Why rigs are going dark
Miners are working through one of the toughest margin environments the industry has faced in years.
According to a recent breakdown, hash revenue for large public miners has fallen from about $55 per petahashes (PH) per day in Q3 to roughly $35 per PH/day today. Their median all-in cost sits near $44 per PH/day. In other words, a significant part of the sector is now mining at a loss.
At the same time, the network hashrate is hovering around 1.0-1.1 zettahash (ZH) per second, which means competition for each block is near record highs.
The punchline is return on investment (ROI): Even brand-new machines now show payback periods above 1,000 days, while the next halving is roughly 850 days away. If nothing changes, many miners buying hardware today may struggle to earn it back before the next halving unless market conditions improve.
This guide walks through how miner economics work in 2025, how to check whether your own machines are underwater and what options you realistically have if they are.
How miner economics work in 2025
Post-halving, every miner is fighting over a smaller pie.
The block subsidy dropped from 6.25 Bitcoin (BTC) to 3.125 BTC in the 2024 halving, cutting the main component of miner revenue in half overnight.
With around 144 blocks per day, that is about 450 BTC in new issuance daily plus fees.
Meanwhile, the network’s hashrate has climbed into the zettahash zone at around 1.0+ ZH/s on recent seven-day averages.
The result is an all-time low hash price, which is the USD revenue per PH/day of hashpower. Some crypto publications and other trackers put recent levels around $35-$38 per PH/day or roughly $0.03-$0.04 per terahash (TH) per day.
Against that, miners juggle:
Capital expenditure (capex): Application-specific integrated circuit machines (ASICs), transformers, racks, networking and land.
Operating expenditure (opex): Power price per kWh, hosting margin, cooling, maintenance, debt service and staff.
To stay alive, you need to clear two hurdles:
Cash flow test: Is daily revenue above daily operating costs at today’s hash price and power rate?
Payback test: Can the rig reasonably earn back its purchase price before the next halving or major hardware obsolescence?
These two metrics tend to be the most useful benchmarks for most setups.
Did you know? In mining, a kilowatt hour (kWh) is the unit you pay for on your electricity bill. A miner drawing four kW consumes four kWh every hour, which makes kWh the metric that ultimately determines your real daily and monthly operating cost.
Why even new-gen rigs struggle to break even
If you are running modern hardware, this is where the story turns uncomfortable.
The current top tier, including machines like Bitmain’s Antminer S21 and the Whatsminer M60 series, delivers around 17-22 joules per terahash (J/TH). It is a major jump from older generations and is now generally treated as the minimum standard for serious-scale deployments.
On paper, that level of efficiency should translate into comfortable margins. In practice:
At a hash price of $35-$38 per PH/day, even the most efficient rigs barely cover electricity costs for miners paying mid-range industrial tariffs.
Analysts estimate about $40 per PH/day as a common break-even level for many operations. Below that mark, every extra hour online eats into reserves.
TheMinerMag and other trackers now show ASIC payback periods stretching beyond 1,000 days at current hardware prices and revenue, which is longer than the time left until the next halving.
Some profitability guides suggest that, at these power rates, buying spot BTC can be more straightforward than mining, though the choice depends on individual conditions.
That is why rigs are going dark. In many setups, every extra block of uptime deepens the losses.
Did you know? A miner’s joules per terahash (J/TH) rating shows exactly how much energy it uses to produce hashing work. A lower J/TH means the machine performs the same terahash for less electricity, which makes it the single best indicator of ASIC efficiency.
How to check if your machines are underwater
Here is a simple framework you can run in 15 minutes.
Collect your numbers:
ASIC model and hashrate
Efficiency (J/TH) from the manufacturer’s spec sheet
All-in power price per kWh (energy, demand charges and hosting markup)
Pool fee and any site-level fees.
Estimate daily revenue:
Take your total hashrate in PH or TH and multiply it by a current hash price feed, such as $35-$38 per PH/day.
If you prefer TH units, remember that $35 per PH/day is the same as $0.035 per TH/day.
Calculate daily power cost:
Convert efficiency to power draw: (J/TH x hashrate in TH) ÷ 1,000 = kW
Multiply kW x 24 x kWh price
Add a 5%-10% buffer for cooling, networking and transformer losses.
Run the cash-flow test:
If revenue is lower than power cost, you are burning cash every day you stay online.
Stress test your setup by checking whether your numbers still hold if the hash price drops 10% and difficulty rises 10%.
If that scenario pushes you negative, you are effectively relying on a short-term BTC moonshot.
Run the payback test:
Take your ASIC purchase price and divide it by net daily profit, which is revenue minus operating costs.
If payback exceeds the time to the next halving, which is about 2.3 years from today, treat any new hardware purchase as a speculative bet rather than a grounded business investment.
If both tests fail, the setup often resembles a costly form of dollar cost averaging rather than a sustainable mining operation.
Your options when mining no longer pays
If the math looks rough, you still have a few levers you can pull.
Throttle or selectively curtail
Underclock machines, shut down the worst performers or run only during off-peak tariff windows. In some markets, grid operators even pay large sites to curtail during stress periods.
Chase cheaper electrons
For hosted miners, this can mean renegotiating contracts or moving to facilities with lower blended power rates. At an industrial scale, the trend is toward behind-the-meter renewables, flared gas and other stranded energy sources that can undercut grid prices.
Repurpose the site
Some operators are experimenting with AI and general high-performance computing workloads, renting spare capacity to inference or rendering clients. It is not a drop-in replacement, since cooling, networking and customer relationships all change, but it can turn a stranded substation into a revenue-producing data center.
Consolidate or exit
For some operators, selling rigs or consolidating can be more practical than continuing through another difficulty epoch.
What shutdowns mean for future miners and for Bitcoin
Miner pain does not automatically translate into protocol risk.
Historically, when enough operators shut down, difficulty adjusts downward and lifts margins for the survivors. The current cycle is more complicated because large public miners with low power contracts and hedging strategies can endure longer, which slows the clean-up.
For anyone considering mining in 2025, the bar is now clear:
Truly cheap power, roughly $0.06 per kWh all in or better
Current-gen efficiency, since sub-20-J/TH hardware is no longer optional
Discipline, with regular break-even checks and a willingness to switch off when the numbers stop working.
For Bitcoin itself, rolling waves of miner shutdowns have so far looked more like a reset, where capital and energy move from inefficient operators to leaner ones.
The uncomfortable takeaway for smaller players is simple: For many smaller operators, the economics often tilt in favor of buying BTC rather than mining, though this varies by power rates and hardware efficiency.
