In the ninety days after Bitcoin's fourth halving, a quiet inversion took hold. Hashrate—the total compute committed to the network—kept climbing toward all-time highs. But hashprice, the dollar value that compute could actually earn, fell below the floor it touched during the 2020 capitulation. The machine got stronger while the work got poorer. Nobody hosted a panel about the gap.
I spent that quarter buried in miner revenue data rather than narratives, and what I found was not a failing network. It was a network quietly becoming something other than what its evangelists still describe from stages. The halving did not break Bitcoin. It converted a philosophical assumption into an empirical question, and the empirical answer is uncomfortable.
Context: the most predictable event in monetary history
The fourth halving was never a surprise. It is hardwired into the 210,000-block cadence that has governed issuance since the genesis block—every four years, the subsidy halves, most recently to 3.125 BTC per block. Every miner, every pool, every ASIC manufacturer has had a decade to model it. And yet the arithmetic remains unforgiving.
The subsidy cut erased roughly half of miner revenue overnight. The theoretical relief valve is the fee market: as subsidies shrink, users are supposed to bid for scarce blockspace, and fees fill the gap. But fees historically account for low single-digit percentages of total miner revenue on ordinary days, and the base block is capped at roughly 1MB of data—four million weight units. For fees to replace a 50% subsidy cut, demand would need to multiply several times over and then stay there. That has not happened. It may not happen for years. The network is running a business model that its own rules say must eventually change, on a timeline that has just gotten shorter.
I have been auditing this kind of gap since 2017, when, as a twenty-one-year-old, I spent six months studying the governance contracts of early DAO prototypes while my peers chased token listings. The lesson from that work was simple, and it applies here: a mechanism is only as trustworthy as its worst-case arithmetic, not its average-case marketing.
Core: where the hash actually goes
Here is the technical reality that "not your keys" maximalists tend to skip. Mining is not a solo endeavor, and it has not been for a decade. Solo block discovery carries brutal variance—a single rig might wait years for a block—so small miners join pools that smooth income by distributing rewards proportionally. That is rational behavior. But it concentrates something far more important than reward: block template construction. The pool, not the individual miner, selects which transactions go into a block and in what order.
I pulled the pool-distribution data across the halving window, and the trend was not ambiguous. Foundry, Antpool, and ViaBTC rotated in and out of a combined 50–55% band of global hashrate. This is not a conspiracy. It is gravity. When margins compress, the pools with the lowest cost of capital and the best hardware pipelines absorb the orphaned hash of smaller operators. The halving did not create concentration; it accelerated it, the way a freeze reveals which pipes were already cracked.
Consider the vertical integration already underway. The largest pools are increasingly tied to hardware manufacturers and to funds that hold the very asset they mine. When the mine, the machine, and the balance sheet belong to the same owner, a pool is no longer merely a cooperative for smoothing variance—it is a vertically integrated financial entity with its own incentives, its own compliance posture, and its own private view of which transactions deserve inclusion. That is a very different institution from the cottage miners of 2013.
Let me be precise about what "three pools" actually means, because "51%" is lazy shorthand. The catastrophic risk—three operators colluding to double-spend—is remote and self-destructive; it would be visible, and it would destroy the asset they hold. The real risk is narrower and more insidious: at three-pool concentration, the marginal cost of censorship drops to near zero. A pool operator can decline to include transactions from a sanctioned address without violating the protocol, without coordinating with anyone, and without any user noticing. Nothing breaks. Blocks form, the chain advances, and one economic actor is simply excluded from the settlement layer.
We audit the code, but who audits the conscience? Here the audit is not hypothetical. It is a quarterly footnote no one reads.
This is not a Bitcoin-only disease. In 2020, during DeFi Summer, I spent three weeks reverse-engineering Harvest Finance's yield optimization and found that its alpha came largely from token emissions for which no sustainable demand existed—growth generated by the mechanism itself, not by the economy around it. I published a dissenting note my team ignored; the yield tokens collapsed months later. In 2024, tasked with explaining Bitcoin ETF approval to grassroots communities, I analyzed the custody solutions of major providers and kept returning to the same conclusion: trust-minimized bridges between TradFi and crypto are possible, but they require the same rigor we apply to smart contracts—and we almost never apply it to miners.
Contrarian: we are measuring the wrong thing
The reflexive response to concentration data is a call for "more decentralized mining"—solo-mining subsidies, Stratum V2 adoption, getblocktemplate reforms. I find most of this theater. It solves the wrong problem.
Bitcoin's decentralization was never really a property of hardware distribution. It was a property of exit. As long as a miner could credibly switch pools, pool power remained contestable, and contestable power is bounded power. But when hashprice compresses and mining becomes a low-margin capital business, exit becomes expensive—and expensive exit is no exit at all. We are not primarily witnessing the centralization of hash. We are witnessing the erosion of the credible threat to leave. That is a different disease, and no amount of better mining firmware cures it.
The same shape repeats across the ecosystem, which is why I stopped treating these as separate stories. Uniswap V4's hooks turn the DEX into programmable Lego, but the complexity spike scares off the developers who would otherwise fork it—concentrating power in the few teams that can navigate the hook surface. KYC regimes impose real friction on honest users while a freshly generated wallet sidesteps them entirely; the compliance cost lands almost entirely on people who were never the problem. In each case, a mechanism advertised as distributing power quietly reroutes that power to whoever can afford the entry fee.
In a sideways market, where price gives no direction and every narrative feels like a coin flip, these structural signals are the only honest way to position. Chop is not noise; it is the window in which you can read what speculation has been hiding.
Takeaway
The fourth halving did not break Bitcoin. It did something quieter and more consequential: it turned a founding philosophical assumption into an empirical question. Build not for the peak, but for the plain—and the plain, right now, is a network where three entities decide what gets settled, while the subsidy that funds their honesty shrinks every four years by law. The next halving arrives in 2028 with an even thinner subsidy and a fee market that will either have matured or will not have. That is the only variable that actually matters. Everything else is price.
