Alkanes Drove 61% of Bitcoin Transactions Across 90 Days of Blocks

SUBFROST published a figure on September 20, 2026 that deserves more scrutiny than it has received. Its indexer counted 59.9 million total Bitcoin transactions between blocks 954,917 and 967,918, spanning June 23 to September 20, and attributed 61.1% of them to Alkanes, the WASM smart contract metaprotocol that settles its state through Bitcoin transaction data. If the number holds, a token protocol most Bitcoin holders cannot name has been the majority user of the world's most expensive settlement network for a full quarter. The stakes are not aesthetic. They run to miner revenue, node operating costs, the Bitcoin Core relay policy fight, and the question of who Asian retail flow answers to when the Financial Services Agency in Tokyo and the Financial Services Commission in Seoul start asking.

The Arithmetic Behind the Claim

Start with the block range, because it is checkable. Blocks 954,917 to 967,918 is 13,001 blocks. The window is 89 days. At a perfect ten minute target that window should have produced 12,816 blocks, so the network ran about 1.4% fast, which is what you expect when hashrate climbs faster than difficulty adjusts. Nothing anomalous there.

Now the transaction count. 59.9 million transactions across 13,001 blocks averages 4,607 transactions per block. That is a high number by historical standards. Through most of 2022 the network averaged closer to 2,000 to 2,500 transactions per block. Blocks only carry 4,000 or more when the average transaction is small, which means few inputs, few outputs, and heavy use of OP_RETURN or witness data rather than large multi-party payment structures. Apply SUBFROST's 61.1% and you get roughly 36.6 million Alkanes transactions, about 2,815 per block, sustained for three months.

Here is the part the headline hides. Transaction count is the least meaningful of the three ways to measure blockspace demand. The other two are weight and fees paid. A protorunes message riding in an OP_RETURN with two small outputs consumes a fraction of the vbytes of a consolidating exchange withdrawal, and it pays a fraction of the fee. A protocol can plausibly hold 61% of transaction count while holding 25% of weight and 15% of fees. Until SUBFROST publishes weight and fee breakdowns, the 61.1% figure tells you how many times Alkanes touched the chain, not how much of the chain it bought.

The second caveat is the source. SUBFROST is not a neutral observatory. It builds frBTC, a synthetic Bitcoin asset that lives on Alkanes, and it has an obvious interest in Alkanes looking like the dominant activity layer on Bitcoin. That does not make the count wrong. Indexing protorunes is deterministic work and any competing indexer can reproduce it. It does mean the burden of replication sits with third parties, and as of this writing the major public dashboards have not matched the claim line for line.

An Asian Order Book, Not a Western One

The regional framing in the headline is the underreported half of the story. Alkanes activity has clustered in Asian trading hours, driven by Korean, Japanese, and Chinese-language communities operating through Telegram groups, self-custodial wallets, and automated market maker contracts rather than through licensed exchanges. This is the same population that drove the 2023 BRC-20 wave and the 2024 Runes launch at the halving block. It is fast, it is fee-insensitive when a mint is live, and it does not route through Coinbase.

Contrast that with where Western Bitcoin demand went in the same 89 days. The American flow is in spot ETF creations, in corporate treasury announcements, and in options on those ETFs. It is custodial, it is disclosed quarterly, and it produces almost no on-chain transactions per dollar deployed. A billion dollars of ETF inflow might generate a few dozen transfers between authorized participants and the custodian. A billion dollars of Asian retail speculation on Alkanes tokens generates millions of transactions.

So Bitcoin now has two demand bases with opposite blockspace footprints. The Western base buys the asset and touches the chain almost never. The Asian base barely holds the asset as a savings instrument and touches the chain constantly. Anyone reasoning about Bitcoin's fee market from ETF flow data has been reading the wrong instrument.

Regulators will notice this before most investors do. Japan's Financial Services Agency has been moving crypto oversight toward the Financial Instruments and Exchange Act framework, which pulls token issuance closer to securities treatment. Korea's regulators have spent two years building out a disclosure and listing regime for digital assets after the Terra collapse. Neither framework has a clean answer for a token that exists only as indexed state derived from Bitcoin transactions, with no issuer, no registered entity, and no chain to seize. The first licensed Asian exchange that lists an Alkanes asset will force that question into the open.

The Security Budget Argument

The bullish read on all of this is simple and it is the one I find harder to dismiss than I expected.

Bitcoin's block subsidy is 3.125 BTC and drops to 1.5625 BTC at block 1,050,000, roughly the spring of 2028. Over the 13,001 blocks SUBFROST measured, the network paid out about 40,600 BTC in subsidy. At a $100,000 Bitcoin that is over $4 billion in three months of security funded by dilution rather than by users. That funding halves every four years on a fixed schedule. Fees are the only replacement. There is no other line item.

Every serious Bitcoin security model has the same uncomfortable conclusion: the network eventually needs users willing to pay real money for blockspace, in volume, every block, through bear markets. Savings-only demand does not produce that. A holder who buys once and never moves coins again contributes nothing to the fee market. The uncomfortable truth is that speculative token activity, whether it is inscriptions in 2023, Runes in 2024, or Alkanes in 2026, has been the most reliable source of non-trivial fee pressure Bitcoin has ever had outside of price melt-ups.

The counter-history is worth stating plainly. In the December 2023 inscription rush, fees ran above a third of total miner revenue on the busiest days, a level the network had touched only briefly in 2017 and 2021. Then it faded. Every wave so far has been a spike, not a floor. Three months at 61% of transaction count is longer than any previous episode, which is exactly why the number matters. A spike funds nothing. A durable floor funds the 2028 halving.

The Relay Policy Fight

The bearish read comes from inside the house, and it is not about price.

Bitcoin Core's removal of the default 80 byte OP_RETURN datacarrier limit, shipped in the version 30 cycle, was the most contested policy change in years. The argument for it was that the limit was trivially bypassable through witness data and Taproot script paths, and that pushing data into harder-to-prune locations was worse than allowing it in a prunable OP_RETURN. The argument against it was that Core was deliberately making it easier to use Bitcoin as a data layer, and that policy defaults are the last soft lever node operators have.

The dissent went to Bitcoin Knots, which retains stricter filtering. Knots' share of reachable listening nodes rose sharply through 2025 and into 2026, and Ocean, the pool associated with that camp, built its template policy around not mining what it considers spam. That is not a fringe position anymore. It is a working minority with its own relay policy, its own pool, and its own template construction.

Here is where I think both camps are being imprecise. The real, measurable externality of metaprotocol activity is not that the transactions are ugly. It is UTXO set growth and initial block download time, because those determine whether a person can still validate the chain on ordinary hardware in ten years. That is a resource question with numbers attached. Content filtering is a proxy for it, and a bad proxy, because it is trivially defeated by any protocol willing to make its data look like an ordinary payment. Alkanes is closer to UTXO-efficient than the inscription wave was, since protorunes-style messaging commits state in OP_RETURN rather than minting a distinct UTXO per asset transfer. If the small-block camp wants to win the argument, it should publish UTXO growth and IBD benchmarks attributable to Alkanes. If the Alkanes camp wants to win, it should publish the same numbers and show they are small. Neither has.

Blockspace Is an Auction, Not a Permission List

Bitcoin's claim on the future is that it is money no government can debase and no committee can gatekeep. Those two properties are the same property. A network that cannot be pressured into blocking a payment also cannot be pressured into blocking a token mint it finds distasteful. You do not get to keep the censorship resistance and discard the parts of it you dislike.

I will take the position. The fee market is the correct arbiter, and it already did its job. Alkanes users bid for blockspace and won it. Nobody was outbid and denied settlement by fiat; they were outbid and denied settlement by price, which is what a market is for. Every mechanism proposed to fix this, whether content sniffing in relay policy or pool-level template curation, requires someone to decide which transactions are legitimate. That role, once it exists, does not stay in the hands of hobbyist node operators. It becomes an address for a subpoena. The Ocean camp is defending something real about resource limits, and it is doing so with a tool that, if it ever worked at scale, would hand every state actor the precedent it has wanted since 2013.

Meanwhile the dollar has been managed by a committee of twelve people for over a century and the arguments for that arrangement still get called serious monetary policy. Bitcoin's alternative is a fixed schedule, an open auction, and no appeals process. The auction producing outcomes the founders would not have picked is not a flaw in the design. It is proof the design has no editor.

What to Watch

Independent replication by the end of Q4 2026. The 61.1% figure needs a second indexer. Watch whether public dashboards land near SUBFROST's number or materially below it. My expectation is that transaction count replicates within a few points and that the weight share comes in under 35%, which would make the headline true and the implied significance overstated.

Fee share of miner revenue on a 90 day trailing basis. This is the only number that decides whether Alkanes mattered. If the trailing 90 day fee share stays above 15% through the 2027 calendar year, the security budget argument is winning. If it drops back under 8% within two quarters of activity cooling, this was another spike and the April 2028 halving arrives with the same unsolved problem.

Knots share of reachable nodes. If it sustains above roughly a fifth of listening nodes, expect visible relay fragmentation, transactions that propagate unevenly, and pressure on miners to publicly declare template policy. That is the scenario in which the policy fight stops being a mailing list argument and starts costing users confirmations.

The first frBTC stress event. Synthetic Bitcoin on a metaprotocol is a custody and bridge question wearing a protocol costume. Every prior cycle ended the same way, with a wrapped asset failing to redeem at par. Watch frBTC supply against disclosed backing. This is the most likely mechanism by which the Alkanes trade ends badly, and it has nothing to do with Bitcoin's base layer.

A licensed Asian listing. The moment Upbit, Bithumb, or a Japanese licensed venue lists an Alkanes asset, the FSA and FSC have to write down a classification. Expect them to treat the token as a security-like instrument and the indexer operators as regulated infrastructure. Expect that to be unenforceable against a protocol with no issuer, and expect the attempt anyway.

Block 1,050,000. Everything above is a rehearsal for that number. The subsidy falls to 1.5625 BTC and the question of who pays for Bitcoin's security stops being theoretical. If the answer turns out to be Asian retail speculating on tokens that most Bitcoin holders consider noise, that will be an awkward result. It will also be a working one.

---

Go deeper: The Bitcoin Halving · Bitcoin ETFs Explained

Source: Bitcoin Magazine