Bitcoin Mining Economics 2026: Hashprice, Power Cost, and Full-Cycle Margins

Bitcoin Mining Economics 2026: Hashprice, Power Cost, and Full-Cycle Margins

Short answer: There is no universal 2026 Bitcoin mining breakeven price. Profit depends on hashprice, fleet efficiency, facility overhead, realized uptime, electricity terms, hardware cost, financing, and whether the analysis counts depreciation. At a $32/PH/day hashprice, a 15 J/TH machine earns about $88.89 per MWh before site costs; a 30 J/TH machine earns about $44.44. The same power contract can therefore support one fleet and shut down another.

Mining analysis often starts with Bitcoin's price and ends with a cost-per-coin estimate. That shortcut hides the mechanism. A miner sells compute into a network-wide revenue market, pays for energy and infrastructure, and owns hardware that becomes less competitive as newer machines arrive.

This guide builds the economics in layers. It uses public information available through July 11, 2026 and company filings for the quarter ended March 31, 2026. The figures are historical or illustrative, not forecasts.

The seven inputs that drive mining returns

Each input can move independently. A more efficient fleet may produce fewer coins than expected if uptime is poor. Cheap electricity can be offset by high hosting fees. A cash-profitable mine can still lose money after hardware depreciation and financing.

Start with hashprice

Hashprice is expected miner revenue per unit of hashrate over time, usually shown as dollars or Bitcoin per petahash per second per day. It combines four network variables:

The April 2024 halving reduced the subsidy from 6.25 BTC to 3.125 BTC. Since then, rising difficulty and weak fee contribution have compressed Bitcoin-denominated revenue per unit of compute.

Hashrate Index estimated that USD hashprice reached an all-time low near $27.89/PH/day in February 2026. Its June 1 roundup reported spot near $32.56. These dated observations show the range of pressure, but a live underwriting model should update hashprice at the time of use.

Why a Bitcoin price target isn't enough

Suppose Bitcoin rises 20%. Hashprice won't necessarily rise 20% for long. Higher margins can bring idle machines back online, attract new deployments, and raise difficulty. More network competition then reduces each machine's expected Bitcoin production.

Fees can also break the relationship. A burst of transaction demand may lift miner revenue without a comparable Bitcoin-price move. In May 2026, however, Hashrate Index reported fees below 1% of rewards, so they offered little cushion.

Use hashprice for operating analysis and model Bitcoin price, fees, and difficulty separately when building scenarios.

Convert hashprice into revenue per MWh

Fleet efficiency turns compute revenue into power economics.

One petahash per second equals 1,000 terahashes per second. A fleet rated at 18 joules per terahash needs about 18 kilowatts to produce 1 PH/s. Over 24 hours, that is 432 kWh before cooling and electrical overhead.

The machine-level formula is:

Gross revenue per MWh = hashprice × 1,000 ÷ (24 × J/TH)

At $32/PH/day:

The final column is not a true breakeven. It is the tariff that consumes 100% of machine revenue before cooling, pool fees, staff, repairs, insurance, property, financing, or hardware cost.

That is why a fixed “17 J/TH redline” is misleading. At a higher hashprice or lower power cost, a 22 J/TH fleet can make cash. At a lower hashprice or expensive hosted rate, a 15 J/TH fleet can struggle.

Add facility overhead

ASIC efficiency is measured at the machine. The utility meter sees the whole site.

Power usage effectiveness, or PUE, expresses total facility energy divided by IT-equipment energy. A mine with 1.08 PUE uses 8% more facility power than its machines alone. Fans, pumps, transformers, networking, lighting, and climate all contribute.

Adjust the earlier formula:

Facility revenue per MWh = machine revenue per MWh ÷ PUE

An 18 J/TH fleet at $32 hashprice earns $74.07 per machine MWh. At 1.08 PUE, that becomes about $68.59 per facility MWh.

Now subtract variable costs:

This example does not claim a standard maintenance rate. It shows where each assumption belongs.

Realized hashrate matters more than installed hashrate

Miner updates use several hashrate labels:

Revenue follows realized hashrate. A 50 EH/s fleet averaging 42.5 EH/s is operating at 85% of the headline figure before considering pool luck.

Low realization can come from scheduled curtailment, extreme weather, machine failures, transformer limits, network issues, repair delays, or economic shutdowns. Some curtailment creates power credits, so lower production may still be financially rational.

Use this bridge:

Expected revenue = average realized PH/s × hashprice × days

Then compare expected pool revenue with reported mining revenue. Differences may reflect fee timing, pool method, Bitcoin valuation, curtailment, or incomplete definitions.

Two breakevens, two decisions

Mining companies correctly point out that depreciation doesn't decide whether a machine should run today. Investors correctly point out that hardware cost decides whether the original investment earned a return.

Both views belong in the model.

Dispatch breakeven

Dispatch breakeven includes costs avoided when machines switch off:

If revenue exceeds these costs, running may produce more cash than shutting down, even if the company reports an accounting loss.

Full-cycle breakeven

Full-cycle breakeven includes:

Use dispatch breakeven for hourly operating choices. Use full-cycle breakeven for buying hardware, building sites, valuing a miner, or measuring shareholder return.

What CleanSpark's filing reveals

CleanSpark's March 2026 Form 10-Q provides a useful split. For its owned facilities in the quarter ended March 31, it reported:

The owned fleet covered direct energy cost. It did not cover reported direct cost after miner depreciation on this measure.

This doesn't mean the company should have switched off. Depreciation would continue. It means the quarter's mining revenue did not recover the period's allocated machine cost under the filing's presentation.

It also shows why company comparisons must use consistent definitions. One miner may report power cost per coin, another cost of revenue excluding depreciation, and another an adjusted metric net of power credits.

MARA: cost per unit of compute versus cost per coin

MARA's first-quarter filing reported 2,247 BTC produced and $172.2 million in Bitcoin mining revenue. Average mined-Bitcoin value was $76,288. The company said global difficulty reduced production despite fleet expansion.

It reported purchased energy cost of $40,047 per BTC for owned sites and a broader cost per petahash per day of $27.60. That broader measure included purchased energy, third-party hosting and other energy costs, and cash operating and maintenance costs, divided by average operational hashrate under its definition.

These metrics answer different questions:

At a $32 hashprice, a $27.60 cash cost per PH/day leaves only $4.40 before costs excluded from the measure. At a $40 hashprice, the same operating base leaves $12.40. The sensitivity is immediate.

MARA also disclosed selling about 20,880 BTC during the quarter to fund operations, growth opportunities, and liquidity management. Treasury policy therefore belongs beside mining margin, not in a separate “holdings” story.

Power credits can change the quarter

Riot reported $111.9 million of Bitcoin mining revenue in the first quarter of 2026. Its self-mining cost of revenue before miner depreciation was $86.8 million before power curtailment credits and $65.7 million after $21.0 million of credits.

TeraWulf reported $14.0 million of gross power cost and $14.1 million of demand-response proceeds for the quarter, producing a net negative realized power cost under its presentation. Cold weather reduced mining consumption and increased demand-response proceeds at Lake Mariner.

These results show why net power cost can be volatile. A site may earn more by curtailing during a stressed quarter than during normal conditions. Analysts should separate:

A negative net power cost in one quarter isn't a permanent tariff. It is a combined power-market and operating result.

For a grid-focused framework, read the <a href="/insights/energy-grid-harmony-model-bitcoin-mining-analysis">Energy-Grid Harmony Model</a>.

The hosting-versus-owned-site decision

Hosted mining replaces much site ownership with a contract. The host may provide power, buildings, operations, and maintenance for a fee.

Hosted cost should include all-in fees, pass-throughs, deposits, curtailment terms, repair responsibility, and termination rights. A 6-cent hosting price isn't comparable with a 6-cent owned-site power price if the former includes operations and the latter does not.

Hosted contracts can also become uneconomic before they expire. If the fee is fixed while hashprice falls, the operator may owe more than the fleet earns.

Hardware return and the difficulty treadmill

An ASIC investment has four economic phases:

Delay hurts twice. The buyer loses early revenue and receives the machine after more network hashrate may have arrived. A machine can remain functional while becoming uneconomic at the buyer's power price.

Model hardware return with a declining share of network rewards, not constant daily Bitcoin production. A simple scenario should include:

Avoid claiming a fixed payback period based on today's hashprice. Forward hashprice markets can provide a market reference for part of the period, though they don't eliminate execution or basis risk. See <a href="/insights/cloud-hashrate-whale-derivatives-analysis">Cloud Hashrate Derivatives</a> for the hedging mechanics.

The AI/HPC opportunity cost

Mining companies increasingly discuss AI or high-performance-compute projects. The economic question isn't whether AI revenue sounds larger. It is whether a specific site can support the workload and whether risk-adjusted return exceeds continued mining.

Mining sites and AI data centers both need power, but AI often requires:

A miner cannot value every contracted megawatt at an AI-data-center multiple. Start with site suitability, signed customer commitments, required capex, financing, delivery milestones, and penalties.

For each MW, compare four uses:

The option to convert can be valuable. It should not be counted as completed AI economics before a project is financed and contracted.

Treasury strategy can dominate operating results

Mining companies may hold, sell, buy, pledge, lend, or hedge Bitcoin. The balance sheet can become more sensitive to Bitcoin than the mine itself.

Track a quarterly bridge:

Opening BTC + mined BTC + purchased BTC + returned collateral − sales − pledged or lent transfers − fees = closing BTC

Then add:

A miner can report positive mining gross margin while funding expansion through dilution. It can also report an accounting loss while its Bitcoin treasury appreciates. Shareholders own both the operation and the financing choices.

Use our <a href="/insights/bitcoin-miner-capitulation-hashrate-ribbon-2026">Miner Capitulation dashboard</a> to distinguish routine production sales from balance-sheet distress.

A reusable 10 MW mining model

Consider an illustrative 10 MW facility with 1.08 PUE, an 18 J/TH fleet, 92% realized uptime, $45/MWh electricity, 2% pool fee, and $2.50/MWh variable maintenance.

Machine load is 10 MW ÷ 1.08 = 9.259 MW. Expected hashrate is about 514.4 PH/s when online. At 92% uptime, average delivered hashrate is about 473.3 PH/s.

The example assumes the whole facility powers down during downtime, so electricity follows uptime. Fixed labor, rent, insurance, hardware, site capital, financing, and tax are not included.

At $28 hashprice, the site remains contribution-positive under these inputs but has little room for fixed cost. A less efficient fleet, higher PUE, or minimum power payment could erase the margin.

How to compare public miners consistently

Use a common worksheet rather than company-adjusted labels.

Operations

Unit economics

Capital and treasury

The <a href="/insights/top-bitcoin-miners-2026">Top Bitcoin Miners 2026</a> article applies a broader operating-quality framework to major public names.

Red flags in mining analysis

FAQ

What was Bitcoin mining hashprice in 2026?

Hashrate Index estimated an all-time low near $27.89 per PH/s/day in February 2026 and reported spot near $32.56 on June 1. Hashprice changes continuously, so every figure needs a date and source.

What electricity price is profitable for Bitcoin mining?

It depends on hashprice, J/TH, PUE, uptime, pool fees, and other costs. At $32 hashprice, an 18 J/TH machine produces about 7.41 cents of revenue per machine kWh before site overhead. At 1.08 PUE, that falls to about 6.86 cents per facility kWh before pool, maintenance, and fixed cost.

What is a good fleet efficiency in 2026?

Lower J/TH is better, but no single cutoff determines profitability. Compare the weighted operating fleet with current hashprice and all-in site power economics. A highly efficient machine can still lose money under an expensive hosting contract.

Why does cost per Bitcoin rise when electricity prices are stable?

Higher network difficulty can reduce Bitcoin production from the same hashrate. Stable power spending divided by fewer mined coins produces a higher cost per coin.

Should depreciation count in mining cost?

Yes for full-cycle investment return, no for a narrow decision about whether to run a machine for the next hour if depreciation cannot be avoided. Analysts should publish both views.

Can curtailment make power cost negative?

Demand-response proceeds can exceed gross power cost in a particular period, as TeraWulf reported for Q1 2026 under its presentation. That result depends on market events and contract rules and should not be assumed to repeat.

Is AI hosting always more profitable than Bitcoin mining?

No. AI can support longer contracts and higher revenue per MW, but it requires suitable sites, customers, reliability, fiber, cooling, and large capital spending. Compare signed project economics rather than sector headlines.

What metric matters most for a mining stock?

Start with cash contribution per PH/day after gross power and direct operating costs. Then add hardware depreciation, corporate expense, capex, debt, treasury activity, and dilution. No single operating metric captures shareholder return.

Source note and limits

Network observations come from Hashrate Index reports available through July 11, 2026. Company examples come from SEC filings for the quarter ended March 31, 2026. Company-defined metrics are not automatically comparable; this article identifies their scope where possible.

The $32 hashprice tables and 10 MW model are CryptosEyes scenarios. They are not current quotes, forecasts, or estimates of a named company. The model excludes taxes and fixed costs where stated.

What to read next

Read <a href="/insights/bitcoin-miner-capitulation-hashrate-ribbon-2026">Bitcoin Miner Capitulation in 2026</a> next to connect this unit-economics model with network shutdowns, difficulty adjustments, company liquidity, and treasury behavior.

The practical takeaway: translate hashprice into revenue per facility MWh, then count every cost before calling a miner profitable.

About the Editorial Team

CryptosEyes publishes independent, source-grounded market analysis. Historical figures are tied to dated sources, and scenarios state their assumptions.