Bitcoin Mining Margins in 2026: Anatomy of a Squeeze

Bitcoin mining has rarely been harder to make money in than it is right now. Bitcoin trades around $59,000 after sliding from an intraweek high near $72,840 earlier in June, the network hashrate sits just under 1 zettahash per second, and difficulty has just printed another record. A year ago Bitcoin changed hands near $107,000, so it is roughly $48,000 cheaper today, and that single fact explains most of the pain in the sector. Put it all together and the margin miners earn for securing the network has been squeezed to one of its thinnest readings since the last halving. This explainer breaks down what a mining margin actually is, the inputs that move it, and why mid-2026 has become a survival test for everyone from home hobbyists to billion-dollar public operators.

What a mining margin actually is

A mining margin is revenue minus cost, but both sides have moving parts that price charts ignore. Revenue has two components: the block subsidy, the newly issued Bitcoin set at 3.125 BTC per block since the April 2024 halving, and the transaction fees users pay to get into a block. A miner earns a share of that revenue in proportion to the hashrate it contributes, after the mining pool takes its cut.

Costs are led by electricity, then hardware depreciation, data-center or hosting overhead, pool fees (typically 1% to 2.5%), staff, and financing. The gross margin most operators quote is simply revenue minus power. The all-in margin, the one that decides whether a business survives, subtracts everything else, including the machines wearing out. One more factor sits on top: uptime. A site that is curtailed during peak grid demand earns less Bitcoin, but it can collect demand-response payments that quietly subsidize the margin. Because revenue is denominated in BTC while most costs are denominated in dollars, a miner is effectively running a leveraged bet on the Bitcoin price every day it keeps the rigs on.

Hashprice, the number that decides everything

If you track one metric, track hashprice: the daily revenue a miner earns per unit of hashrate, usually quoted in dollars per petahash per second per day. It folds price, subsidy, fees, and difficulty into a single figure, which is why operators live and die by it. Rising difficulty pulls hashprice down, because the same machine wins a smaller share of each block; a rising Bitcoin price or a fee surge pushes it back up.

On June 8, 2026, hashprice sat at $28.94 per PH/s per day, down 11.3% in a week, according to Luxor’s Hashrate Index. It has drifted lower since, after difficulty jumped 7.15% on June 26, one of the largest upward moves of the year. For context, hashprice traded well above $50 during the rally that followed the 2024 halving. At the current reading, a large share of the global fleet is at or below cash breakeven, which is what makes the remaining inputs decisive.

The 2024 halving still casts a long shadow

Roughly every four years, or every 210,000 blocks, the block subsidy is cut in half. The fourth halving landed at block 840,000 on April 20, 2024, dropping the subsidy from 6.25 BTC to 3.125 BTC. Overnight, the single largest line of miner revenue was halved.

What kept many sites afloat that week was a brief fee windfall. The halving block alone collected 37.67 BTC in fees, worth about $2.4 million, because the Runes token protocol launched in the same block. That spike faded within days. On a typical day in 2026, fees make up only a low single-digit percentage of block revenue, leaving the subsidy to do the heavy lifting. The lesson for margins is blunt: subsidy cuts are permanent, fee spikes are temporary, and the gap has to be filled by a higher Bitcoin price or cheaper, more efficient production.

Why electricity is the make-or-break input

Power is the largest variable cost in mining and the one input an operator can shop for. It is quoted in cents per kilowatt-hour, and the gap between 3 cents and 8 cents is the gap between a healthy margin and a shuttered site. The table below shows the breakeven power price for popular machine classes at the early-June hashprice of about $29 per PH/s per day. Pay more than the figure shown and the machine loses money on electricity alone, before any other cost.

Two caveats. This is power-only breakeven; true all-in breakeven is higher once depreciation, hosting, and pool fees are added. And every figure moves daily with hashprice and the Bitcoin price. The ranking, though, is stable: the more efficient the machine, the higher the power price it can survive.

Rig efficiency and the depreciation trap

Machine efficiency, measured in joules per terahash (J/TH), is the second big lever, and lower is better. Bitmain’s S21 XP Hydro runs at 12 J/TH and 473 TH/s, while the newer S23 generation pushes the frontier toward 9.5 J/TH. The S19 machines that dominated the last cycle sit far behind, at 21.5 to 34.5 J/TH.

At today’s hashprice, that gap decides who is mining and who is parked. An S19j Pro needs power below roughly 4 cents per kWh just to break even on electricity, which prices most of the older fleet out unless it is plugged into stranded or curtailed supply. For a home miner paying residential rates of 12 to 20 cents per kWh, only the very newest hardware clears breakeven at all, and even then heat reuse, not profit, is often the real reason to run a rig. The deeper trap is depreciation. ASIC miners are wasting assets: they lose share as the network grows and lose resale value as better machines ship. A rig bought near the top of the last cycle can be worth a fraction of its price two years on, and that loss is a real cost the gross-margin headline never shows.

What the public miners’ books reveal

Listed miners publish detailed figures, and the latest CoinShares mining report put the weighted-average cash cost to produce one Bitcoin at about $79,995 in the fourth quarter of 2025. With Bitcoin now near $59,000, that average sits above spot, meaning much of the listed sector would be underwater on a cash basis at today’s price if nothing else changed. The spread between operators, though, is enormous.

Read these with care. The all-in numbers include depreciation, stock compensation, and overhead, and for hybrid operators they are distorted by AI buildout costs charged against a shrinking amount of mined Bitcoin. TeraWulf’s extreme figure reflects a company that has largely stopped optimizing for Bitcoin output. Cash cost is the cleaner read on pure mining economics, and even there the best operators (Hut 8 and IREN, near $50,000 to $58,000) sit a world apart from the laggards.

The AI pivot is rewriting the margin story

The cleanest answer to thin mining margins in 2026 has been to stop depending on them. Miners control exactly what the AI boom is short of: signed power contracts, substations, land, and cooling at gigawatt scale. So they are leasing it to AI tenants instead of pointing all of it at Bitcoin.

The sums are large. More than $70 billion in cumulative AI and high-performance computing contracts have now been announced across the listed mining sector, and S&P Global expects AI to supply as much as 70% of listed-miner revenue by the end of 2026, up from roughly 30% at the start of the year. Core Scientific already books 39% of revenue from AI colocation; TeraWulf has signed more than $12.8 billion in long-term contracts and saw HPC leasing overtake mining income in the first quarter; IREN landed a five-year, $9.7 billion AI cloud deal with Microsoft in late 2025. Some operators are even selling mined Bitcoin to fund the switch.

For margins, the effect cuts two ways. AI hosting offers contracted, dollar-denominated revenue at steadier and usually higher margins than mining. But every megawatt sent to GPUs is a megawatt not pointed at Bitcoin, and if the most efficient operators keep reallocating power, network security leans on a smaller, more concentrated group of miners.

The accounting and SEC overlay

Margins are not only an engineering problem; they are a reporting one. US-listed miners file 10-K and 10-Q reports with the Securities and Exchange Commission, and those filings are where cost-to-mine, fleet efficiency, and power deals become public and comparable. The SEC has also pressed crypto-exposed companies to spell out the risks behind those numbers.

A quieter rule change reshaped how the margins look on paper. Under FASB’s ASU 2023-08, effective for fiscal years beginning after December 15, 2024, companies measure their Bitcoin holdings at fair value, booking gains and losses each quarter rather than only writing them down. That makes a treasury-heavy miner’s earnings swing with the Bitcoin price, which separates the operating margin of mining from the mark-to-market of coins on the balance sheet. The practical takeaway when reading miner results: a strong quarter can be a price rally, not a cheaper kilowatt-hour.

What to watch in the second half of 2026

Three variables will decide whether margins recover from here.

• Bitcoin price: the single biggest lever. Every move in spot flows straight into hashprice and the breakeven math above.
• Difficulty and hashrate: with the network near 1 zettahash per second, sustained price weakness should eventually push inefficient capacity offline, easing difficulty and lifting hashprice for the survivors.
• Fees: a durable return of on-chain demand, through Ordinals, Runes, or something new, would cushion the post-halving revenue gap, though no operator should budget for it.

The structural story is simpler. The next halving, due in 2028, will cut the subsidy again to 1.5625 BTC, so the pressure on margins only builds from here. Miners who cannot reach the efficiency frontier or lock in sub-5-cent power are increasingly taking the other exit and pointing their substations at AI. Either way, the era of easy Bitcoin mining margins is over.

By the HOGE Wire editorial desk, covering mining and staking economics.