Three percent. That is the number that should stop every Bitcoin observer cold. BIP-110, a proposal to force miner compliance through mandatory signaling, entered its enforcement phase with less than three percent of the network's hashpower offering a supportive signal. In any other governance system — a parliamentary vote, a shareholder ballot, a strike authorization — that margin would have been called a rout. The proposal marched forward anyway.
This is not ancient history, nor is it a footnote. It is the clearest case study in Bitcoin's unwritten constitution: a governing framework without courts, without formal ballots, without an obvious mechanism for counting consent. The BIP-110 saga answers a question that institutional observers are only now beginning to ask with full seriousness: who actually decides what Bitcoin is? The answer, it turns out, is stranger than either the maximalists or the critics would have you believe — and it becomes more pertinent with every passing cycle.
The Context: An Experiment Born in Schism
BIP-110 occupies a peculiar corner of Bitcoin's technical genealogy. Proposed during the inflection years of 2015 and 2016, when the blocksize debate was splitting the ecosystem into hostile camps, it belongs to the lineage of mandatory signaling ideas: mechanisms by which full nodes, rather than miners, enforce protocol upgrades. The technical logic was straightforward. If enough node operators ran a version of Bitcoin Core that refused to accept blocks missing a specific version bit, miners would eventually have no commercial choice but to comply.
The design philosophy flows directly from the user-activated soft fork (UASF) tradition. Its premise is that legitimacy in Bitcoin flows from the people who run nodes, not from the people who run ASICs. The version-bit system proposed in BIP-110 was an early attempt to operationalize that premise — an attempt that predates BIP-9, the mechanism that ultimately became the standard deployment framework for soft forks such as SegWit and Taproot.
BIP-9's design differs in one crucial respect: it requires 95 percent miner approval across a full difficulty period before activation. It is a consensus-seeking instrument. BIP-110, by contrast, is a consensus-enforcing instrument — a hammer, not a handshake. It is the difference between asking someone to a meeting and serving them an eviction notice. That BIP-110's mandatory signaling phase went live at all, with a miner support rate below three percent, tells us something essential about the limits of node power in a proof-of-work system.
This was the twilight of the "gentlemen's agreement" era of Bitcoin governance. In 2017, the New York Agreement — a publicly signed accord among miners, exchanges, and companies to activate SegWit2x — would later reveal how fragile that model was. BIP-110 belongs to the shadow history of that civil war: the more obscure skirmish that preceded the famous battles. The future would not belong to BIP-110. But the questions it raised never went away. They are the same questions now echoing through debates about ETF custody, mining-pool consolidation, and the emergence of AI-driven economic agents.
The Core: The Architecture of Divided Power
Let me be precise about what was actually at stake. Bitcoin's design grants two distinctly different kinds of power to two distinctly different constituencies. Miners hold the power to produce blocks — to write history into the ledger. Node operators hold the power to validate blocks — to decide which history is legitimate. Under normal conditions, these powers align, and the system operates with the calm predictability of well-oiled machinery. The collapse of that alignment was the true subject of BIP-110's signaling phase.
To understand the mechanics, you have to look at the deployment state machine that governs such proposals: defined, started, locked in, active. BIP-110 compressed these stages into something closer to a decree. After a specific window, enforcing nodes simply rejected non-compliant blocks. The design did not wait for a threshold; it enforced a position. That shortcut is precisely what made it structurally different from BIP-9, which formalized a patient state machine to avoid direct confrontation. Shortcuts in governance produce measurements, not consensus.
And the numbers tell an economic story far better than any polemic. A miner support rate below three percent is not a symptom of apathy. In a market where mining is a hyper-competitive commercial operation with thin margins, relentless capital expenditure, and facility-level debt, it is a coordinated signal of refusal. The miners looked at BIP-110 and asked a different question than the developers. The developers asked: is this the right governance model? The miners asked: what does this get us?
Follow the money, not the noise. The answer was, precisely, nothing. BIP-110 did not improve transaction throughput. It did not reduce fees. It did not extend a carrot to offset its stick. It was pure governance restructuring — a reallocation of authority from the hashpower side of the network to the node side. The miners' economic calculation was entirely rational. Why voluntarily surrender influence without compensation? The three percent signal was the market's verdict, delivered in the only language that matters in a proof-of-work system: hashpower.
This is where the philosophical heart of the matter beats. The activation of a mandatory signaling threshold in the absence of miner consensus creates a network with two competing consensus views: nodes that recognize one chain extension, miners producing blocks that the enforcing nodes refuse. The result during BIP-110's experiment was not a swift victory for either side. It was a revelation that "consensus" in proof-of-work is not purely mathematical. It is a political settlement between internal constituencies with different economic interests — a settlement prosecuted through software, resolved through markets, and settled by capital.
I recognize this pattern from personal experience. During my 2017 due diligence work in the ICO boom, I spent weeks auditing a cross-border payments token that boasted "community governance." The reality beneath the rhetoric was a multisig treasury wallet controlled by known venture capital signatories. The community could vote on cosmetic changes; the VCs decided the architecture. BIP-110's story is the inverse mirror: the community of full node operators wanted to assert governance, but the capital-intensive actors — the miners — simply refused. Governance without economic weight is a committee without a budget.
The developer ecosystem absorbed this lesson better than most observers credit. After BIP-110's forced signaling experiment demonstrated its own weakness, Bitcoin moved decisively toward BIP-9's model — one that institutionalized 95 percent miner thresholds and avoided direct confrontation. It was, in retrospect, a quiet admission: the code-driven activism of forced signaling could not substitute for capital that was never going to be deployed. The market does not negotiate with proclamations; it negotiates with production.
I saw this dynamic amplify during DeFi Summer in 2020, when I was tracing liquidity mechanics for a report on Latin American remittances. Yield farmers did not participate in protocol governance as citizens. They behaved as capital allocators. When a governance proposal threatened their yields, they exited. When the ideological stewards of an ecosystem lacked capital behind their proposals, their voices became whispers in a thunderstorm. The miners' behavior during BIP-110 was the original version of that refusal — the first time Bitcoin's internal negotiation between rule and capital became externally visible in telemetry.
The Economic Disincentive History Forgot
There is a deeper economic layer to the BIP-110 story that most retrospectives ignore. The proposal's failure was not merely a governance setback; it was an early signal of a structural misalignment that would define Bitcoin's security economics for years to come. For most of the 2010s, miners derived the overwhelming majority of their revenue from the block subsidy — newly minted bitcoin that made up roughly 90 percent of their income. In that environment, any protocol change that did not touch the subsidy was, at best, a distraction, and at worst, a threat to the delicate cost structure of their operations.
Now consider the current landscape. With the block subsidy having halved multiple times and transaction fees growing in relative importance, the calculus has shifted. The recent wave of inscription activity — Ordinals and related protocols — injected a new fee narrative into Bitcoin's base layer, and with it, a new revenue stream for miners. My view, informed by years of tracing payment flows across Latin American corridors, is that this fee injection has been a genuine lifeline for Bitcoin's security model. Without the inscription wave, the divergence between security costs and security rewards would have become an existential question by now. This is the lens through which BIP-110's failure should be read. Proposals that fail to reward the actors who must implement them will always struggle, regardless of technical merit. The miners who ignored mandatory signaling were not being obstructionist. They were responding to the same incentive structure that now makes them eager to process inscription-containing blocks. Miner cooperation is not given; it is purchased through the alignment of interests.
The Market's Missing Appetite for Governance Risk
What did the market make of all this? The honest answer is: very little, which is itself a data point. Protocol governance disputes of this kind tend to be priced subtly. During the 2017 SegWit2x saga, the market priced in the risk of a contentious hard fork — the premium appeared in futures spreads and options implied volatility, not in the spot price. The same dynamics applied to BIP-110, scaled down. The broader lesson for institutional participants is uncomfortable: markets are terrible at pricing long-horizon governance risk because such risk rarely manifests on a predictable schedule. It arrives in sudden bursts of chain reorganizations or exchange announcements — precisely the kind of event that no model can predict in advance.
One framing has shaped my reading of such events: markets are regime-uncertainty averse, not event-uncertainty averse. They handle the question "what will happen on this block height" poorly but adequately; they handle the question "what rules will govern the network in 2030" very poorly. BIP-110 was a mechanism question, not an event question. Its low direct market impact was therefore predictable. But the accumulation of unresolved mechanism questions is exactly how governance debt builds up. And governance debt, like technical debt, compounds silently until it triggers a liquidation event.
My 2024 regression analysis of ETF flows across 15 major altcoins showed something counterintuitive: governance disputes originating in Bitcoin itself had almost zero measurable effect on altcoin liquidity distribution. The market has learned to segment "Bitcoin governance noise" away from everything else. That is both a maturity signal and a vulnerability. Complacency is precisely what makes the next dispute unpriceable. During the 2022 bear market, as I watched leveraged protocols collapse across the industry, I came to understand that the market's relationship with governance risk is fundamentally asymmetric. In bull markets, governance disputes are noise — or worse, they become marketing material framed as "vibrant decentralization." In bear markets, the same disputes become existential narratives that accelerate capital flight. BIP-110's timing, during a period of relative price stability, meant that its governance signal was absorbed by the market's attention economy. This was a fragile blessing. The next mandatory signaling episode will not be so lucky.
The Contrarian Angle: Failure as Fault Injection
The conventional reading — that BIP-110 failed because miners ignored it — is far less interesting than the truth. The truth is that BIP-110 succeeded as a testing instrument precisely because it failed as a governance mechanism. Its design included a hard-fork rollback pathway, a contingency that presupposed the possibility of failure. The mandatory signaling phase was not a desperate attempt to seize power from miners. It was a shadow boxing match designed to measure the network's response curve — a deliberate injection of stress into the system to observe where the cracks would appear.
This reframing has a name in engineering: fault injection. You do not wait for a bridge to collapse to learn its load capacity; you apply increasing loads, observe the deformation, and document the limit. BIP-110 applied a governance load that Bitcoin had never experienced. The three percent support figure was not the failure of the proposal. It was the measurement of the network's true power structure — a measurement available exclusively to those willing to launch an intentionally controversial flag.
The decoupling thesis is worth dwelling on here. While the public debate framed BIP-110 as a battle between developers and miners, the structural transformation that mattered was happening elsewhere: in the emergence of mining pools as economic coordination vehicles outside both groups. Pools are neither nodes nor individual miners. They are capital aggregation machines, and their collective indifference to BIP-110 is the real data point. The three percent support rate almost certainly did not reflect three percent of individual miners making independent choices. It reflected the refusal of major pools to even consider adaptation. In Bitcoin, power does not reside in code or in hashpower alone. It resides in the coordination layers that can mobilize hashpower — a structural reality that no amount of protocol design can circumvent.
Looking back from the vantage point of 2026 makes this timelier than ever. As the ETF era institutionalized Bitcoin, the governance question migrated from node operators and miners to custody providers and SEC-regulated money managers. BIP-110's mandatory signaling was a rehearsal for a drama that now plays out in boardrooms rather than blocks. The actors have changed; the dynamics have not. When a handful of custody providers coordinate policy responses through their scale, that is not governance, but it exercises more authority than any version-bit vote could. The lesson of the three percent revolt is that in any system, whoever creates the infrastructure at the base also sets the rules at the top — whether through an ASIC or a trust department.
There is a further darkness here that I do not want to gloss over. The BIP-110 spectacle — the appearance of a democratic activation process with almost no participation — is structurally identical to the governance theater that continues to afflict DAOs in 2026. Voter turnout in almost every major DAO remains below five percent. Proposals are written by core teams, passed by whales, and enforced by multisig signers whose identity barely changes from year to year. There is a regulatory irony the industry has not fully confronted. Projects preach decentralization from every stage, publishing governance dashboards and delegation metrics that suggest participatory vitality. Meanwhile, team wallets and foundation holdings remain traceable on-chain, and the DAO itself functions as a compliance shield — a layer of plausible deniability between decision-makers and decisions. BIP-110 was never a DAO, of course. But the pattern of manufactured consent is the same: the appearance of a process, the reality of a veto. At least the miners had the clarity to refuse. The DAO members simply do not show up. The result is a governance market that has learned to manufacture consensus rather than measure it.
The Takeaway: What the Ghost of BIP-110 Leaves Behind
So what did BIP-110 leave behind? Not an activated feature. Not a canceled upgrade. Not even a coherent precedent. It left something more valuable: a map of Bitcoin's political terrain, drawn in the only valid way — through direct engagement with the system rather than wishful thinking about how it should work. Bitcoin survived the experiment not despite the low miner support but because of it. The three percent figure was a lesson in humility for the developers, a lesson in coordination for the miners, and a lesson in structural reality for the rest of us.
As AI agents enter the economic equation in earnest — agents with their own wallets, their own capital, and their own negotiation strategies — the governance debates of the next decade will replay BIP-110's central question with amplified stakes: can rules ever be forced upon the actors who hold the capital to resist? I suspect the answer will be the same as it was in 2015. No. But the attempt to draw the map is what matters. Consensus is not a technical term; it is a political settlement, renegotiated every time a version bit flips. Volatility is the tax on impatience. Follow the money, not the noise — and when the money refuses to move, listen carefully to what that refusal says about where power truly resides.