Forced Lightning channel closes cost a median 925 sats in 2026, preprint finds
In brief
- Lightning force closes in a 2026 public-channel sample cost a median 925 satoshis, per Kumar's preprint.
- Broadcasters waited a median 2.5 days before they could sweep their own balances.
- Cooperative closes cost a median 197 satoshis in 2026, a 4.7-to-one gap.
- Kumar's cost cohort covers 14,435 public force closes through October 4; the preprint isn't peer-reviewed.
What the preprint measured
Kumar reported 973,581 channel closes from January 2018 through October 31, 2026. Of those, 523,815 were force closes. The detailed 2026 cost cohort is narrower (14,435 public force closes through October 4), which left four weeks to follow their outputs.
The fee total covers more than one transaction. It includes the commitment transaction, required balance sweeps, pending-payment resolution, owner fee bumps seen in the same block and later owner anchor cleanup. Fees for batched spending were split among inputs. Private cooperative closes and unclassified simple-taproot spending details aren't counted.
Cooperative closes cost less. The paper put the 2026 cooperative-close median at 197 satoshis, which puts the force-close median 4.7 times higher.
Where the wait comes from
A unilateral close settles a channel on Bitcoin with a commitment transaction that was signed in advance. Under Lightning's BOLT 3 specification, the output that returns the broadcaster's own balance carries a delay called to_self_delay. The delay gives the peer time to penalize a broadcaster that submits a revoked channel state.
That's the source of the 2.5 days.
The scope matters here. The figure measures the wait until the balance becomes eligible to be swept. It doesn't measure when funds reach a wallet, and it says nothing about how fast ordinary Lightning payments are. Across the whole study period, the reported median wait was 1.7 days.
A cheap exit still ties up capital
Kumar compares the two costs with a break-even annual return (the rate at which the broadcaster's waiting cost on its locked balance would equal the force-close fees). By his calculation, that rate was 1.2% in 2026, compared with 16.7% over the full period. It's an illustrative number. It compares per-channel fees for both parties with the broadcaster's locked balance, leaves out pending-payment amounts and doesn't assume operators actually earned that return.
CryptoSlate's point is that a cheap exit can still leave capital unavailable for days, and the fee alone doesn't show how much usable liquidity is left.
BOLT 3's zero-fee commitment design works differently on fees. A child transaction pays the commitment's mining fee through child-pays-for-parent, and transactions that resolve pending payments (HTLCs) pay their fees by adding inputs. The broadcaster's delayed balance remains part of that design. That's protocol background, though, and the 2026 cohort didn't test it.
Frequently asked questions
Why does a forced Lightning close make the broadcaster wait?
A unilateral close settles the channel on Bitcoin with a commitment transaction that was signed in advance. Under BOLT 3, the output that returns the broadcaster's own balance carries a delay called to_self_delay. The delay gives the peer time to penalize a broadcaster that submits a revoked channel state.
How much more does a forced Lightning close cost than a cooperative one?
Kumar's preprint reports a 2026 median of 925 satoshis per channel for public forced closes, counting both parties. The 2026 cooperative-close median was 197 satoshis. That makes the force-close median 4.7 times higher. The paper hasn't been peer-reviewed.
Does the 2.5-day wait mean funds take 2.5 days to reach a wallet?
No. The 2.5-day median measures the wait until the broadcaster's balance becomes eligible to be swept. It doesn't measure when funds reach a wallet or how fast ordinary Lightning payments are. Across the whole study period, the median wait was 1.7 days.
