BIP-110 at 2.6%: The Vote That Failed and the Fee Signal That Remains

The timestamp is block 961,632. The network is online. The proposal is not. In the latest BIP-9 signaling window, only 2.6% of Bitcoin mining hashpower signaled support for BIP-110, the temporary soft fork designed to impose seven consensus restrictions on non-payment data. The activation threshold is 90%. This is not a close vote. It is not a slowly forming majority. It is a census: miners were asked whether they wanted to restrict inscriptions, and 97.4% answered with silence. The ledger does not lie, only the storytellers do. The number is not the real story. The real story is what 2.6% hides. For those who spend their days inside the mempool, the context is familiar. Inscriptions use Taproot’s extended script paths to embed arbitrary data in bitcoin transactions. Images. Text. Metadata. The practice is not a new currency; it is a new way to consume block space. BIP-110 is a response. It is a temporary soft fork with an effective window of roughly one year, introducing seven consensus-level constraints on data-heavy transactions. It uses the BIP-9 version-bit mechanism: miners signal support by setting a value in the block header’s version field. If 90% of hashpower signals during a difficulty period, the soft fork activates. If not, the proposal dies by calendar. Block height 961,632 serves as the deadline for the signaling experiment. That makes this event uniquely testable. Most governance debates in crypto are vague; this one has a timestamp. In my audit work, I have learned to treat deadlines as data collection instruments. The date forces a decision into a finite window. The result is clean: one epoch, 2.6%. A 2.6% signal rate is below statistical noise. To put it in context, an individual pool can move the number by a full percentage point in a single day if one of its main nodes changes configuration. A movement with real momentum should produce a sustained line on a chart. Instead, the chart is flat. There is no slowly assembling coalition. There is no last-minute dash toward 90%. There is a small cluster of version stamps, then nothing. I follow the bytes, not the headlines, and the bytes are saying that BIP-110 is not going to activate in the foreseeable future. But why did the miners say no? This is where structural hypothesis testing matters. I ran two competing hypotheses. Hypothesis A: miners rejected BIP-110 because they share a philosophical view that bitcoin should remain a pure monetary network. Hypothesis B: miners rejected BIP-110 because inscription fees are income, and they are not inclined to vote against their own revenue line. The on-chain evidence favors Hypothesis B. Let the fee data speak. In recent epochs, non-payment transactions have become a meaningful share of block space. An individual inscription can consume 200 kilobytes or more of a block. It pays in completely ordinary sat/vB fees. The miner does not ask whether the data is a jpeg or a settlement layer; the miner sees the fee column. When the fee column is thick, a proposal that kills that fee stream is a tax increase on the miner. The version field is the only legal way to vote against that tax. This is not an exotic theory. During the 2020 DeFi summer, I spent three months back-testing yield strategies on Ethereum transaction logs. The pattern was identical: protocols that generated fees from “useless” activity were defended by the same validators who had publicly condemned the activity. Economic exposure beats ideological preference almost every time. BIP-110 is the bitcoin version of that lesson. It is not a constitution; it is an election with a fee schedule. Compliance Brief: For institutional readers, the risk translation is straightforward. The support rate has no machinery behind it, so the proposal does not create near-term systemic risk. Bitcoin’s consensus layer remains stable. But the regulatory surface does not shrink simply because a BIP fails. As long as inscription bytes are a recurring percentage of block fees, the network is carrying a data market inside a monetary layer. That is not a systemic failure; it is a structural fact that regulators will eventually notice. The market has not priced that fact yet. Now the contrarian angle, and it is important. Low miner support is being read as a decisive ideological rejection of the monetary-purist camp. That is the wrong conclusion. The number is not about ideology; it is a snapshot of the fee market at this exact difficulty period. The same miners who silently refused to signal yesterday will signal tomorrow if the fee column changes. They do not have to love inscriptions. They just have to need the revenue. The absence of a BIP-110 majority tells us nothing about the durability of the pure-money narrative. It tells us that, at this moment, survival math and inscription fees are aligned. This is the correlation-causation error hiding inside the narrative. Because 2.6% is low, the market concludes the inscription problem has been defeated. But the “problem” was never going to be solved by one failed vote. Taproot is still live. BIP-341 is still a data highway. A thousand new inscriptions can be created the day after BIP-110 dies. The only thing that died is a specific proposal with a specific timestamp. The underlying tension remains encoded in every block that carries a non-payment payload. History repeats, but the code changes the rhythm. In 2017, a similar signaling debate produced SegWit after a long and bitter fight. It took years, multiple version-bit proposals, and a user-activated soft fork threat. The final implementation looked nothing like the original draft. I expect the same arc here. The first attempt to restrict inscription data failed. It will not be the last attempt. The next proposal will likely be narrower, better engineered, and more carefully timed — which also means it will be harder to dismiss. The window for that is twelve to twenty-four months, not the next ninety days. For now, the practical consequence is that inscription-related assets are facing a temporary removal of the “regulatory boot.” The probability that BIP-110 activates is low. So the immediate risk to Ordinals and related ecosystems is also low. That is the opportunity buried in the failed vote: the market can trade the next six months without a forced consensus change. But do not mistake a delay for an acquittal. The proposal’s failure does not make the issue disappear. It only pushes the conflict further down the version field. If support somehow crosses 10%, inscription assets will move violently; the alert is already coded into the fee column. Forensic Footnote: BIP-9 requires 90% of hashpower to signal within a difficulty adjustment period for activation. The 2.6% figure comes from the last completed signaling window as observed through public pool announcements and mempool.space statistics. The proposal may be temporary, and the exact BIP number in the canonical bips repository may differ from the label used in market discussions. Inscription share is defined here as the percentage of block bytes and fees contributed by non-payment transactions using Taproot script paths. Node behavior expectations at block 961,632 remain opaque; I have seen enough soft fork debates to know that vague rules produce confused verifiers. My index has boundaries; the ledger does not lie, but the ledger also does not label itself. Takeaway: stop watching the vote count and start watching the fee book. If inscription bytes continue to rise toward 50% of total block fees, miner incentives will harden and BIP-110 stays dead. If that fee stream collapses, expect a quieter, more surgical soft fork within two years. The next signal will not arrive as a tweet. It will arrive as a miner’s version field flipping from 0 to 1. Block height 961,632 is still approaching. When it gets here, the question will be simpler than the debate: were you looking at the polling page, or the fee column? Precision is the only hedge against chaos. I know which column I am watching.