BIP-110 at Block 961,632: The Soft Fork That Turned Into a Liquidity Test

At block 961,421, only 47 of 1,806 blocks carried the BIP-110 versionbit. That is 2.60 percent. Down from 2.70 percent the day before. Over the prior 392 blocks, exactly eight new signals appeared. The activation window is nearly closed. The highest achievable total in the current period is 263. The threshold is 1,109. This is not a close race. This is a failed referendum happening in slow motion. Yet two prominent voices are screaming opposite conclusions. Dathon Ohm, the pseudonymous author of BIP-110, says Bitcoin Core will become insecure at block 961,632. Michael Saylor says the proposal's backers should stand down before they hurt themselves. Both are describing the same piece of software, the same block height, and the same timing. They cannot both be right. But both are telling you something about the state of Bitcoin governance. BIP-110 is a soft fork that caps the amount of data a Bitcoin transaction can carry. It ships in Bitcoin Knots, the smaller fork of Bitcoin Core maintained by Luke Dashjr. The rule change is framed as a curb on arbitrary data — the endless "blockspace spam" debate that has consumed more Twitter feed than technical merit. But in the last 48 hours, it has turned from a policy argument into a settlement risk. At block 961,632, roughly two days after Ohm's initial warning, mandatory signaling kicks in. Miners must set versionbit 4 in every block header. Blocks that fail to set that flag become invalid to any node running BIP-110-enabled software. That includes Bitcoin Knots with the patch. It excludes vanilla Bitcoin Core, which has no such rule. What did Ohm say precisely? He told miners and users to install Knots. And then he went further: "It is not recommended to run Bitcoin Core, as it will become insecure when mandatory signaling begins." Miners who generate templates from Core, he warned, may produce invalid blocks on "an incoherent chain that keeps being wiped out, along with any earnings." That is a strong claim. Nothing on the official BIP-110 website calls Core insecure. Ohm also calls the proposal a fix for "critical vulnerabilities" — a phrase absent from the site, which merely positions it as an anti-spam measure. Let's audit the numbers like a balance sheet. The BIP-110 monitor — a page the project itself points to — shows a trend that points nowhere near activation. At block 961,022 on August 4, Saylor counted 38 signaling blocks out of roughly 1,400, or 2.70 percent. By block 961,421, the monitor logged 47 of 1,806, or 2.60 percent. That is not just a stall; it is a decline. Now run the math. Early lock-in for BIP-110 requires 1,109 signaling blocks in a single two-week difficulty period. With 217 blocks remaining in the current period, the maximum achievable total is 263. Even if every single miner flipped their versionbit this instant — which is the equivalent of every corporate treasury buying BTC at a local top — BIP-110 would still fall 846 signals short. And no completed two-week stretch since December has finished above 1.29 percent. The pattern is not acceleration; it's a flatline. The only reason we are still talking about this is that mandatory signaling does not require lock-in to begin. At block 961,632, BIP-110 nodes will start rejecting blocks that do not carry the flag. This is not a question of consensus threshold. It is a forced choice: miners either set the bit or their blocks vanish from one particular view of the chain. That is the moment both Saylor and Ohm are pointing to. The difference is what follows. Luke Dashjr is certain the outcome is settled. Any miner who refuses to signal, he told BeInCrypto, loses their block rewards entirely. The invalid blocks they produce will only mislead nodes that have not yet updated. He added that shipping a Knots release without BIP-110 would not alter the timeline, and claimed there is no material opposition. That is a bizarre thing to say when the signaling rate is under three percent. But it is also how a certain kind of protocol idealist talks: "opposition" is measured by what the code says, not by what the hash rate does. Saylor reads the same mechanics and reaches the opposite conclusion. At 961,632, he says, BIP-110 nodes will reject non-signaling blocks. Unless major miners reverse course, Bitcoin continues normally while BIP-110 stalls or forks into irrelevance. His advice: backers should stand down. For weeks he has warned about Bitcoin neutrality — the risk that any activation scheme, even by a minority, could be weaponized to change the rules without consensus. Adam Back has joined the fray with characteristic bluntness: the reason BIP-110 lacks consensus, he said, is that it is a stupid idea, it does not work, and it fails technical consensus. He called it the polar opposite of SegWit, which had both technical and ecosystem consensus. That is not just an opinion; it is a structural comparison. SegWit had months of technical review, implementation, and a clear majority of signaling miners. BIP-110 has 2.6 percent signaling and a timeline measured in hours. But let me slow down and look at this from the mechanical side, because that is where my mind goes after years of watching systems break. The real issue is not whether BIP-110 activates. The real issue is the incentive matrix for a miner at block 961,637. You are a miner on Bitcoin Core. Your template has no versionbit 4. You produce a valid block on the Core chain. BIP-110 nodes reject it. Your block reward sits on a chain that a subset of the ecosystem considers invalid. If that subset is tiny, you lose nothing. If it grows, you lose everything. That is a convex payoff. Miners hate convex payoffs. They will do whatever reduces variance. The safest thing is to set the flag, even if you have no intention of supporting the rule change. It costs nothing. It avoids the risk of orphanage. That is why "signaling" is a meaningless measure of support. It is a measure of risk aversion. So the real data point is not the 2.6 percent. It is the cost of compliance. If setting versionbit 4 were truly neutral, every miner would be doing it already. They are not. The fact that 97.4 percent of miners are refusing to set a free flag tells you they either consider it a commitment trap or they expect the BIP-110 chain to die. Either way, the signal is clear: this soft fork is a dead asset. We didn't need a monitor to tell us that. In 2020, during the DeFi yield arbitrage, I learned that the most expensive thing in crypto is being on the wrong side of a liquidity mismatch. I spent three nights stress-testing slippage models against Ethereum gas spikes. The lesson: depth is the only truth. BIP-110 has no depth. It has a handful of versionbits and a developer with a megaphone. Yields don't lie, and neither do block headers. A chain that cannot even attract passive signaling in a bear market is not a chain. It is a proposal. Here is the contrarian angle, and I want to be careful, because it cuts against the easy narrative that BIP-110 will simply die. The conventional take says: let it die, no harm done. But the mechanical reality is nastier. The mandatory signaling threshold at block 961,632 means the split is not dependent on activation. It is dependent on software choice. Every node running BIP-110 will reject non-signaling blocks from that moment on. That is not a fork proposal; it is a static partition. The question is whether the Knots node set is large enough to produce persistent orphan blocks, even if those orphans are few and far between. That is the hidden friction. In a low-liquidity environment, even a small orphan rate can create a gap in the longest chain, and then you have a real chain split with real exchange implications. The market is pricing this as noise. I am not so sure. In 2022, when Terra collapsed, we didn't look at the death spiral; we looked at Celsius's balance sheet. The second-order effects killed more firms than the initial depeg. The same logic applies here: the second-order effect of BIP-110 is not activation. It is the precedent that a minority node cohort can force an economic choice onto miners at a specific block height. That is a governance coup, regardless of whether the flag is set. It may fail this time. But the template is open-source. Every future soft fork can use it. This is also a bifurcated market lesson. In 2024, when I tracked the liquidity bridge between BlackRock's IBIT and on-chain reserves, I noticed that ETF inflows weren't moving spot liquidity. The institutional pool and the retail pool were decoupling. Same thing is happening here. The institutional pool — the exchanges, the custody vaults, the ETF market makers — will largely run Core. The retail pool — the hobby miners, the node purists, the Knots loyalists — may run BIP-110. If the split happens, the price discovery will happen on the institutional pool. The question is whether the retail pool can keep a chain alive long enough to confuse the settlement layer. That is a classic friction point, and in a bear market, friction is a killer. So where does Bitcoin land at 961,632? My base case is Saylor's: Bitcoin continues on Core, BIP-110 stalls into irrelevance, and the weekend is a non-event for price. But the structural scar is real. The node ecosystem just learned that a small, well-organized group can force a global settlement test with almost zero support. That is a new attack surface. Watch the orphan rates on Monday, not the tweet storms. Position with a long-vol bias if you must hold exposure. But mostly, hold your keys. The chain settles. The narrative doesn't.