The number is 185. That was the block count remaining when the August 7 monitoring data flagged the countdown: Bitcoin BIP-110's mandatory signaling period was about to begin rejecting blocks. The signal support rate behind this rule change sits at 2.62 percent. Bitcoin Core, the network's dominant node implementation, closed the relevant pull request on March 26 without merging it. Bitcoin Knots, the minority implementation, is activating the rule anyway. No comparable activation has ever reached this point with such minimal endorsement.
This is not an upgrade. Upgrades carry consensus. This is unilateral enforcement of a validity rule that miners, developers, and the broader ecosystem never agreed to. Trust the code, but verify the architecture. The architecture in question is not BIP-110's Merkle path compression scheme. It is the foundational assumption that every Bitcoin node implementation reads from the same rulebook.
Context: A Soft Fork With a Decorative Threshold
BIP-110 is a version-bit soft fork targeting block data compression and SPV verification efficiency. The idea is not new; it descends directly from the SPV optimization proposals that circulated during the 2015-2017 block size wars. What is new is the activation path, which abandons every guardrail that made BIP-9 workable.
BIP-9 was the activation mechanism behind the last successful soft fork wave, and its failure mode was community fragmentation. BIP-110 keeps the version-bit language but drops the consent math entirely. Standard BIP-9 activation requires roughly 95 percent hashrate signaling over a 2,016-block window. BIP-110 lowers the threshold to 55 percent, or 1,109 blocks. That alone is a red flag. But the structural deviation is worse: the mandatory signaling period begins at block 961,632 regardless of whether the threshold is reached. Execution nodes reject blocks without version bit 4 starting at that height. If no lock-in occurs, the data reduction rules still activate at block 965,664. The threshold is decorative.
Bitcoin Knots implements BIP-110. Bitcoin Core does not. Core contributor Antoine Poinsot stated in June, in a personal capacity, that Core would not execute the proposal. The Core PR for BIP-110 was closed unmerged in March. Meanwhile, OCEAN mining pool switched its default endpoint to a BIP-110-compliant configuration on July 15 — a full month before the activation window. That timing is not an accident; it is coordinated preparation. In any governance framework, coordination without disclosure is how minority control happens. The coordination pattern matters. OCEAN announced the endpoint switch on July 15, signaling intent a full month before the rule activates. That gives compliant block producers time to align infrastructure. It also gives the rest of the network time to pretend nothing is happening.
Core: The Risk Is Divergence, Not the Rule
I have spent five years inside decentralized governance systems, most recently building a modular compliance layer for an institutional custodian during the ETF integration wave. The pattern here is familiar: a technically confident minority pushes a rule change through a broken process, and the cost lands on downstream infrastructure. The same dynamic appears in any governance system where execution speed outpaces consent.
The rule's technical design is unremarkable. The risk is what happens when two node implementations disagree on block validity.
Risk one: chain stall and islanding. After 961,632, if miners keep producing blocks without bit 4, Knots nodes reject them. Those nodes cannot extend the chain. They stall at the last compliant block or repeatedly reorg against a chain they consider invalid. The only escape is a pool like OCEAN continuously producing compliant blocks. OCEAN controls an estimated one to two percent of network hashrate. That is enough to keep Knots nodes alive. It is nowhere near enough to sustain a parallel chain. The result: a "valid" Knots chain with trivial work behind it, and a "valid" Core chain with 97.38 percent of hashrate. Two definitions of valid. One network.
Risk two: upgrade state machine hazards. BlockSlop reproduced a narrow failure in regtest. Switching from a BIP-110-executing Knots node to a non-executing implementation retained previously accepted blocks in the data directory. On restart, the node briefly operated with inconsistent rule state — accepting an inheritance of history that the new rules would not tolerate. Knots merged a header-scanning protection that invalidates violating blocks and triggers a reorg. But transaction-level violations invisible in block headers still require full reconnection validation, and some cases require a reindex. The fix exists, but it arrived weeks before activation. That is a compressed audit cycle for a consensus-level rule change.
Risk three: the compatibility gap. From 961,632 onward, Bitcoin Core and Bitcoin Knots theoretically disagree on block validity. This violates the core safety principle that node implementations converge on a single consensus state. Bitcoin's security argument is the longest valid chain. When "valid" depends on which binary you run, that argument loses its force. The Knots team's own August 7 warning admits it: old non-executing software may leave an unsafe chain state. That is a consensus-level admission, and it is rare. Bitcoin's security model was never designed to adjudicate disputes between implementations. It was designed to be unambiguous.
The signal data is worse than the threshold math suggests. Across the monitoring window, only 2.62 percent of blocks carried bit 4. That is not a near-miss; it is a wholesale rejection. BIP-34, BIP-66, and BIP-65 all cleared their 95 percent thresholds within months of signaling. BIP-110 is attempting to activate a rule that 97.38 percent of hashrate declined to endorse. In DAO governance, this would be a failed proposal — rejected, archived, and revisited with better framing. The only reason BIP-110 is moving is that Knots controls the execution layer for a subset of nodes, and OCEAN controls block production for a sliver of the network. That is not consensus. It is a supply chain.
The governance failure compounds the technical risk. During the 2022 crash, I ran an emergency governance intervention: pausing a flawed voting mechanism, implementing quadratic voting to block whale dominance, and running fifty community calls in two weeks. The lesson was brutal — pre-defined rules matter more than goodwill. BIP-110's activation path has no pre-defined integrity. The 55 percent threshold is not oversight; it is a formality designed to be bypassed. Efficiency without oversight is just faster risk.
Taken together, these three risks describe a system whose failure mode is institutionalized. The problem is not that BIP-110 exists. It is that nothing in the current governance framework stopped it from reaching this point. No activation gate, no implementation veto, no binding hashrate consent. The threshold was designed to be bypassed, and the clock is already running.
Contrarian: The Stress Test We Did Not Ask For
Now the counter-intuitive read. This may be the healthiest stress test Bitcoin has run in years. Node implementation pluralism means divergence is possible — and the network gets to see what that looks like. The market's dismissal is rational. At 2.62 percent signaling, with no exchange support, no wallet support, and no meaningful hashrate, a fork without infrastructure is not a fork. It is a hobby chain.
Economic value follows hashrate. The 2017 Bitcoin Cash split had miner support, exchange support, and a clear leadership faction — and still captured only a fraction of Bitcoin's value. BIP-110 has none of those. Price impact stays below one percent unless OCEAN's compliant blocks produce visible infrastructure disruption. That does not make it harmless. It makes it a warning.
But the darker pragmatism is precedent. Non-consensus soft forks are now a demonstrated move. Any implementation with enough conviction can force a validity divergence, and the cost is borne by exchanges, wallets, and indexers downstream. The minority faction believes it is right. Sometimes it is. The problem is that the process died — and the precedent outlives the correction.
Takeaway
BIP-110 will fail silently or produce a negligible chain that a few dozen Knots operators maintain. What persists is the knowledge that a minority implementation can force a validity divergence on the network. That knowledge will be reused. Governance is not a feature; it is the foundation. If Bitcoin does not standardize activation rules across implementations, this scenario repeats. The ledger remembers what the community forgets. In the crash, only structure survives.