The fourth Bitcoin halving, which occurred in April 2024, reduced the block subsidy from 6.25 BTC to 3.125 BTC per block. For mining operations that had not structurally adjusted before the event, the revenue impact was immediate. For those that had — through hardware upgrades, power purchase agreements, or firmware-level efficiency improvements — the halving represented a compression of margins rather than their elimination. The difference between those two outcomes comes down to which variables in the profitability equation an operator has addressed and which they have left unchanged.
This post breaks down the post-halving profitability equation, identifies the variables operators can and cannot control, and focuses on where operational decisions have the most measurable impact on the bottom line.
What the Halving Actually Changed
The halving does not change Bitcoin’s price, the network’s transaction fee environment, or any individual operator’s electricity costs. What it changes is the number of BTC issued per block — and by extension, the baseline revenue that the entire mining industry competes for.
The immediate effect of a halving is a compression of hashprice: the amount of revenue generated per terahash per second per day across the network. When the block subsidy drops by 50%, and Bitcoin’s price does not simultaneously double to compensate, every terahash in operation earns less. Miners with higher operating costs relative to their revenue either reduce operations or exit, which eventually causes the network difficulty to adjust downward. This correction has followed every halving in Bitcoin’s history, but the timing and depth of the adjustment vary, and operators cannot rely on it as a planning assumption.
The 2024 halving followed the same structural pattern. Network hashrate had grown substantially in the months leading up to April 2024 as miners deployed new hardware ahead of the event. The resulting difficulty level, combined with the immediate 50% subsidy reduction, compressed hashprice significantly in the weeks following the halving. Operations running on older, less efficient hardware or carrying high electricity costs felt this most acutely.
The Mining Profitability Equation
Mining profitability is determined by a relatively small number of variables. Understanding which of them an operator can control — and to what degree — is the starting point for any post-halving strategy.
Revenue side:
- Bitcoin price (cannot be controlled)
- Network difficulty and hashrate (cannot be controlled)
- Block subsidy (cannot be controlled — currently 3.125 BTC per block)
- Transaction fees as a share of block reward (cannot be controlled directly)
- Pool fee structure (can be selected)
- Fleet hashrate output (partially controllable through firmware and hardware)
Cost side:
- Electricity rate in $/kWh (partially controllable through siting, PPAs, and curtailment participation)
- Hardware efficiency in J/TH (controllable through hardware selection and firmware optimization)
- Overhead — facilities, cooling, labor, maintenance (controllable through operational decisions)
The variables that operators cannot control represent the majority of revenue exposure. This is a structural feature of Bitcoin mining, not a temporary condition. It means that the operational levers available to improve profitability are finite, and the ones that exist deserve rigorous attention.
Hashprice: The Single Number That Ties It Together
Hashprice — typically expressed as USD per TH/s per day — is the most useful single metric for monitoring mining revenue conditions in real time. It captures the combined effect of Bitcoin’s price, network difficulty, block subsidy, and transaction fees in a single number that reflects what a terahash is actually worth on any given day.
During periods of high hashprice, hardware efficiency matters less because revenue is strong enough to cover higher operating costs. During periods of compressed hashprice — as experienced in the weeks and months following the 2024 halving — hardware efficiency becomes the central profitability variable, because the revenue side is fixed by market conditions and the only room to improve margin is on the cost side.
An Antminer S21 XP operates at a fundamentally different break-even point than an Antminer S19 XP. For instance, S21 XP runs at approximately 13.5 J/TH, while the S19 XP runs at approximately 21.5 J/TH. At $0.07/kWh, S21 XP spends roughly $0.023 per TH per day on electricity, compared to approximately $0.036 per TH per day for the S19 XP — about 59% more. In a compressed hashprice environment, that efficiency gap can be the difference between maintaining positive operating margin and mining at a loss.
Where Efficiency Gains Come From Without Replacing Hardware
Hardware replacement is the most direct path to better J/TH, but it is also the most capital-intensive. For operations that manage existing S19x fleets, the more immediate path to efficiency improvements is firmware-level optimization.
Stock Bitmain firmware applies fixed, conservative operating settings calibrated for the broadest possible range of conditions. This approach prioritizes stability and warranty compliance over peak efficiency. The consequence is that most Antminers running stock firmware operate at a voltage-frequency combination that is safe — not at the optimal point for each chip individually.
Custom Antminer firmware addresses this through per-chip and per-board tuning. By reading each chip’s actual frequency response and adjusting accordingly, custom firmware can reduce per-unit power consumption while maintaining the same hashrate output — improving J/TH without changing the hardware. In practice, efficiency improvements of 10–25% over stock settings are achievable depending on hardware condition, operating environment, and the tuning approach applied.
For a fleet of 500 Antminer S19j Pro units drawing an average of 3,050W on stock firmware, a 15% efficiency improvement reduces average draw to approximately 2,590W per unit. At $0.07/kWh, that is a reduction in daily power cost from approximately $2,561 to approximately $2,176 — a saving of $385 per day, or roughly $140,000 per year, with no capital expenditure on new hardware.
The relevance of that figure depends on hashprice. In a healthy hashprice environment, $140,000 per year in operational savings is meaningful but not decisive. In a compressed post-halving environment where margins are narrow, it can determine whether a fleet operates profitably or at a sustained loss.
Electricity Rate: The Variable With the Highest Leverage
Among the variables operators can influence, the electricity rate has more impact on profitability than any other single factor. A 1-cent reduction in $/kWh across a large fleet produces savings that compound continuously regardless of Bitcoin price or network conditions.
The levers available for reducing effective electricity cost vary by market and operation size:
Power purchase agreements (PPAs)
lock in a fixed rate for a defined period, providing revenue predictability and protecting against spot price increases. Large-scale operations in markets with abundant generation capacity — Alberta, Texas, Quebec, parts of the Pacific Northwest — can negotiate rates well below grid average through direct agreements with generators or utilities.
Curtailment and demand response participation
allow miners to reduce load during high-demand grid periods in exchange for credits or compensation from grid operators. In markets like ERCOT, where demand response programs are well-developed, curtailment revenue can meaningfully offset electricity costs, effectively lowering the net rate paid per kWh during non-curtailment hours. This approach requires firmware that can ramp fleet power up and down quickly and predictably in response to automated signals — a capability that stock Bitmain firmware does not support.
Behind-the-meter generation
involves co-locating mining operations with power-generation assets — natural gas, hydro, wind, or solar — and purchasing power directly at or near the cost of generation rather than at the retail rate. This model is increasingly common in new large-scale deployments and can produce electricity at costs that are structurally lower than any grid-connected operation can achieve.
Hardware Lifecycle Decisions in a Post-Halving Market
One consequence of a halving is that it changes the economic calculation around when to retire older hardware. A machine that was profitable at $0.07/kWh when the block subsidy was 6.25 BTC may fall below break-even at the same power rate with the subsidy at 3.125 BTC — depending on its efficiency and the prevailing hashprice.
The standard decision framework is straightforward: if a machine’s daily electricity cost exceeds its daily revenue, it should be shut down unless there is a credible near-term basis for improvement (a Bitcoin price increase, difficulty adjustment, or cost reduction). Holding unprofitable hardware online in anticipation of price recovery is a capital allocation decision, not an operational one, and the carrying costs are real.
For hardware that sits near the break-even threshold, firmware optimization can shift the calculation. An S19j running at 30 J/TH on stock firmware may be marginally unprofitable at a given hashprice. The same machine, tuned to 25 J/TH via custom firmware, may be marginally profitable. Whether that difference justifies continued operation depends on the operator’s specific power rate and hashprice at the time, but the adjustment is achievable without capital expenditure and reversible if conditions change.
For a detailed breakdown of how UMC OS tuning algorithms affect efficiency at the chip and hashboard level, see the existing guide on tuning algorithms in custom Antminer firmware.
Transaction Fees as a Growing Component of Revenue
The 2024 halving renewed attention on Bitcoin transaction fees as a share of block reward. With the subsidy now at 3.125 BTC, transaction fees represent a larger proportion of total miner revenue than at any previous point in the network’s history — not in absolute terms on any given day, but structurally as a percentage.
Fee revenue is highly variable. Periods of high on-chain activity — driven by Ordinals, BRC-20 token activity, or increased network usage — can produce fee spikes that significantly supplement the base subsidy. Periods of low activity can produce fee revenue that is nearly negligible. Operators cannot predict or control fee revenue, but it is a component of the revenue side that has become harder to ignore as the subsidy continues to decrease with each halving cycle.
The implication for long-term planning is that mining operations built entirely around subsidy revenue are structurally dependent on Bitcoin price appreciation to remain viable through successive halvings. Operations with the lowest possible cost structures are better positioned to remain profitable during periods of low fee activity and compressed hashprice.
Operational Efficiency Beyond Firmware
Firmware and electricity rates are the two highest-leverage variables, but operational overhead is a meaningful cost line item for large facilities. Labor, facilities management, cooling infrastructure, and equipment maintenance all contribute to the total cost of mining per BTC. Post-halving, operators running lean operations with high automation have a structural cost advantage over those with higher overhead.
Automation of routine fleet management tasks — firmware updates, performance monitoring, fault detection, and ramp control — reduces the labor cost per unit managed. For fleets running UMC OS, the RigRunner deployment tool and the full REST API available through UMC OS allow operators to manage large fleets with less manual intervention than stock firmware setups require. This does not produce the same magnitude of savings as electricity cost reduction or firmware efficiency gains, but it compounds with them.
Cooling infrastructure efficiency is another operational variable. Air-cooled facilities carry ongoing costs for fan maintenance, air handling, and thermal management of the building envelope. Immersion and hydro cooling systems have higher upfront capital costs but lower ongoing thermal management overhead, and they allow hardware to run at higher efficiency points because chip temperatures are lower. For new deployments being evaluated post-halving, the total cost of ownership calculation should include the cooling method as a variable, not just the electricity rate and hardware cost.