SolDataLab

SOL Weekly News Review — September 28–October 4, 2026

2026-10-05 · Weekly News Review · SolDataLab Research Desk

The week after a record is a test of what the record meant, and this one returned a split verdict. SOL closed at $119.60, down 1.9% against a Bitcoin that gained 0.4% and an Ethereum that was flat - the first weekly decline in three weeks, and the first week since July in which SOL lagged both majors. Underneath the flat tape, the demand channel that carried the quarter nearly stalled: US spot Solana ETFs took in just $2.43 million after $188.21 million the week before, and September 30 brought the complex's first daily net outflow in weeks. The consensus calendar, meanwhile, taught everyone a lesson in discipline - September 28 came and went without Alpenglow, exactly as the fine print said it would. Below, the week's stories, each with the numbers and the take.

The week in one paragraph

SOL walked from $121.96 on Monday to a $118.04 low close on Thursday, then climbed three straight closes to finish at $119.60 - a 1.9% weekly decline, 2.3 percentage points behind Bitcoin and 1.9 behind Ethereum. TVL slipped 1.5% to $6.52 billion, DEX volume fell 11.1% to $16.28 billion on a restated base, but network fees held nearly flat at $111.67 million (-1.0%) and Jito tips edged up to $2.91 million. Stablecoin supply rose 0.5% to $16.82 billion, with USDC printing a series high of $7.43 billion. ETFs netted $2.43 million on the SoSoValue tally (Farside: $0.8 million) - the fourteenth consecutive positive week, by a margin of one decent Monday. Every figure is machine-fetched from DefiLlama or the sources attributed below.

The week's data snapshot

MetricValueSource
SOL weekly close$119.60 (-1.9%)DefiLlama
Weekly high / low close$121.96 / $118.04 (band compressed 64%)DefiLlama
Relative performance-2.3pp vs BTC, -1.9pp vs ETHDefiLlama
TVL$6.52B (-1.5%)DefiLlama
DEX volume (7d)$16.28B (-11.1% on a restated base)DefiLlama
Network fees (7d)$111.67M (-1.0%)DefiLlama
Jito MEV tips (7d)$2.91M (+1.4%, series high)DefiLlama
Stablecoins total$16.82B (+0.5%, third straight record)DefiLlama
USDC / USDT on Solana$7.43B series high (+1.8%) / $2.84B (+6.4%)DefiLlama
ETF net inflows (Sep 28 - Oct 2)+$2.43M (SoSoValue); +$0.8M (Farside)SoSoValue / Farside

The through-line: nothing broke this week, but everything showed its dependency. The ETF streak survived because $1.3 million showed up on Friday; the price level held because three quiet closes in a row rebuilt the floor; the fee line held while volume fell, which is either resilience or the bot share doing its work. Flat weeks are where a market shows you which of its series are load-bearing - and this week pointed at fees, not volume.

1. The September 28 Alpenglow date came and went - because it was never a date

The mainnet did not activate Alpenglow on September 28, and no serious source ever said it would. The date circulated because it appears in Anza's Agave v4.3 release schedule as the point when mainnet resumes feature activation - a routine runtime upgrade window - and nowhere does that schedule reference Alpenglow. Anza's head of research Roger Wattenhofer, whose message became the headline 'No Alpenrush', and Solana co-founder Anatoly Yakovenko both said the protocol would not be rushed; Alpenglow (SIMD-0326) completed its testnet activation on September 24 and devnet on September 25, and mainnet has no announced date. A whiteboard primer from Solana Developers with Wattenhofer and Jacob Creech landed on September 28 - the educational channel working overtime while the rumor channel corrected itself. Sources: India Crypto Research's weekly review, October 5; Anza's Agave 4.3 schedule; Solana Developers on X, September 28.

The useful part of this story is what it says about how activation will actually arrive. Alpenglow migrates over epoch boundaries, requires validators to acknowledge the switch with supermajority stake - testnet needed 82% - and has no fixed activation time by design. The signal to watch is not a calendar entry; it is a governance action naming a mainnet activation epoch, or Firedancer shipping Votor (more on that below). Until one of those prints, every 'Alpenglow date' is someone else's schedule wearing a better story.

My take: The phantom date is the most useful thing that happened all week, precisely because nothing happened. A calendar entry that never mentioned Alpenglow had become 'the Alpenglow date' by sheer narrative gravity, and it expired without a gap in the book - which tells me the people who trade size were never pricing it. From here on, my standard is mechanical: I believe a mainnet activation date when it names an activation epoch, or when Firedancer ships Votor. Everything else is someone else's schedule wearing a better story.

2. ETF flows fell from a record $188.21M to $2.43M - and the streak survived on the last day

US spot Solana ETFs took in $2.43 million of net creations across September 28 - October 2 on the SoSoValue tally, a 98.7% drop from the prior week's record $188.21 million, and the narrowest positive week of what is now a fourteen-week streak dating to late June. The daily shape tells the whole story: +$12.7M on Monday and +$5.4M on Tuesday did the week's work, September 30 brought the first daily net outflow in weeks at -$11.1M (Bitwise's BSOL led redemptions at -$8.9M while Fidelity's FSOL took in +$2.8M), October 1 lost another $5.9M, and October 2's +$1.3M was all that kept the week positive. A second tally on different settlement conventions, Farside Investors, puts the week at +$0.8 million with a -$12.5M September 30 print; this site leads with SoSoValue and states the spread rather than picking a winner. Sources: SoSoValue daily flow data via TradingNews and 24/7 Wall St; Farside Investors SOL ETF page.

Context makes the number less alarming than it looks - and the composition makes it more interesting. Bitcoin ETFs took in $241 million the same week, XRP funds saw weekly inflows fall 94%, and the Solana complex's cumulative net inflows stand at $1.61 billion with net assets around $1.9 billion - about 2.7% of Solana's market value, roughly 15.8 million SOL. For 2026 the funds have taken in $842 million, including $480 million in the third quarter and $272 million in September alone. Concentration remains the structural fact: one product, BSOL, accounts for about $1.2 billion of the $1.6 billion lifetime total - roughly 75% - and it is a staking fund, which means its flows are a yield product as much as a directional bet. A flat week in a yield-driven complex reads more like quarter-end rebalancing than a verdict.

My take: A 98.7% collapse in weekly flow that keeps a streak alive is a genuinely ambiguous object, and I refuse to pretend it is clearly one thing or the other. The bear reading: the record week was the anomaly, quarter-end trimmed it, and new money is waiting. The bull reading: this is a yield vehicle (one product holds roughly 75% of lifetime flows) churning with the rate calendar, exactly as its structure predicts. The number that resolves it is not this week's - it is whether the fifteenth week goes negative. One red day taught us nothing; a red week will.

3. SOL closed flat-to-down at $119.60 - the first decline in three weeks

The daily-close series (DefiLlama, each point the price at 00:00 UTC of the labelled date) stepped from $121.96 on Monday to $118.82 on Tuesday (-2.6%), wobbled sideways through $119.09 and $118.04, then printed three consecutive gains into Sunday: $118.38, $118.59, $119.60. That is a 1.9% weekly decline on the rounded figures, a market cap of roughly $70.5 billion on 589.1 million circulating SOL (CoinMarketCap), and a position 59.4% below the $294.33 all-time high of January 19, 2025. Against the majors, SOL lagged by 2.3 percentage points on Bitcoin (+0.4% for the week) and 1.9 points on Ethereum (flat) - the first week since early September in which it trailed both. Source: DefiLlama daily series.

The shape matters more than the sign. Intraday, the week tested the $122-124 zone twice without closing above it and held the $117 floor, per TradingNews's technical read - and on this site's daily-close series the low close ($118.04, Thursday) sat $6.9 above the prior week's $111.15 open. A week that gives back 1.9% after +11.8% and +9.2% is consolidation, and consolidation that holds three-quarters of its gains while the ETF channel goes quiet is the constructive version. The reading flips only on a weekly close below the $117-118 band, which would put $111.15 - the prior week's open - back in play.

My take: This is the week the market showed me which series are load-bearing. The entire week traded inside a $3.92 band - 64% tighter than the week before - the low close held nearly $7 above the prior week's open, and three rising closes ended the week while the ETF channel was printing its two outflow days. Spot demand absorbed the absence of institutional flow without flinching. The part I take most seriously is the relative flip - lagging both majors for the first time in five weeks - because leadership changes announce themselves in flat weeks, not in trending ones.

4. Q3 in the books: a sixty-percent quarter, record activity, and a $180 million app-revenue month

The quarter closed on September 30 with SOL up roughly 60% - from about $74 at the end of June to about $118 - though some outlets quoting from early July print closer to 48%. September added about 14.6% and August roughly 41%, per India Crypto Research's quarterly accounting; on this site's daily-close series the September calendar month ran $103.01 to $119.09, +15.6%. The comparison set for the quarter: Ethereum gained about 71% and Bitcoin about 43%, so SOL's quarter was strong in absolute terms and mid-pack against its majors - a fact the SOL/ETH ratio's roughly 6% quarterly decline makes visible. Source: India Crypto Research, October 5.

The activity records behind the price are the quarter's real ledger. Q3 carried an estimated 14.2 billion non-vote transactions on Solana, a quarterly record (India Crypto Research). September's application-layer revenue reached about $180 million - the strongest month since the start of the year and roughly 32% of aggregate blockchain app revenue, per MoneyCheck's September tally, with a single-day print near $8 million on September 11. And cumulative DEX activity on Solana crossed $3 trillion all-time (MoneyCheck). These are attributed third-party figures rather than this site's own series, and they are flagged as such - but they all point one direction: the quarter's price gains were invoiced by usage, not just by flows.

My take: I distrust quarterly superlatives, so let me be precise about which records I am endorsing. The ~60% quarter and the 14.2B non-vote transactions are attributed third-party figures, and I print them as such. What I can verify on my own series is the September fee total - $416 million, up 33.1% on August, with the fee-to-volume ratio rising all month - and fees are paid transactions, the least gameable series on the chain. The price did not outrun the usage in Q3. That is the quarter's real ledger.

5. Stablecoins: USDC printed a series high and the majors took the wheel back

Total stablecoin supply on Solana rose 0.5% to $16.82 billion at the week-ending snapshot - a record on this site's DefiLlama basis, and the second consecutive weekly record. The composition moved more than the total: USDC gained 1.8% to $7.43 billion, its highest print in the eight weeks this site tracks, while USDT jumped 6.4% to $2.84 billion, a second straight weekly gain after the September contraction. The third bucket - every issuer beyond the two majors - fell 3.2% to $6.55 billion, taking its share from 40.4% to 38.9%. Source: DefiLlama stablecoin charts for Solana.

The composition swing is the story. Last week's record was third-bucket-led - diversification away from a two-issuer structure; this week's is majors-led, with the two largest issuers adding a combined $300 million while the long tail shrank. A swing like that inside a flat total is repositioning, not growth - and majors-led repositioning historically skews toward trading and settlement flow rather than retail payments. The series to watch is USDT: one strong week after three of decline does not make a trend, but if it repeats, the dollar layer's shape has changed for the second time in a month.

My take: The composition swing is more interesting than the record. Last week the dollar base diversified - the third bucket hit a 40.4% share; this week the majors took the wheel back, with USDT bouncing 6.4%, a second straight weekly gain while the tail shrank. I will not invent intent from a supply series, but the shape fits a specific story: repositioning for trading and settlement ahead of the next tape, not retail payments flow. The discriminating datapoint is USDT's follow-through - a second consecutive gain and the September composition narrative has a live counter-story.

6. DEX volume fell a third time in four weeks - and the base was restated upward again

DEX volume totaled $16.28 billion for the week, down 11.1% - but the denominator moved too: DefiLlama has since restated the prior week to $18.32 billion, up from the $17.48 billion published a week ago. Against that published vintage the decline reads -6.9% instead. Either way it is the third weekly decline in four, and the daily path shows where the air is: weekday prints of $2.49-2.76 billion stepped down to $1.55 billion on Saturday and $1.71 billion on Sunday, a genuine weekend churn dip rather than a reporting gap. For the longer arc, cumulative DEX activity on Solana is now past $3 trillion all-time (MoneyCheck). Source: DefiLlama DEX dashboard.

The honest frame is the one this site has run since September: the DEX series is the softest currency on the chain. Roughly 58% of Solana DEX volume is bot-like flow by Bitquery's estimate, one aggregator has already demonstrated that its volume line can move independently of its fee line, and the source itself restates weeks retroactively - this is the fourth consecutive week with a revision in the tape. What makes this week's decline real anyway is the weekend: $1.55 billion on a Saturday is not a measurement artifact, it is fewer humans trading. The number that answers whether that matters is in the next story.

My take: For the fourth consecutive week there is a revision inside the DEX tape - the prior week was restated upward to $18.32B - so I want to be careful about what I call real. What I call real this week is the weekend: $1.55B on a Saturday and $1.71B on a Sunday are not measurement artifacts, they are fewer humans trading. That is the honest part of the decline. The cumulative $3T all-time figure (attributed) is the longer arc, but it does not change what a quiet Saturday means.

7. The fee line refused to follow volume down - and the fee-to-volume ratio hit a series high

Network fees came in at $111.67 million, down just 1.0% against a volume series that fell 11.1%, and Jito MEV tips rose 1.4% to $2.91 million. The daily fee band stayed inside its recent range - $14.68M to $17.40M on weekdays, $12.93M on Saturday, $16.23M on Sunday - with no day collapsing the way the DEX dailies did. Dividing fees by volume gives a fee-to-volume ratio of roughly 0.69% ($111.67M against $16.28B), the highest of the eight weeks this site tracks, up from about 0.62% the week before on the same vintage. Source: DefiLlama fees and DEX dashboards.

A rising fee-to-volume ratio in a falling-volume week means each dollar of measured volume carried more paid priority - more adversarial flow, more urgent transactions, or a mix shift toward small urgent tickets. Whatever the mechanism, it is the opposite of the signature a dying chain produces, and it is why this site's standing practice is to lead with fees and tips and quote volume with caveats. The discipline cuts both ways: if next week's fee line follows volume down instead, the 'fees are the harder currency' argument loses its streak, and the activity question reopens in earnest.

My take: This is my favorite datapoint of the week and the one I would build the month's read on: fees held within 1% of the series high while volume fell 11%, pushing the fee-to-volume ratio to roughly 0.69% - the highest of the eight weeks I track, and the second straight week the ratio has risen. Paid priority cannot be double-counted the way routed volume can. The discipline cuts both ways, though: if next week's fee line follows volume down, this entire framework loses its streak, and I will have to say so.

8. From the adoption file: a bank-grade token goes live, treasuries compound, and tokenized stocks printed a $4.4B week

Five items, each small, none noise. Fiserv - the payments giant - moved its Roughrider Coin to Solana, where it went live (India Crypto Research, October 5). DeFi Development Corp's treasury passed 2.5 million SOL, up about 10% in six weeks, while Forward Industries holds on the order of 6.9 million SOL - the corporate SOL treasury track keeps accumulating through the quiet tape (same source). Kamino opened a GPU-loan-linked sUSDai collateral market letting holders borrow USDC at up to 80% loan-to-value. Metaplex unveiled Vantage, a customizable trading terminal. Tetra Digital launched CADD, a Canadian-dollar stablecoin with Solana support (TradingNews). And tokenized-stock trading on Solana set a weekly volume record around $4.4 billion (TradingNews).

The pattern across the file is breadth without a headline: a payments company, two treasuries, a lending market, an infrastructure launch, a foreign-currency stablecoin, and a record in the tokenized-equity complex - all landing in the same five sessions as the quietest ETF week since June. That is what a base-building week looks like underneath: the marginal buyer steps back, the marginal builder does not.

My take: The breadth is the signal in the adoption file. A payments giant's token went live, two treasuries compounded, a lending market opened, an infrastructure launch shipped, a foreign-currency stablecoin arrived, and tokenized stocks printed a record $4.4B week - all inside the quietest ETF week since June. The marginal buyer stepped back and the marginal builder did not. Base-building weeks look exactly like this underneath, and they are why I read the flat tape as digestion rather than exit.

Three stories that mattered most

Eight items make a long list, so here is where I would spend attention if I could only follow three:

  1. The ETF margin (item 2). A fourteen-week streak that survived on a +$1.3M Friday is either a yield vehicle doing what yield vehicles do, or a demand channel rolling over - and the fifteenth week is the experiment that tells them apart.
  2. The fee-to-volume ratio at its eight-week high (items 6-7). Volume fell 11% and fees refused to follow; paid urgency is growing while churn shrinks. If that divergence survives a second soft-volume week, it becomes the strongest usage evidence on the board.
  3. The consensus discipline (item 1). September 28 passed without Alpenglow, without damage, and without the market having priced the phantom - a free lesson in what activation evidence looks like. The next real signal will be a governance epoch or a Firedancer Votor ship, not a calendar entry.

What I'm watching next week

A flat week hands the next one a clean scoreboard: every question below has a number that resolves it.

Source: DefiLlama - Solana prices, TVL, DEX volume, fees, stablecoinsSoSoValue - Solana ETF daily net flows, week of Sep 28 - Oct 2, 2026Farside Investors - Solana ETF flow table (second tally, settlement-convention differences)TradingNews - ETF by-fund breakdown, tokenized-stock record, levels and macro tape (Oct 2026)India Crypto Research - weekly review: phantom Alpenglow date, Q3 close, DvP launch (Oct 2026)24/7 Wall St via Yahoo Finance - $2.4M ETF week, 2026 YTD and September totals (Oct 2026)MoneyCheck - September app revenue $180M, DEX cumulative past $3T (Oct 2026)Solana Developers on X - Alpenglow whiteboard primer with Wattenhofer and Creech (Sep 28, 2026)CoinMarketCap - SOL circulating supply, 24h volume

Did Alpenglow go live on Solana mainnet on September 28?

No. The September 28 date is the tentative point when mainnet resumes feature activation under Anza's Agave 4.3 schedule - a routine runtime upgrade window that does not reference Alpenglow anywhere. Alpenglow (SIMD-0326) completed its testnet activation on September 24 and devnet on September 25; mainnet has no announced date, and Anza's Roger Wattenhofer summarized the position as 'No Alpenrush'.

How much did Solana ETFs take in this week?

$2.43 million of net inflows across September 28 - October 2 on the SoSoValue tally, down 98.7% from the prior week's record $188.21 million. A second tally with different settlement conventions (Farside) puts it at $0.8 million. Either way it is the fourteenth consecutive week of net inflows - by the narrowest margin yet, saved by a +$1.3M Friday.

Why did SOL fall this week while everything else was mixed?

It barely fell: -1.9% on the daily-close series, against a Bitcoin that gained 0.4% and a flat Ethereum. The decline is consolidation after +11.8% and +9.2% weeks, and the low close of $118.04 held $6.9 above the prior week's $111.15 open. The more notable shift is relative: SOL lagged both majors for the first week since early September.

Why does the fee-to-volume ratio matter so much on this site?

Because it separates measured activity from measurement artifacts. This week's ratio of roughly 0.69% ($111.67M fees on $16.28B volume) is the highest of the eight weeks tracked: fees held within 1% of a series high while volume fell 11%. Volume can be double-counted or redefined - routed and bot flow especially - but paid fees cannot, which is why a rising ratio in a falling-volume week is read as resilient real demand.

What is the single most important thing to watch next week?

Whether the ETF streak reaches fifteen weeks. The complex's flows are the only demand series that has grown every week since late June, and the fourteenth survived by $1.3 million. A first weekly outflow would mark the end of that layer's uninterrupted growth - not the end of the institutional case, but the end of the part of it that never had to be re-argued.

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.

Tags: SOL ETFStablecoinNetworkDeFi