Solana Weekly Intelligence — September 7–13, 2026
Every week I answer the same five questions with the same rule: real data or no data. This week the answers changed shape — price fell 4.5% but fees jumped 28.3%, the stablecoin mix kept rotating from USDT toward USDC, and the ETF tape printed its first clearly negative flow day of the month. None of these is a verdict on its own; together they describe a market deciding whether August's breakout was a re-rating or just a rally.
Question 1 — What did the market actually do?
SOL opened at $106.49 (its weekly high daily close), fell through the week, flushed to $98.63 on Friday and closed Sunday at $101.75 — down 4.5%. Context makes it fairer: Bitcoin fell 3.8% the same week, so 0.6 points of SOL's decline was idiosyncratic and the rest was beta. Against Ethereum the gap was worse: ETH rose 0.45%, leaving SOL −4.9pp. After August, when SOL beat both majors by double digits over the breakout weeks, two facts are new: this is the first week of the period where SOL lagged ETH by that much, and the second week in a row where it lagged BTC.
The level that matters from the daily series is $98.63 — Friday's close, the week's low. The weekend bounce ($102.39, then $101.75) means buyers defended the high-$90s. If that level breaks on real volume next week, the August breakout is fully retraced to the low $90s where it began.
Question 2 — Is DeFi still growing?
| Week ending | TVL | WoW | DEX 7d | WoW |
|---|---|---|---|---|
| Aug 16 | $4.82B | — | $10.9B | — |
| Aug 23 | $5.57B | +15.6% | $19.8B | +81.5% |
| Aug 30 | $5.92B | +6.3% | $18.5B | -6.2% |
| Sep 6 | $5.92B | 0.0% | $16.5B | -10.9% |
| Sep 13 | $5.91B | -0.2% | $18.5B | +11.8% |
TVL has gone flat at $5.91-5.92B for three consecutive weeks after the breakout run. That is not deterioration — a 4.5% price drop with flat USD TVL means deposits replaced the price effect — but it is a pause. DEX volume's +11.8% bounce to $18.5B reverses the previous week's drop and confirms the $16.5-19.8B band as the post-breakout normal. The honest summary: DeFi on Solana stopped growing this month, and started clearing.
Question 3 — What is happening to the stablecoin base?
| Asset | Aug 16 | Aug 30 | Sep 13 | Trend |
|---|---|---|---|---|
| USDT | $2.90B | $2.84B | $2.54B | Falling |
| USDC | $6.70B | $6.88B | $7.30B | Rising |
| Total | $15.76B | $15.96B | $16.46B | Flat-to-up |
The total is drifting up — $16.46B at the week's close, up 4.4% over five weeks — but the composition tells the real story: USDT has fallen from $2.90B to $2.54B while USDC climbed from $6.70B to $7.30B. That is a $360M rotation out of USDT and $600M into USDC on this chain. USDC now backs more than 44% of Solana's stablecoin float versus about 43% for USDT and others combined. Rotations like this usually reflect venue and user preference, and their persistence (four of five weeks down for USDT) is what makes them worth respecting.
Question 4 — Is fee revenue real or borrowed from the rally?
Real, and this week proved it. Network fees hit $108.3M — the highest weekly print of the five-week series — in the same week price fell 4.5% and DEX volume rose only 11.8%. Fees growing three times faster than volume means transaction count per dollar is climbing: small trades, more of them. The structural backdrop is August's record 5.2 billion non-vote transactions (an all-time monthly high, up 19% from July, per the official Solana account citing Blockworks). Jito tips rose to $2.08M from $1.81M, the same direction at smaller scale. If next week holds fees above $100M, the August activity was a level change, not a spike.
Question 5 — What would change my mind?
Three falsifiable things, in order: a second consecutive negative ETF week (September 11 printed −$278.84K against a $1.42B base, the first clearly negative day of the month per SoSoValue); a daily close below $98.63; and a fifth straight weekly USDT decline. Conversely, two things would strengthen the bull read: fees above $100M again, and a tokenized-equity supply print above the record $684M. I also track the RWA complex more broadly — Solana led all chains with $348M in 30-day net RWA inflows and $4.23B total per a September 7 report, and the tokenized-equity records this week suggest that leadership is compounding.
Health check
| Area | Status | Why |
|---|---|---|
| Price trend | Amber | -4.5% week, second straight week lagging BTC |
| DeFi depth | Green | TVL flat at $5.91B through a price drawdown |
| Trading activity | Green | DEX $18.5B, +11.8%, back inside the $16.5-19.8B band |
| Fee revenue | Green | $108.3M, series high, rising against price |
| Stablecoin base | Amber | Total drifting up but USDT down 4 of 5 weeks |
| ETF flows | Amber | First negative day of the month (Sep 11); base $1.42B |
| Regulatory | Green | SOL named a core ETF asset Sep 5; no adverse actions |
The ambers are watch-items, not warnings: a price pullback inside an uptrend, a stablecoin rotation with a growing total, and one choppy ETF day after weeks of inflows. Nothing in the data this week contradicts the structural story — record August transactions, record tokenized-equity supply, an anchor ETF past $1B — but the tape needs to reassert itself for the greens to outnumber the ambers again.
Three datasets that disagreed this week - and what each one measures
The week's defining feature was disagreement between datasets, and disagreement is only useful if you know what each instrument measures. Price (-4.5%) measures sentiment and market beta - it is the noisiest series here and the most gameable. Fees ($108.3M, the series high) measure paid transactions - the least gameable series on the chain, because every fee dollar is a transaction someone chose to pay for. ETF flows (-$278.84K on one day against a $1.42B base) measure institutional positioning cadence - slow, sparse, and meaningless at daily resolution except as sequence.
Weighted by instrument quality, the week's reading inverts its price action. The least gameable series printed its high while the noisiest fell - and my Question 4 answer this week ('real') is an application of exactly that weighting. The slow series agreed: TVL flat through the drawdown, stablecoins within 1% of their series high. Only the fast, noisy instruments (price, relative performance) told the bearish story.
What disagreement does not mean is error - all three instruments were measuring the same week honestly. It means the week contained two different events: a repricing of the token and a continuation of the chain's usage. The five-question framework forces both into the same answer set, which is why the health check shows greens and ambers simultaneously instead of collapsing into one grade. The resolution conditions are already encoded in Question 5 - a second negative ETF week, a close below $98.63, or a fifth USDT decline would start converting the ambers; fees above $100M again, or a tokenized-equity print above $684M, would confirm the greens.
Sources
Source: DefiLlama — TVL, DEX, fees, stablecoins, pricesSoSoValue — ETF flowsBlockworksCoinMarketCap — Solana
What are the five questions this series answers?
Market, DeFi depth, stablecoins, fee revenue, and what would change the thesis — the same five every week, so each week's answers are comparable.
Why is the health check more amber this week?
Because three areas moved the wrong way at once: price lagged both majors, USDT fell again, and ETF flows printed a negative day. None is severe on its own; the count is the honest summary.
Is $98.63 a technical level I should care about?
It is the week's lowest daily close and the level the weekend bounce defended. It matters because a weekly close below it would retrace the breakout to its starting range.
What does 'fees rising against price' mean?
Dollar fee revenue rose 28.3% while SOL fell 4.5%. It means activity (transactions) grew faster than dollar volume — typically small-trade intensity — and it is the strongest real-usage signal of the week.
Is the USDT decline bad for Solana?
The total stablecoin base is still growing; the rotation is from one issuer to another. I don't assign a cause the data doesn't show, but four declines in five weeks makes it a genuine watch item.
Where do the numbers come from?
DefiLlama public APIs (prices, TVL, DEX, fees, stablecoins), fetched September 14, 2026, plus attributed reports from SoSoValue, Blockworks and CoinMarketCap for ETF and event data.
When datasets disagree, which one wins?
Rank by gameability and noise: fees (paid transactions) over TVL and stablecoins (slow deposits), and both over price and daily ETF prints. This week the least gameable series printed its high while price fell - so the divergence reads as a repricing of the token happening alongside a continuation of usage, not as a data conflict.
Is the ETF's red day more important than the fee record?
Not yet. One daily print of -$278.84K against a $1.42B base is sequence, not trend - the same complex printed +$10.19M on September 1. The fee record is a series high on the least gameable series the chain has. The ETF print matters if it becomes a second consecutive negative week; until then it is a watch item.
Need on-chain energy without the price tag?
Rent TRON Energy at Tronsell →A service we run and trust: a ~400M TRX self-operated energy pool, with 60-90% savings versus on-chain energy costs.