Dogecoin Weekly Data Report — September 28–October 4, 2026

1. What changed this week

Three things changed, and for once the price is the least of them. First, usage broke its floor: daily transactions averaged 17,860, down 25.9% — the lowest weekly average in our records, and far below the 23,000 threshold we published two weeks ago as the reclassification line for September's squeeze-churn read. Second, the security margin cracked: day-end difficulty fell from 40.20 million to as low as 28.93 million (October 2) before recovering to 32.89 million — 18.2% week over week, with implied hashrate down 18.1% — against a price that moved 1.0%. Third, the price itself: DOGE closed at $0.09592, −1.0%, in the quietest trading week since our records began. The disconnect between a flat price and falling usage-and-hash is this report's subject.

Context against the nine weeks we track: the two-week uptrend is over. After W38 (+6.0%) and W39 (+11.0%) printed the first back-to-back green weeks of the dataset, W40's −1.0% leaves the close at $0.09592 — 1.0% below the W39 pivot of $0.0969, and 4.1% below the $0.10 line that one September Tuesday closed above. Nothing about the week was violent on price; both series underneath it were.

2. The dashboard

Every figure above is a 7-day average or a window aggregate computed from the raw daily series — never a Sunday print standing in for the week. Sources and derivations are in the appendix.

3. Price: seven sessions inside a 3.4% band

The week's path, day by day, from our own close series:

Two readings matter. First, the close equals the week's best close — there is no fade-gap this week; the Sunday session did the week's work. Second, the close sits 2.2% below the intraday high ($0.0981, September 30) and 4.1% below $0.10: the breakout line from last week is now a level price is drifting away from, not testing.

4. Usage: the floor gives way

Daily transactions averaged 17,860, down 25.9% from 24,106 — breaching the 23,000 threshold we published on September 27 as the reclassification line. Active addresses averaged 34,985, down 7.6%. Both are now the lowest weekly averages in our dataset, which begins July 24. The full series:

The series continues below. The composition distinguishes this week from September's churn: there was no spike day carrying a price event's timestamp — the decline ran all week, from 20,348 on Monday to 16,020 on Sunday. Last week's +1.9% turn, which we scored as 'turned, not recovered', has fully unwound; the series is 43.4% below its August peak of 31,562. The squeeze-churn reclassification we pre-committed two weeks ago has now triggered, and the honest question shifts from 'when does usage recover' to 'where is the floor'.

The fee line adds one cross-check the collapse passed: miners' total fee intake fell only about 15% in DOGE terms (roughly 8,086 DOGE a day against 9,551) even as transactions fell 25.9%, because the average fee per transfer rose 14.3% (CoinMetrics FeeTotNtv ÷ TxCnt, both weeks same-caliber). A data glitch that deleted transactions would drag the fee series down with it; a market that stopped transacting leaves the per-transfer fee intact while the count falls. The two series moved exactly as a real pullback moves — together in dollars, apart per transfer — which is the kind of internal consistency a fabricated break could not fake.

5. Issuance and supply

The chain produced 1,363 blocks per day on average (9,538 blocks across the week), issuing 13.63M DOGE per day — worth approximately $1.31 million at the $0.09592 close. The supply series confirms the arithmetic to the coin: CoinMetrics puts supply at 156,190,892,866 DOGE at the window's end, an increase of exactly 95.38 million DOGE — 9,538 blocks × 10,000 DOGE, the fixed per-block subsidy. We cross-checked the block count independently against the chain itself: Blockchair's day-end block heights run 6,392,301 (September 27) to 6,401,839 (October 4) — a difference of 9,538. Two sources, one number.

The DOGE-denominated subsidy rose 0.2% (1,363 vs 1,360 blocks) while its dollar value fell 0.9% — the price did the work, as it always does. The security budget is price-led; the mint is not.

6. Hashrate and difficulty: the security margin, measured

This is the week's most consequential section, measured entirely inside the window. The primary series is the day-end difficulty — Dogecoin retargets every block, the September 28 blocks alone ranged from 26.80M to 70.15M within one day, and the difficulty of each day's last block is a dated, individually checkable print:

Three observations. First, magnitude: a day-end drawdown of 28.0% at the trough and 18.2% at the week's close, on a price that fell 1.0% — the security margin moved by an order of magnitude more than the market. Second, direction: the week ends with two consecutive rebound prints (October 3 and 4), lifting the day-end print +13.7% off the trough — the return leg has started inside the window, though on two prints it is young. Third, what did not move: the chain produced 1,363 blocks a day, an implied block time of 63.4 seconds against the 60-second target, within a whisker of last week's 63.5 — the clock held while the margin underneath it compressed. This trips the security-budget check we published last week ('a >2% single-week difficulty move against flat price') at fourteen times its threshold; the mechanics are this week's deep dive. The hashrate companion series is derived, not sampled: day-end difficulty × 2³² ÷ block time gives ~2.72 → ~2.23 PH/s (−18.1%), and applied to the September 28 snapshot difficulty the same formula yields 3.80 PH/s against the explorer's published 3.83 — so the derivation reproduces the snapshot methodology on inputs that never leave the window.

7. The fee ledger

The fee ledger switches this edition to CoinMetrics' daily total fees (FeeTotNtv) — a series checkable for every day of the window, replacing the explorer's point-in-time median print. Miners collected about 8,086 DOGE per day, an average of 0.4527 DOGE (~$0.0434) per transfer, worth on the order of $760 per day — roughly 0.06% of subsidy revenue. The per-transfer average rose 14.3% from the prior week's 0.3962 (same caliber) even as total fee income fell about 18% in dollars, because transactions collapsed 25.9% faster than per-transfer fees drifted up. Averages run above the median prints of prior editions because the fee distribution is skewed; the caliber is now consistent week to week. There was no congestion: the block clock never filled, holding 1,363 blocks a day at 63.4 seconds against the 60-second target. If congestion ever arrives, it shows up in the clock first; we will report it when it does.

8. Market value ratios, three ways

At the $0.09592 close and 156.1909B DOGE outstanding, the market values Dogecoin at $14.98 billion. Divided by the week's activity, that is $838,900 of market cap per daily transaction and $428,200 per daily active address. The transaction ratio jumped 33.7% this week — but read the mechanism before the number: the market cap fell 1.0% while the transaction denominator fell 25.9%. A valuation ratio expanding because its usage denominator shrinks is the unhealthy version of that move; last week's +9% came from the price side. The sensitivity, so no single activity assumption carries the read:

9. Scenario math: what $0.090–$0.107 does to the security budget

The band narrows this week to the levels in play — $0.090 as the defended floor, $0.09592 as the actual close, $0.100 as the line above, $0.107 as the analysts' zone from last week's coverage. Issuance is held at the observed 13.63M DOGE/day; only the price moves:

The invariants hold as always: the percentage column never moves — dilution is 3.18% a year at any price — and the security budget swings by $85 million between the bottom and top rows. After this week's difficulty drawdown, the price side of miner economics matters more than it did: a −1% week took the daily subsidy value to $1.31M while the hash pursuing it fell by more than a quarter before rebounding.

10. Volume: the quietest week since the series began

Gate.io turnover for the week totalled approximately $1.44 billion, down 46.5% from $2.70 billion — the lightest week in our records apart from the series' first ($1.16 billion, August 10–16), and less than half the squeeze week. The pullback is mechanical: last week's $2.70B carried the squeeze's liquidation churn, and nothing this week replaced it. The five-week arc — 1.96B, 2.70B, 1.44B — is the volume shadow of the price event this site keeps describing: a two-week flare, then quiet.

A composition note on the quiet: $1.44 billion across seven sessions is roughly $206 million a day, against roughly $385 million a day last week and $272 million a day in the series' first week. The quiet is therefore a level, not a one-day artifact — and it is the level against which every story in this week's news review should be read: two US-regulated derivatives on-ramps, a testnet launch and a fund exit all landed inside the lightest tape our dataset has measured.

11. What we still cannot source

Four gaps, one resolved this edition. Pool share: still no verifiable pool-hash distribution source at writing, so the table stays out. Per-day ETF flows: the week's prints carry a one-day attribution conflict between outlets (September 28 vs 29 for the same ~$879K), published as-is in the news review rather than averaged. Whale net positions: the 545M-DOGE accumulation claim circulating this week has no named tracker; the netflow series we actually score remains unsourced. Point-in-time snapshots: the explorer's median-fee and hashrate prints carry no dated history we can reproduce — which is why this report's fee ledger now runs on CoinMetrics' daily FeeTotNtv and its hashrate on the day-end-difficulty derivation, keeping every figure inside the report's window, and why it uses day-end block prints for difficulty — each one individually checkable against Blockchair's block explorer.

12. Appendix: data snapshot