Bitcoin SegWit Adoption

What it shows

SegWit adoption is the percentage of each day’s transactions that spend at least one SegWit input — a transaction that carries witness data. SegWit (Segregated Witness, activated August 2017) moved signature data into a separate witness structure, cutting fees and enabling Taproot and the Lightning Network. The higher this line, the more of the network’s activity comes from upgraded wallets and services.

It’s a network-health and ecosystem-maturity gauge, not a valuation metric. It climbed steadily from 0% at activation into the mid-90s% — most Bitcoin activity is SegWit today.

How it’s measured

We read the count of SegWit transactions per block directly from the Bitcoin blockchain (getblockstats’ swtxs field), divide by the block’s total transactions, and aggregate by UTC day. A transaction counts as SegWit when it carries witness data — which happens precisely when it spends at least one SegWit output. No third-party feed is involved; the number is recomputed from the chain each day.

Why the 2026 dip — Alkanes and anchors

In July 2026 the line fell from ~95% to ~88%. Wallets did not abandon SegWit — the number of SegWit transactions kept rising. What changed is the denominator: starting 2 July 2026, an automated mint wave from Alkanes — a smart-contract token protocol built on Runes — flooded the network with tiny transactions minting the DIESEL token. Each writes its data into an OP_RETURN “runestone” and spends a keyless pay-to-anchor output, which is spent with an empty witness — so the transaction carries no witness data and correctly does not count as SegWit. As of early August 2026 these were ~98% of all OP_RETURN transactions; non-witness transactions more than tripled (roughly 19k to 83k per day), diluting SegWit’s share even though real wallet usage of SegWit was unchanged — a compositional shift in what’s on-chain, not a reversal in adoption.

Why you might see a different number elsewhere

“SegWit adoption” isn’t a single definition. We follow Bitcoin Core’s convention: a transaction is SegWit when it carries witness data. That’s the same definition Glassnode uses, and both read ~88% during the 2026 Runes surge. Some trackers instead count a transaction as SegWit whenever it spends a SegWit-type output — including the pay-to-anchor outputs these Runes transactions use, even though they’re spent with an empty witness. Under that definition the number stays near ~97%. Neither is “wrong”; they answer slightly different questions. We show the witness-carrying measure because it tracks genuine wallet upgrades and moves when the network’s real composition moves.

Caveats

This is a transaction-count share, so it’s sensitive to whatever is filling blocks — a spam wave or a data-carrier protocol can move it sharply without any change in wallet behaviour, as the 2026 Runes surge shows. It doesn’t weight by value or by size. And because it counts witness presence, it treats keyless anchor spends as non-SegWit by design. Informational, not advice.

Related: Script types & adoption · all on-chain charts

Frequently asked

Did the 2026 dip mean people stopped using SegWit?
No. The number of SegWit transactions kept rising. A surge of Runes “data-carrier” transactions — which carry no witness — swelled the total transaction count and diluted SegWit’s share. It’s a change in what’s filling blocks, not a reversal in wallet upgrades.

Why does SiftingSats show ~88% when another site shows ~97%?
Different definitions. We (like Glassnode) count a transaction as SegWit when it carries witness data — Bitcoin Core’s swtxs measure. Some trackers count a transaction as SegWit whenever it spends a SegWit-type output, including keyless pay-to-anchor outputs spent with an empty witness, which keeps their number near saturation.

What are Alkanes, Runes, and pay-to-anchor outputs?
Runes is a Bitcoin token protocol that stores instructions in an OP_RETURN “runestone”. Alkanes is a smart-contract metaprotocol built on Runes; the 2026 surge was Alkanes mass-minting the DIESEL token. Pay-to-anchor (P2A) is a small keyless output spent with an empty witness — chaining these let a bot fire mints cheaply, with no signing or funding per transaction.

Where does the data come from?
Straight from the Bitcoin blockchain (getblockstats), recomputed from the chain daily. No third-party feed.