Three Pools, One Coin: The Mining Bill Bitcoin's Sideways Market Is Hiding

Somewhere between block 840,000 and this morning's flat tape, Bitcoin stopped being mined by "the network" and started being mined by three rooms full of ASICs. Foundry USA, AntPool, and ViaBTC have spent the last quarter clearing more than half of every block between them. No shock triggered it. No exploit. Just the slow, boring math of a post-halving subsidy grinding against a price that refuses to move. The chart has gone sideways so long that most desks have stopped watching it — and that is the tell. While everyone waits for a direction, the thing that actually secures the chain is quietly consolidating into fewer hands. Here is the arithmetic that matters. After the fourth halving, the block subsidy fell to 3.125 BTC. Hashrate keeps setting new all-time highs. Revenue per unit of hash keeps sliding toward multi-year lows. Those three facts cannot coexist forever. One of them has to break — and the sideways market is the pressure vessel where the break is being cooked. I have watched this movie before. Chasing the white whale in the 2017 ether rush taught me that the loudest signal is never the price — it is the plumbing. In 2017 it was the ICO queue and the Telegram scramble for allocations. In 2020 it was slippage in the yield aggregators. Right now it is the hashrate ribbon and the pool leaderboard, and almost nobody is reading either. That is the gap I want to close. Context: why the halving bill arrived late Bitcoin's security model is simple on a whiteboard and brutal in practice. Miners spend real electricity to guess a number. Whoever guesses first gets the block subsidy plus the transaction fees in that block. Every 210,000 blocks — roughly four years — the subsidy is cut in half. That is the halving, and it is a pay cut handed to the entire industry at once, on a schedule everyone knew about in advance and almost nobody hedged against properly. The 2024 halving dropped the subsidy from 6.25 to 3.125 BTC. The bull case was that fees would pick up the slack. Ordinals and inscriptions had juiced the fee market through 2023 and into 2024, and for a few weeks the mempool was a genuine substitute for the lost subsidy. Then the narrative faded. Runes did not sustain the fee base. By 2026, the fee share of miner revenue is back to a thin slice, and the subsidy is doing almost all the work again — except the subsidy is now half of what it was. So a miner's revenue per terahash has been cut twice in real terms: once by the halving, and once by the fact that every surviving operator responded to the halving by plugging in more machines. Hashrate is not a fixed pie. When revenue per hash falls, rational miners do not all shut down — the efficient ones expand to defend their margin, which pushes difficulty up and squeezes the inefficient ones out. It is a slow-motion arms race, and the arms are getting very expensive. That is why the pain took two years to surface. Difficulty adjusts every 2,016 blocks — roughly two weeks — and it lags the hashrate. An operator running thin on margin can survive a few bad adjustments by dipping into reserves. But you cannot dip forever. By the time the difficulty curve catches the hashrate curve, the operator who was "temporarily" underwater is permanently gone, and his rigs have been sold to someone with cheaper power. The halving did not kill the industry in April 2024. It scheduled the execution and let the calendar do the work. The Core: where the margin actually goes Let me put the numbers where they belong. When I audited the post-halving economics for a handful of mid-size operators last year, the break-even spread was the story. An S19-class rig — the workhorse of the 2020-2021 build-out — needs a price and a difficulty combination that no longer exists to stay profitable at retail power rates. Even the newer, more efficient machines only clear their cost of capital if they are running on sub-4-cent power or hedged energy contracts. At residential or even small-industrial electricity, the math is a slow bleed that turns into a cliff the moment difficulty ticks up again. That bleed is the reason the pool leaderboard looks the way it does. Mining rewards are variance, and variance kills small players. A solo miner with ten machines can go weeks without a block and starve. A pool smooths that variance, which is why everyone joins one — and why the pools that already have scale keep getting bigger. Foundry USA, AntPool, and ViaBTC are not dominant because they are sinister. They are dominant because they are the rational answer to a variance problem that the halving made worse. Decentralization on Bitcoin has quietly become a rounding error in a spreadsheet of pool market share. Here is the part the price chart hides. Hashrate is still climbing. If miners were truly capitulating — the classic "miner capitulation means bottom" signal — hashrate would fall as rigs went dark. It is not falling. It is rising while revenue per hash falls. That means the marginal rig being plugged in is not being paid for by Bitcoin alone. It is being paid for by something else: cheap stranded energy, debt raised against treasury holdings, or — increasingly — a pivot toward AI and high-performance compute that treats Bitcoin mining as a flexible load rather than a pure bet on BTC. That is the structural change nobody priced. A large share of new hashrate is now subsidized by data-center economics that have nothing to do with the block reward. The miners who survive the chop are not the ones with the best BTC thesis. They are the ones who can sell compute to someone else when BTC revenue dips — flipping their megawatts between model training and hashing depending on which pays more that week. The ones who cannot do that are the ones selling coins into every rally to cover the gap. And the fee market — the supposed savior — never showed up at scale. Inscriptions were a spike, not a base. The fees that propped up miner revenue in early 2024 were a sugar high, and the industry is now living on the subsidy alone. When I ran the fee-to-subsidy ratio for a sample of recent blocks, it looked almost identical to the post-2020 baseline: fees are a garnish, not a meal. Anyone who built a mining model on the assumption that fees would replace the halving built on sand, and the tide has already come in. So where does the value actually accrue? Not to the miners, and not to the holders who think a rising hashrate is automatically bullish. It accrues to whoever controls the pool template, because the pool decides which transactions get included and in what order. Three pools clearing half the blocks is not a 51% attack — it is worse in a subtler way. It is a veto. A coordinated policy decision by two of those three could censor transactions at the template level, and the network would keep producing blocks as if nothing happened. The chart would not even flinch. This is the same lesson the RWA crowd keeps refusing to learn on the other side of the market. Everyone wants to tokenize Treasuries and real estate onto a public chain and call it institutional adoption. But institutions do not need your permissionless validator set. They need finality, reversibility, and someone to call when a transaction goes wrong — and the more they get those things, the less the chain resembles the thing that made it valuable. Bitcoin's mining concentration is the same trade-off wearing a different hat: the more the security budget consolidates, the more "decentralized consensus" becomes a consensus of three well-capitalized companies with compliance departments and a shared interest in not upsetting anyone who matters. I have been hunting spreads while the market sleeps for fifteen years, and the pattern is always identical. The crowd watches the price. The professionals watch the capacity. Right now the capacity is telling you that the mining industry has already repriced itself for a world where the subsidy is a rounding error — and it did it without the price ever confirming the thesis. Speed kills slower than greed, but it still kills, and the slowest thing in this market is the realization that the supply side has already changed shape. The Contrarian Angle The consensus trade is that a flat market is a waiting room. Accumulate, wait for the breakout, ride the expansion. I think that is backwards for Bitcoin specifically. The sideways chop is not a pause in the mining story — it is the mining story. Every week of flat price is a week of forced distribution from operators who cannot hold, into the hands of whoever is still solvent enough to absorb it. That is why the range keeps holding: there is a steady, price-insensitive seller on one side and an ETF-driven bid on the other, and they are roughly matched. The chop is not indecision. It is a settlement. The blind spot is the "capitulation equals bottom" heuristic. It worked in 2018 and 2022 because hashrate fell when miners died. This cycle, hashrate does not fall, because the marginal miner is subsidized by AI compute and cheap power, not by BTC. So the signal that used to mark the bottom is now disconnected from the price. If you are waiting for a hashrate collapse to tell you when to buy, you are waiting for a phone that will not ring. Volatility is just noise until it becomes signal — and right now the signal is in the pool share, not the candles. There is a second-order risk almost nobody is modeling. The pools that dominate hashrate are also the pools that run the most compliance-sensitive infrastructure. If a regulator decides that a mining pool is a money transmitter — a live question in more than one jurisdiction — the largest pools are the most exposed, precisely because they have the most to lose. Concentration is not just a censorship risk. It is a single point of regulatory failure that the whole network now leans on. The Takeaway Watch three things over the next quarter, and none of them are the price. Watch the difficulty adjustment for the first real decline — that is the moment the AI-subsidy thesis cracks and the hashrate story finally turns. Watch pool share, because a fourth entrant regaining ground would be the only genuine decentralization story left on the board. And watch miner treasury sales, because that is the supply that decides whether this chop resolves up or down. The halving bill came due two years ago. The market is only now deciding who pays it — and the three pools holding the pen are not in a hurry to show you the invoice.