% in a week while fees rise 15% is a measurement artifact, not an activity signal. The report's standing practice is to lead with fees and tips, quote DEX volume with its caveats attached, and never let a single soft series drive a conclusion about whether Solana is being used.

This is also where the Bitquery ~58% bot-volume finding (covered in the News Review) lands as confirmation rather than surprise. An estimate with a wide error bar, yes - but its direction is exactly what a stable 0.65% fee-to-volume ratio implies. The consequence for readers is mechanical: if you use DEX volume as a proxy for activity, discount it, and if you need a harder number, use fees and tips, which this week said the chain was busier than it has been in the seven weeks tracked.

The stablecoin layer: a record, and a composition shift

The record total is the headline; the composition is the information. USDT's 25.9% jump is its first sharp expansion since the August stretch and ends three weeks of contraction. The non-USDC/USDT bucket reached $6.77 billion, a series high, and now holds 40.4% of the base - the dollar layer keeps diversifying away from a two-issuer structure. Third-party trackers cited higher totals midweek (the Solana Foundation's $17.3B, SolanaFloor's $17.39B on September 25); this report stays on its stated DefiLlama basis so the series remains comparable across weeks, and notes the spread as methodology rather than picking the biggest number.

The read: a record total and a record third-bucket share at the same time is resilience, not just size. A 40% share outside the two majors means Solana's stablecoin base no longer depends on a duopoly - a structure last cycle's chains never had. The USDT-led composition also fits a trading-driven week: USDT minting and the long tail of tokens tend to lead in a 9.2% price week, while USDC settlement grows more slowly. That is the pattern you would expect, and it is visible only because the series stays on one methodology.

Staking and the network

The staking layer is the steadiest table on this site: stake added 0.6M SOL while price did the moving, holding the staking rate at roughly three-quarters of circulating supply. On a mainnet query of September 27, annual inflation was 3.6278% against a total supply of 634.76M SOL, with 440.55M delegated - a gross staking yield of about 5.23% before commissions (CryptoTicker's mainnet arithmetic, recomputed by this site). Transaction counts from the PublicNode sample window dipped 3.6% week over week; the window is roughly twelve hours normalized per day, so small dips inside a range are noise, not signal - the network did not meaningfully slow even as it set price and fee records.

The validator count ticked from 676 to 675 - a rounding-level move that this report mentions only to pre-empt the question. Client diversity and the Firedancer track remain a parallel maturity story that did not move this week; Alpenglow's testnet switch (covered in the News Review) is the consensus track, and it is not the same work as a second validator client reaching mainnet.

Data integrity: what this report does and does not do

This week's disclosure is about the source, not this site's errors. Between the September 21 and September 28 fetches, DefiLlama restated its historical Solana fee series down by 4-13% for the weeks from August 23 through September 20, and raised DEX volume for September 14-20 from the $18.49B published at the time to $19.84B. This report uses the September 28 vintage throughout - the seven-week table above is internally consistent on that basis - while archive articles keep the vintages they were published with, each stating its own basis. The September 27 fee day was finalized during the September 28 fetch window and may still see minor revision; that is the fourth consecutive week in which this source revised a recent day after first publication.

One aggregator-level caveat sits inside the DEX series: Jupiter's reported volume fell roughly an order of magnitude more than its fee line week over week, a divergence that fee data - which held its normal range - does not support as a usage collapse. This report leaves the series as published and flags the anomaly rather than adjusting a number it cannot recompute. Beyond that, the omissions rule is unchanged: no perps volume (DefiLlama's perps endpoint returned no usable series on fetch day), no priority-fee average, no active address counts, no compute-unit totals, no pump.fun-specific series. Metrics without a free and reliable source are omitted rather than estimated.

Methodology constants: weeks run Monday 00:00 to Sunday 23:59 UTC; price is the DefiLlama daily close at 00:00 UTC of the labelled date; TVL and stablecoins are week-ending level snapshots; DEX volume, fees and tips are summed across the seven daily buckets; market cap is circulating supply times close; week-over-week percentages are computed on the rounded values displayed in the tables.

Risk watch

Levels that would convert this week's headlines into a trend, or turn them back into spikes:

• A weekly close below $111.15 - the week's opening print - would hand back the entire move and restore the September band; below $114.99, Thursday's test level, the consolidation reading weakens
• A second consecutive DEX decline below $17B would turn the measurement caveats into a trend question; the number to watch is whether the fee line keeps its $15-17M daily band even as volume softens
• Stablecoin supply above $17B on this site's basis would confirm the record as a level change rather than a spike; a fade back below $16.4B would read as mint-and-redeem churn
• Any Alpenglow mainnet scheduling signal would move the consensus story from roadmap to calendar

Source: DefiLlama - Solana chain TVL, DEX volume and fee series (September 28 vintage)DefiLlama - stablecoin supply charts for Solana (USD-pegged basis)DefiLlama Yields - liquid staking pool APYSolana RPC via PublicNode - stake, validators, performance samplesCoinMarketCap - Solana circulating supply, market cap and all-time highCryptoSlate (CMC data) - Solana market data snapshot, Sep 27, 2026CryptoTicker via CoinMarketCap - mainnet inflation and staking yield arithmetic, Sep 27, 2026

Tags: DeFiStablecoinNetworkStaking