DOGE Mining and Fee Market — September 2026 Monthly Report

September in the mining and fee market

September was a month with two loud days and a quiet month around them — and, underneath the quiet, the first sustained move in Dogecoin's security margin since this site began measuring it. The price closed at $0.09457, up 14.1% from August's $0.0829, on the strength of a two-session squeeze that printed the first $0.10 daily close in our records. The chain, meanwhile, produced between 1,349 and 1,376 blocks every day, issued the same 10,000 DOGE per block, charged fees through every regime without a single congestion event — and saw its day-end difficulty fall 27% from the month's mid-point peak, the signal that October inherits. This report aggregates the month's weekly prints into one view: the price shape, the usage response, the mining economics, the security margin, the institutional ledger, and what the final week handed to October.

Price and the shape of the month

Dogecoin opened September at $0.0829 (the August 31 close) and spent the first two weeks going lower, not nowhere: the month's worst close printed at $0.08007 on September 15, with the intraday floor at $0.07835 on September 16. Then two sessions repriced the month: +14.3% on Monday, September 21 (a short squeeze worth $5.66M that day, per our weekly report's sourced accounting) and a +$0.6% Tuesday that closed at $0.10042 — the first daily close above $0.10 in the dataset — after trading to an intraday high of $0.10589. Wednesday gave 7.7% back, and the month's final week drifted between $0.0928 and $0.0969 before closing at $0.09457 on September 30.

Weekly changes are computed on unrounded closes (Gate.io DOGE/USDT daily candles). The month's shape in one line: a two-week slide, a two-day squeeze, a one-day fade, and a week of plateau.

The two sessions that made the month

The squeeze's anatomy, from our September 21–27 report: roughly $5.66 million of short liquidations on the Monday, about $12.7 million across the move (Yahoo Finance's accounting), open interest at a one-month high, and — the institutional echo — a record $2.89 million ETF week (SoSoValue), the best since the complex launched in November 2025. Usage spiked with the tape: 51,563 active addresses on the Monday, the highest single-day count in our dataset. None of it was a fee event or a usage event in the durable sense — by the month's end, both series had round-tripped, and the transaction series finished below where it started.

For miners, the ordering matters: the impulse was mechanical (liquidations), the echo was institutional (funds), and the production side of the network was unchanged through all of it. The mint does not know what price did.

Activity: usage followed the tape up for a day, and down for a month

The month's on-chain series told a blunt story. Daily transactions averaged 24,079 for September (722,368 in total) — down 15.7% from August's 28,586 — and the decline was a staircase, not a cliff: 31,562 (the August-peak week) → 26,865 → 25,807 → 25,503 → 23,660, with the squeeze week's 24,106 a one-week pause before the month's final days broke the series open (September 29 printed the month's quietest day at 17,830). Active addresses averaged 37,421, down 3.5% from August, with the squeeze Monday's 51,563 the month's — and the dataset's — single-day high.

The scale check, stated once for the month: a $14.77B month-end market cap over 24,079 daily transactions is roughly $613,000 of market value per daily transaction. September sharpened August's finding: usage in this asset is a high-beta companion of price, not an independent adoption series — and the beta is getting more expensive, because the price held its gains while the activity did not.

Mining economics: a fixed line, marked to market

The production side of September: the chain produced an average of 1,363 blocks per day (CoinMetrics BlkCnt; daily range 1,349–1,376), an implied block time near 63.4 seconds against the 60-second target, issuing 408.8M DOGE over the month — worth approximately $36.4 million valued day by day at the Gate.io closes, or about $1.21M per day on average. The monthly arc in dollar revenue, marked at each week's close:

Month over month, the two halves of the model: coin issuance fell 3.4% (408.8M vs August's 423.1M — the block count is never fixed), while dollar revenue rose 12.0% ($36.4M vs $32.5M) because the close rose 14.1%. Annualized at the month-end close, the subsidy flow is worth on the order of $470M — roughly 4.97 billion DOGE per year, about 3.18% supply inflation against 156.1B outstanding, no maximum supply, no halvings. The supply series confirms the issuance arithmetic to the coin: 40,877 blocks × 10,000 DOGE = exactly the 408.77M DOGE CoinMetrics shows added between August 31 and September 30.

The chain clock — and the month the margin started moving

The clock was September's steadiest series: 1,363 blocks a day, never below 1,349 or above 1,376, self-correcting through difficulty exactly as designed. But the difficulty underneath the clock told the month's under-reported story. Using the difficulty of the last block of each day (each print individually verifiable on Blockchair's block explorer):

Difficulty built through mid-September (51.62M on September 13), then fell 26.9% by month-end — a decline that began before the squeeze and continued through it. The final week of September handed October a moving series: the month closed at 37.74M, and our October reports measure the path from there. Two transparency notes. First, we publish day-end prints rather than a blended daily average because each print is individually checkable; a single snapshot in this regime can mislead — Dogecoin retargets every block, and single days have printed ranges from 26.8M to 70.2M. Second, we do not attribute the cause: merged hash follows Scrypt economics across Dogecoin and Litecoin, and no source we can verify explains the drift. We measure; attribution requires evidence.

Expressed as hashrate — implied, not sampled: day-end difficulty × 2³² ÷ block time — the month's security margin ran from roughly 3.48 PH/s at the September 6 day-end print to a mid-month peak near 3.50 (September 13), then slid to 2.56 by September 30: a 26.5% compression that matches the difficulty path because block times barely moved (1,363 blocks a day all month). The derivation reproduces the explorer's published snapshot methodology to within 1% on the print we can check it against, and every input — each day-end difficulty, the block count — sits inside the month and is individually checkable. For the month's miners, that is the under-reported series in one line: the pool of work per block compressed by more than a quarter while the clock never noticed.

The structure it clocks: pure Scrypt proof-of-work, merge-mined with Litecoin, no staking, no validator set. What the month added: the first measurable stress test of the margin — and the schedule did not notice.

The fee market: congestion-free through every regime

This edition's fee ledger runs on CoinMetrics' daily FeeTotNtv — a series checkable for every day of the month, replacing the point-in-time snapshot prints of our weekly editions. September's miners collected 272,558 DOGE in fees across the month, about 9,085 DOGE per day — an average of 0.3773 DOGE per transfer at the month's activity, worth on the order of $810 per day against ~$1.21M of daily subsidy: roughly 0.07%. The distribution was flat: the busiest fee day of the month was the squeeze Monday itself (September 21, 11,840 DOGE), the quietest was the month's final Sunday (September 27, 7,364), and no day printed anything resembling congestion — the block clock held 1,349–1,376 blocks a day all month. Fees, like everything else in this report, were a design constant wearing a market's clothes: usage rose and fell, the squeeze came and went, and the fee line barely moved.

The institutional ledger: the best month since January

The ETF complex's month, week by week (SoSoValue, via the coverage cited in our weekly reports):

September's net: +$3.71 million — the complex's best month since January 2026. The composition is as telling as the total: Grayscale's GDOG took the bulk (its assets ended the window near $14.11M against $16.34M cumulative inflows), 21Shares' TDOG ran −$593,000 for the month, and Bitwise fixed the exit dates for BWOW (about $801,400 in assets): final trades October 14, cash distribution October 22 — the channel goes from three funds to two. The verdict echoes August's, with better numbers: the channel flipped from absent to its best month ever and remains under 0.11% of the asset — the record week equaled roughly two days of issuance at the new price. Scale, not sentiment, is the finding.

What a DOGE fee buys — and the cheaper rail

Dollar for dollar, Dogecoin remains one of the cheapest transfers in crypto: an average fee of about $0.034 (September average, CoinMetrics FeeTotNtv ÷ transactions) to move value in about a minute. The honest comparison for stablecoin movers is against TRON, where a USDT transfer costs about $0.20 in TRX bandwidth when paid directly — or about $0.02 using rented energy:

That roughly 90% saving on TRON is the entire proposition behind Tronsell's energy rental — same transfer, same finality, a fraction of the cost. We publish the comparison because a fee report that only quotes one chain is marketing, not data.

What October inherits

Five dated markers, in order. The difficulty path: a month-end print of 37.74M, 26.9% under the September 13 peak, with a cause we cannot source — the day-end path is now a weekly line item in our data report. October 14: BWOW's last trading day; the complex becomes two funds. October 22: BWOW's final cash distribution. The $0.10 line: September's high ($0.10589) and first close above it ($0.10042) sit 4–10% above the September 30 close of $0.09457; a monthly close above $0.10 would complete a two-month structure. DogeOS mainnet: the testnet launched September 30 with no mainnet date — the first dated check the ecosystem has given our transaction series. Beneath all five sits the month's most durable finding: a network whose clock, issuance and fees did not vary by a measurable amount through a 14% price round trip, a 43% usage decline from peak, and the first security-margin stress test we have recorded.