NFT Weekly Intelligence — Sep 14–20, 2026

This is the weekly intelligence briefing for Sep 14–20, 2026. Organic sales were $37.54M while buyer addresses rebounded 174.04% to 114,977 — and for the first time in this series the capture's own comparison base matches ours to within three addresses.

1. Market: How Much Traded, and Who Participated?

Organic NFT sales were $37.54M for the week ending Sep 19, down 15.28% on the ~$44.31M prior-week base cited inside the Sep 19 capture and down 19.75% against our own Sep 12 capture of $46.78M. Transactions fell to 808,432. Buyer addresses rebounded 174.04% to 114,977 and seller addresses 151.72% to 108,037.

The two comparison bases closed from a 42% gap to 5.3% — the narrowest of this series — and the address columns now reconcile almost exactly between them. The capture's own +174.04% implies a prior buyer count of 41,956, against the 41,959 we recorded on Sep 12: a three-address difference. Seller addresses reconcile less tightly (+149.81% on our base against +151.72% reported). Value and transaction counts still do not: the capture's −9.08% implies about 889,000 trades against the 917,549 in our capture, a 3.1% base gap that leaves the two transaction columns 2.8 percentage points apart.

The per-unit figures carry the real information, and this week they unwound. Average value per transaction fell from $50.98 to $46.44, down 8.9%. Transactions per buyer address fell back to 7.0 from 21.9, and value per buyer address from about $1,115 to $327. Both of last week's anomalous ratios normalised in a single week, which is the strongest evidence yet that the 41,959 print was an attribution artefact rather than a market state. It is not a recovery: 114,977 buyer addresses is still 58% below the 273,655 recorded on Sep 5. The participant base stopped contracting; it has not come back.

2. Chains: Where Did Volume Sit, and How Much Was Wash?

Ethereum leads at $15.32M — 40.8% of organic volume — despite a 2.66% reported decline; on our Sep 12 base the fall computes to 9.0%. Polygon holds second at $7.09M and Bitcoin third at $4.33M after a 53.99% drop, the steepest of the six. BNB Chain is fourth at $2.58M, its second consecutive decline and about 92% below the level its own 1,042% week reached. Base was the only chain to grow, up 4.06% to $2.16M; Solana slipped 11.08% to $1.89M.

What separates this week from last is participation. Buyer addresses rose on all six named chains — Base 253.04%, BNB Chain 384.73%, Solana 164.90%, Bitcoin 141.29%, Polygon 91.47%, Ethereum 63.82%. A rebound across every chain is hard to read as market behaviour and easy to read as a metric reverting, which is the reading the base reconciliation in section 1 supports. Chain-level W/W percentages are as reported in the Sep 19 capture. Wash volume was published this week for Ethereum, Polygon and Base only; the remaining chains are shown without a figure rather than an estimate.

3. Blue Chips: Where Did Floors Land?

Four of five floors fell again and BAYC was the lone gainer — the second week running that a single blue-chip floor rose against a falling cohort, after Doodles was the sole gainer the prior week. BAYC added 0.91% as reported (2.24% between our two dated snapshots) to 6.85 ETH; CryptoPunks was next-most resilient at −0.74% reported (−0.34% snapshot) to 29.69 ETH. Doodles was weakest at −7.02%, with Pudgy Penguins at −6.29% and Azuki at −4.62%. ETH rose 1.73% over the window, from $2,514.85 to $2,558.32 — a second consecutive week of ETH-denominated floor declines through a rising unit of account, though against a 1.73% rally rather than last week's 6.94%.

Volume and floor direction now diverge inside the cohort. Doodles traded 52.96% more volume (52.57 ETH across 141 sales) into its weakest floor, while CryptoPunks traded 62.15% less (286.70 ETH) into its mildest decline and BAYC 48.94% less (206.90 ETH) into the only gain. Rising turnover against a falling floor is sellers meeting bids; falling turnover against a firm floor is an absence of trade rather than a sign of strength. The briefing's listed snapshots (CryptoPunks deepest at 1,105 listed, 11.1%) remain point-in-time inventory, not weekly velocity.

The briefing's USD floors imply an ETH price of about $2,549, marginally below the $2,558.32 endpoint it cites; we quote its USD figures as published rather than re-converting them. Implied cap = floor (USD) × supply. It assumes the entire supply could clear at the floor — an upper bound, not a valuation.

4. Wash Trading: How Much Reported Volume Was Organic?

Polygon's ratio snapped back to 2.55× — $18.07M of wash against $7.09M organic — one week after easing to 2.22%. Three of the four weeks tracked now sit at or above 2.5%, so the single dip did not retire the structural-wash label, and Polygon headline volume should still not be read as demand. Base's wash figure returned after a one-week gap at 2.22× — $4.80M against $2.16M organic, its widest ratio yet, and above organic in every week the figure has been published. Ethereum's wash share fell to 2.9% ($443,802 on $15.32M) from 4.7%.

Wash figures for Bitcoin, BNB Chain and Solana were not published in the sources we cite, a gap that has now persisted for two weeks. We show blanks rather than estimates, and the working rule is unchanged: rank chains on organic volume, treat Polygon's headline as unusable, and read a missing wash figure as unknown rather than zero.

5. Outlook: Last Week's Calls, Scored — and This Week's Watch Items

Last week's outlook made five falsifiable calls. Scoring them against the Sep 19 capture and the Sep 18 briefing:

1. Buyer addresses rebounding above 100,000 — happened, decisively. They printed 114,977. More important than the level is the reconciliation: the capture's own +174.04% implies a prior count of 41,956 against our 41,959, which retires the alternative reading that the 84.67% collapse was a definition change. The metric kept its meaning; one print moved — and both of last week's spiking ratios, 21.9 transactions per buyer and $1,115 per buyer, fell back to 7.0 and $327.
2. Ex-BNB stabilisation — did not happen. The market ex-BNB printed $34.96M against $42.76M, a break 18.2% below the ±10% band around $43M that would have confirmed a floor. Two flat weeks have given way to a decline, with BNB itself contributing a 36.26% fall inside a market that fell 15.28%.
3. Polygon wash below 2× — did not happen. The ratio snapped back to 2.55×, restoring the structural-wash reading that last week's 2.22× dip had put in question.
4. Solana's third growth week above $2.5M — did not happen. Solana slipped 11.08% to $1.89M. Its two growth weeks (11.17%, then 32.91%) are followed by a decline shallower than the market's but not a trend, so the OpenSea relaunch effect remains unmeasured rather than disproven.
5. A new SEA timeline — did not happen. OpenSea's token still carries no date and the silence has passed its sixth month; MetaMask's token repeated the pattern this week.

This week's falsifiable items for the Sep 26 capture:

1. Buyer addresses above 100,000 for a second week — would convert a one-week rebound into a level; a fall back toward 42,000 would reopen the artefact question.
2. A third consecutive week of falling sales — $37.54M makes two; a print below $35M would establish a downtrend rather than a normalisation.
3. Ex-BNB above $38M — a reclaim would mark the 18.2% break as a one-week event; below $34.96M extends it.
4. Polygon's wash ratio above 2.5× for a second week — would end the ambiguity left by the 2.22× dip and re-anchor the structural label.
5. Base's wash above organic holding — the figure has exceeded Base's organic volume in every week it was published (1.05×, 1.34×, 1.13×, then 2.22×); a fifth print would confirm it as structural.