The next BTC halving is 312 days out. The miner stress test starts now.
312 days to the 5th BTC halving. Subsidy drops 3.125 to 1.5625 BTC at block 1,050,000. CleanSpark, Riot, and Marathon will all face a hashprice that needs to clear $0.062/TH/day to break even.
Bitcoin’s fifth halving is currently scheduled for approximately 7 April 2028, at block height 1,050,000 — 312 days from today based on the rolling block-time average of 9 minutes 47 seconds tracked by mempool.space. The subsidy will drop from 3.125 BTC to 1.5625 BTC per block. Daily issuance falls from 450 BTC to 225 BTC. At Tuesday’s spot of $101,840, that is the difference between $45.8m and $22.9m of new supply per day reaching miners’ wallets. Subtract the $22.9m and the marginal seller in the asset class changes character overnight. The countdown is now short enough that miner balance sheets are being repositioned in real time, and the equity market is already trading the implied stress. CleanSpark, Riot Platforms, and Marathon Digital have each underperformed BTC by 18-31% year-to-date despite spot trading near six figures.
What is at stake is whether the public miner sector survives the next halving with its current capital structure intact, or whether the 2024-2026 ASIC capex cycle proves to have been mistimed against a halved subsidy and a hashprice that has not yet recovered. The cleanest measure of miner stress is hashprice — the revenue per terahash per day, currently $0.062 per the Luxor hashprice index. At the halving, that number mechanically halves unless transaction fees double or spot price doubles. Neither is the base case. The miners that came into 2026 with the wrong fleet, the wrong power contracts, or the wrong debt covenants are going to be visible by Q1 2027.
The math of the subsidy drop
The subsidy schedule is enforced by every full node validating the block reward at each height. The fifth halving cuts the per-block subsidy from 3.125 BTC to 1.5625 BTC. That is roughly 99.7% of total miner revenue at current fee levels — transaction fees averaged 5.4% of block reward over the trailing 90 days per mempool.space, but ranged from 1.2% (low-mempool weekend lull) to 18% (the 9 March ordinals spike). At today’s spot and today’s hashrate of 712 exahash per second per Glassnode, the post-halving daily issuance value falls from $45.8m to $22.9m. Total miner revenue, fees included, falls from approximately $48.3m to $25.2m per day, assuming fees grow proportionally.
| Metric | Today | Day after halving (BTC flat) | Day after (BTC +50%) |
|---|---|---|---|
| Block subsidy (BTC) | 3.125 | 1.5625 | 1.5625 |
| Daily issuance (BTC) | 450 | 225 | 225 |
| BTC spot (USD) | $101,840 | $101,840 | $152,760 |
| Daily subsidy revenue | $45.8m | $22.9m | $34.4m |
| Daily fee revenue (est) | $2.5m | $2.5m | $3.8m |
| Hashprice (USD/TH/day) | $0.062 | $0.034 | $0.051 |
| Breakeven J/TH (at $0.05/kWh) | ~28 J/TH | ~17 J/TH | ~24 J/TH |
Hashprice is the only number that matters
Hashprice — revenue per terahash per day — collapses everything into a single comparable number. Today’s $0.062/TH/day is roughly the 30th percentile of the 2024-26 range. The 2024 post-halving low was $0.038, hit in September 2024 when difficulty had risen 14% and spot was at $58,000. The current level is constructive but not loose; miners running 25-30 J/TH ASICs (Antminer S21 generation) clear above $0.04/TH/day net of power at $0.05/kWh. Miners running 38+ J/TH legacy fleets (S19j Pro and older) clear barely above breakeven. At halving, those numbers compress by 50% before considering any difficulty adjustment.
Difficulty is the wildcard. If hashrate stays at 712 EH/s through the halving, difficulty stays flat and hashprice halves on a 1:1 basis with the subsidy cut. If hashrate falls 15-20% post-halving — the typical pattern as marginal miners capitulate — difficulty adjusts down two weeks later and surviving miners recover roughly 15-20% of the hashprice loss. The 2024 halving saw a 6% hashrate drop in the first month, milder than 2020’s 22% drop because newer ASIC efficiency cushioned the blow. The 2028 halving is expected to see a similar 8-12% drop based on the current fleet age distribution per Cambridge’s CBECI dashboard.
Public miner balance sheets, ranked by halving readiness
The three large US-listed public miners enter the halving with very different positioning. CleanSpark holds the cleanest fleet — average efficiency 19.4 J/TH per their Q1 disclosures — and runs essentially all S21 generation hardware. Their hashprice breakeven at $0.045/kWh power is roughly $0.022/TH/day, which means they clear the halving under almost any spot scenario above $80,000. Riot is mid-pack: fleet efficiency 23 J/TH, but their Rockdale and Corsicana sites have power contracts at $0.029/kWh which gives them the lowest cash cost per coin among the listed group. Marathon is the question mark: fleet efficiency 26 J/TH, more geographically dispersed, and a larger BTC treasury position which means their P&L is increasingly a leveraged bet on BTC price rather than operational margin.
The smaller listed miners are a different category. Hut 8, Bitfarms, Iris Energy, TeraWulf, and Core Scientific all run fleets with average efficiency in the 26-32 J/TH band. Their halving math is brutal at flat spot: hashprice of $0.034/TH/day against power-included breakeven of $0.027-0.031 leaves operating margins in the low single digits. Any operational disruption, any power-price spike, or any 10% drop in BTC spot puts the marginal miner of the group at negative cash margin. The historical pattern suggests two to three public miners will either be acquired or restructure within twelve months of the halving. The 2024 cycle saw Core Scientific emerge from Chapter 11 and consolidation in private mining; the 2028 cycle will produce a similar shakeout, larger in dollar terms because the sector is larger.
The fee market has not done what it needed to
For seventeen years the Bitcoin community has talked about the transition from subsidy revenue to fee revenue. In practice it has not happened. Average transaction fees as a percentage of block reward sit at 5.4% over the trailing 90 days, down from a peak of 23% during the April 2024 ordinals frenzy. The Bitcoin Runes protocol launched at the 2024 halving has produced sporadic but not sustained fee demand. Lightning Network capacity at 5,400 BTC handles real economic activity but generates negligible base-layer fees. The mempool spends most days below 50,000 transactions queued, well within block capacity.
That matters because the long-term miner security model requires fees to eventually replace subsidy. At today’s fee level of $2.5m per day, fees would need to grow roughly 18x to fully replace the post-halving subsidy at flat spot. That growth path requires either a sustained spike in on-chain activity from new protocols, a much larger ordinals-style speculation cycle, or a structural change in how Bitcoin is used as a settlement layer. None of those is the base case for 2028. The likely outcome is that the post-halving security budget falls in real dollar terms, hashrate compresses 10-15%, and the network operates at a lower aggregate security level until the next bull cycle bids spot high enough to restore miner economics.
The private miner cohort is the unseen variable
Public miners account for roughly 28-32% of network hashrate depending on whose attribution model you trust. The remaining 68-72% is private and largely unrated — Russian and Kazakh industrial-scale operations, US private miners like Cipher and Stronghold’s non-public peers, Middle Eastern sovereign-linked operations, and a long tail of smaller operators in low-power-cost geographies. That cohort enters the halving with very different economics than the listed sector. Russian operators with sub-$0.03/kWh hydroelectric or stranded-gas power survive the halving comfortably at any spot above $50,000. Middle Eastern operations operating effectively at-cost as part of national diversification strategies have no profit-driven capitulation point. Conversely, US-based private hosting facilities running older fleets on commercial-grade power are the most vulnerable cohort in the entire universe, and they are not visible in any public dataset until they shut down.
That asymmetry matters for the post-halving hashrate trajectory. The public miners we can model are the ones least likely to capitulate first; the private cohort that capitulates first is invisible until the difficulty adjustment data confirms it has happened. The 2024 halving saw a 6% network hashrate drop, almost entirely from private miners — public-miner hashrate actually grew through the halving as listed operators absorbed the freed-up power capacity. The same pattern is likely to repeat in 2028, with the additional wrinkle that AI hosting demand for grid capacity is now competing aggressively with mining hosting demand. Power-contract renewals in 2026-27 will be the leading indicator of which private miners survive.
The AI-data-center pivot is the wildcard
Roughly a third of public miner capex over the last twelve months has redirected toward AI/HPC hosting. Iris Energy’s Childress facility, Core Scientific’s CoreWeave contract, TeraWulf’s Lake Mariner expansion — these are bets that the higher-margin AI hosting business can cushion the halving stress. The unit economics are real: AI hosting generates $1.50-2.50 per kWh of gross revenue against $0.30-0.50 for mining at current hashprice. The catch is that the conversion capex is high — roughly $1.5-2.0m per MW for AI-grade liquid cooling, transformer upgrades, and tier-3 networking, against $0.4-0.7m per MW for mining-grade buildout. The miners that can pivot meaningful capacity will outperform; those that cannot will be squeezed by both the halving and the rising AI-power competition for grid capacity.
What to watch from here
Three variables matter. First, public miner Q2 and Q3 earnings calls — listen for guidance on fleet refresh capex and AI hosting conversion timelines. A miner that has not committed to a fleet refresh by end-2026 enters the halving with structurally non-competitive economics. Second, the hashprice index — watch for any sustained drop below $0.050/TH/day, which is the level where the marginal listed miner stops generating cash margin and balance-sheet stress becomes visible. Third, fee market behaviour — any sustained mempool backlog above 200,000 transactions queued would signal that a new fee-generating protocol has launched and could change the post-halving math meaningfully.
For asset allocators, the halving math has secondary implications beyond mining equities. The subsidy cut removes 225 BTC of daily issuance, or roughly $22.9m at flat spot — a structurally meaningful reduction in new supply that historically has correlated with spot strength 6-18 months post-halving. That said, the four prior halvings each saw spot weakness in the immediate three months pre and post, as miners front-ran the event by selling treasury and as the difficulty adjustment cycle played out. Position accordingly. Track the live halving countdown on our halving page, the hashprice and miner-flow data in the market hub, and miner earnings dates on the events calendar.
The honest read: the 2028 halving will reveal which miners actually have a sustainable business and which were riding a 2024-26 capex cycle that did not produce the operational efficiency they sold to capital markets. The countdown has started. The first sign of stress will not be price-driven — it will be a fleet-utilisation announcement in a 10-Q. Read the filings, not the press releases.