Lightning in September 2026 Is Unmeasured: mempool.space's Stats Froze on 8 July at 4,532 BTC, Amboss Now Shows 3,781 BTC (-34% Since January), the Newest Volume Figure Is Still River's November 2025 $1.17B, and USDT-on-Lightning Has No Public Number
The cleanest thing that can be said about the Lightning Network in September 2026 is that nobody can currently measure it well. The tracker most articles quote, mempool.space, has not published a Lightning statistics row since 8 July 2026: its latest endpoint and its three-month series both return that single date, showing 40,588 channels and 4,531.6 BTC of public capacity. Amboss's public GraphQL API, queried on 22 September 2026, shows 36,084 channels and 3,780.8 BTC (about $326M at $86,300). That is 16.6% below mempool's frozen July figure and 34% below the 5,742.9 BTC mempool recorded on 1 January 2026. 1ML, queried the same day, sees only 19,908 channels and 2,663 BTC and reports capacity down 1.7% and channels down 2.0% over the trailing 30 days. The 30% gap between Amboss and 1ML is not rounding: the trackers crawl different gossip views and prune dead channels on different schedules, and none of them has published a reconciliation.
The stablecoin story is where the measurement gap bites hardest. Tether's USDT went live on Lightning via Taproot Assets on 21 March 2026, and Lightning Labs shipped tapd v0.8 with a client SDK on 23 June. Neither event has produced a single public volume, channel or supply number. Taproot Asset channels are ordinary Taproot outputs whose asset contents never appear in Lightning gossip, so none of the three trackers above can see them, and the v0.8 announcement itself contains no adoption figures at all. The one wallet publishing numbers, Speed, describes its USDT-L as a wrapped token backed by Ethereum USDT held in multisig rather than native Tether issuance, and its page as displayed on 22 September claims 15,300 nodes, 51,000 channels and $300M of capacity, a channel count higher than any tracker has ever recorded for the whole public network. Treat it as marketing copy, not telemetry.
Payment volume has the same problem in slower motion. The most recent network-wide estimate anyone can cite is still River's: $1.17 billion across 5.22 million transactions in November 2025, an average ticket of $223 versus $118 a year earlier, with dollar volume roughly 4x from $286M while transaction count fell from the 6.6 million peak of August 2023. River published that in February 2026 and no operator has released a 2026 monthly figure since. Every "north of $1B a month" line published this year is that same ten-month-old data point being re-quoted, usually without the date.
Put together: measured public capacity is down roughly a third from its December-January peak on every tracker still updating, channels are drifting down about 2% a month on 1ML, and the growth that is being claimed (USDT, larger tickets, exchange rails) is happening either in private channels or in asset channels that are structurally invisible. That may be perfectly fine for users. It is not fine for anyone quoting network health numbers as if they were audited.
What to watch: whether mempool.space resumes its Lightning series and whether the resumed figure lands near Amboss's 3,800 BTC or the old 4,500; a second River-style volume release covering the first half of 2026, since November 2025 is the last real datapoint; any Taproot Assets explorer or Tether disclosure of USDT-on-Lightning supply, which is the only way the stablecoin claims become auditable; and 1ML's 30-day capacity change turning positive, which would be the first tracker-visible sign that the drawdown has ended.
Sources (6)
AI Research
Key Takeaway
Every Lightning tracker still updating shows public capacity down about a third from its January 2026 peak, the most-quoted tracker has been frozen since 8 July, the newest payment-volume figure is ten months old, and USDT-on-Lightning is structurally invisible to all of them.
The cleanest thing that can be said about the Lightning Network in September 2026 is that nobody can currently measure it well. The tracker most articles quote, mempool.space, has not published a Lightning statistics row since 8 July 2026: its latest endpoint and its three-month series both return that single date, showing 40,588 channels and 4,531.6 BTC of public capacity. Amboss's public GraphQL API, queried on 22 September 2026, shows 36,084 channels and 3,780.8 BTC (about $326M at $86,300). That is 16.6% below mempool's frozen July figure and 34% below the 5,742.9 BTC mempool recorded on 1 January 2026. 1ML, queried the same day, sees only 19,908 channels and 2,663 BTC and reports capacity down 1.7% and channels down 2.0% over the trailing 30 days. The 30% gap between Amboss and 1ML is not rounding: the trackers crawl different gossip views and prune dead channels on different schedules, and none of them has published a reconciliation.
The stablecoin story is where the measurement gap bites hardest. Tether's USDT went live on Lightning via Taproot Assets on 21 March 2026, and Lightning Labs shipped tapd v0.8 with a client SDK on 23 June. Neither event has produced a single public volume, channel or supply number. Taproot Asset channels are ordinary Taproot outputs whose asset contents never appear in Lightning gossip, so none of the three trackers above can see them, and the v0.8 announcement itself contains no adoption figures at all. The one wallet publishing numbers, Speed, describes its USDT-L as a wrapped token backed by Ethereum USDT held in multisig rather than native Tether issuance, and its page as displayed on 22 September claims 15,300 nodes, 51,000 channels and $300M of capacity, a channel count higher than any tracker has ever recorded for the whole public network. Treat it as marketing copy, not telemetry.
Payment volume has the same problem in slower motion. The most recent network-wide estimate anyone can cite is still River's: $1.17 billion across 5.22 million transactions in November 2025, an average ticket of $223 versus $118 a year earlier, with dollar volume roughly 4x from $286M while transaction count fell from the 6.6 million peak of August 2023. River published that in February 2026 and no operator has released a 2026 monthly figure since. Every "north of $1B a month" line published this year is that same ten-month-old data point being re-quoted, usually without the date.
Put together: measured public capacity is down roughly a third from its December-January peak on every tracker still updating, channels are drifting down about 2% a month on 1ML, and the growth that is being claimed (USDT, larger tickets, exchange rails) is happening either in private channels or in asset channels that are structurally invisible. That may be perfectly fine for users. It is not fine for anyone quoting network health numbers as if they were audited.
What to watch: whether mempool.space resumes its Lightning series and whether the resumed figure lands near Amboss's 3,800 BTC or the old 4,500; a second River-style volume release covering the first half of 2026, since November 2025 is the last real datapoint; any Taproot Assets explorer or Tether disclosure of USDT-on-Lightning supply, which is the only way the stablecoin claims become auditable; and 1ML's 30-day capacity change turning positive, which would be the first tracker-visible sign that the drawdown has ended.