Power Usage Effectiveness (PUE) is a facility-level metric that shows how much total electricity a site consumes for each unit of electricity delivered to its computing equipment. In a Bitcoin mining farm, the computing load is the ASIC fleet, and because ASICs typically run as a near-continuous load, any overhead captured by PUE compounds into measurable electricity cost over time. This article explains what PUE measures, how it differs from ASIC efficiency and pool-side hashrate, and how to translate a PUE figure into an electricity-cost comparison.

What Is PUE in a Bitcoin Mining Farm?

PUE originates from data-center engineering and is now defined under ISO/IEC 30134-2 as a facility key performance indicator comparing total facility energy consumption with IT-equipment energy consumption (ISO). In a mining context, "IT equipment" primarily means the ASIC fleet, along with supporting IT equipment such as network devices. External cooling systems and power-distribution losses are facility overhead, not IT loads. PUE does not describe hashrate, mining revenue, or hardware quality. It describes how much additional electricity the site needs to keep that hashrate running—through cooling, ventilation, power distribution, and related infrastructure.

A site can deploy highly efficient ASICs and still spend substantially more on electricity than a comparable site if its cooling and electrical infrastructure consume a large share of additional power. PUE isolates that overhead so it can be measured, compared, and managed separately from the hardware itself.

PUE Formula: Total Facility Energy vs. ASIC Energy

The standard formula is:

PUE = Total Facility Energy ÷ IT Equipment Energy

Rearranged, total facility energy equals IT equipment energy multiplied by PUE. PUE is dimensionless, and a value closer to 1.0 indicates that less non-IT energy is required to support the same computing load.

The overhead that PUE captures can be expressed as:

Non-IT facility energy relative to IT energy = PUE − 1

For example, a PUE of 1.10 means non-IT facility energy equals 10% of IT-equipment energy; a PUE of 1.25 means it equals 25%. This overhead is not limited to cooling. Depending on how the site boundary is defined and metered, it can include ventilation fans, pumps, transformers, power-distribution losses, uninterruptible power supplies where present, lighting, and other facility loads.

Why Cooling Raises Electricity Cost

Where electricity is billed at or upstream of the facility meter, the operator pays for total facility energy, not just the energy delivered to the ASICs. For a flat electricity rate, the general calculation is IT Equipment Energy in kWh × PUE × Electricity Rate per kWh. The ASIC-based version below assumes the ASIC fleet accounts for the entire IT load. Use the average ASIC power over the modeled operating hours and the PUE for that same period and measurement scope. If other IT loads are material, use the full IT-equipment energy instead:

Electricity Cost = (ASIC Load in kW × PUE × Operating Hours) × Electricity Rate

Because ASIC fleets typically run continuously, even a modest increase in PUE accumulates into a meaningful daily and annual cost difference. This is the central reason PUE matters for mining economics: it is not an abstract engineering ratio but a direct multiplier on billed electricity.

PUE 1.10 vs. 1.25: A 10 MW Cost Example

The following example is illustrative only and uses assumed figures rather than a market quote. Assume a 10 MW ASIC load representing the entire IT load, continuous 24-hour operation, and an electricity rate of $0.06/kWh.

At the same 10 MW ASIC load and the same electricity rate, moving from a PUE of 1.10 to 1.25 raises total electricity cost by approximately 13.64% per day, or roughly $788,400 over 365 days of uninterrupted operation at an unchanged rate. Actual results will differ based on uptime, curtailment schedules, tariff structure, demand charges, and the exact point at which electricity is metered. The example is meant to demonstrate the scale of impact, not to represent a typical mining-farm PUE or electricity rate.

PUE Is Not ASIC J/TH or Pool-Side Hashrate

PUE is frequently discussed alongside ASIC efficiency and pool-side hashrate, but the three measure different things and should not be substituted for one another.

ASIC efficiency, expressed in J/TH, measures the energy a miner or miner fleet consumes relative to the hashrate it produces. It is a hardware-level metric, typically sourced from manufacturer specifications or from compatible measured power and hashrate data taken over the same period and the same equipment boundary.

Pool-side hashrate is an estimate the mining pool calculates from accepted shares over a given time window; it is not a direct electrical measurement. ViaBTC's documentation distinguishes a mining machine's local hashrate from the pool-estimated hashrate shown on the pool dashboard, noting that the latter is derived from accepted work and elapsed time rather than measured power draw (ViaBTC). Deriving an ASIC's J/TH from pool-side hashrate alone is therefore not reliable unless a matching, compatible power measurement is also available for the same time window.

An operator can use PUE alongside J/TH for internal facility planning—for example, approximating facility-inclusive energy per terahash as J/TH multiplied by PUE—but this combined figure should not be presented as the ASIC's rated efficiency. It is a facility-level planning calculation and is only meaningful when the underlying device-efficiency measurement, ASIC-load measurement, and PUE cover the same equipment boundary and the same operating period.

How to Measure Mining-Farm PUE Correctly

A credible PUE figure depends on consistent measurement, not just the ratio itself. Four conditions support a comparable calculation:

1. Total facility energy and IT-equipment energy cover the same site. The IT measurement includes the ASICs and other IT loads; whole-miner input measurements include integrated fans. External cooling equipment and distribution losses remain facility overhead rather than being freely assigned to the IT denominator.
2. Facility energy and IT-equipment energy are measured over the same period, rather than combining an instantaneous power reading with a monthly energy total.
3. Energy is recorded in kWh or MWh over a defined interval, not inferred from nameplate capacity.
4. The reporting party discloses the measurement level, meter locations, and included loads. An upstream meter may capture distribution losses along with IT consumption, so its reading should not be treated as equivalent to equipment-input metering without accounting for the difference.

These distinctions follow The Green Grid's guidance on measurement levels and component classification (PUE: A Comprehensive Examination of the Metric).

Published figures also require context. Soluna reported a PUE of 1.03 for Project Sophie for the year ended December 31, 2024 (SEC filing). This is a project-specific result, not a company-wide or mining-industry benchmark. Confirm measurement scope and reporting period before comparing it with another site.

What to Check in a Hosting or Power Agreement

PUE only affects an operator's cost when it corresponds to the way electricity is actually billed. Before applying a PUE figure to a cost model, it is worth confirming:

• Whether electricity is billed at the site meter, in which case total facility energy—including cooling and infrastructure overhead—is billed directly.
• Whether a hosting agreement charges a single all-in rate per kWh or per MW that already incorporates cooling and infrastructure energy, in which case applying a separate PUE adjustment on top would double count that cost.
• Whether cooling electricity is billed as a separate line item; if so, it should be added once to the ASIC electricity cost, not added again through a PUE multiplier.
• Whether the tariff includes demand charges, curtailment payments, or fixed fees that are not captured by a simple per-kWh calculation.

Clarifying these points before modeling costs prevents the most common error in applying PUE: counting the same overhead twice, once in the contracted rate and once in a separately applied PUE figure.

Using Mining-Revenue Estimates Alongside Facility Electricity Costs

PUE-based cost modeling is one half of a mining-economics analysis; estimated revenue is the other, and the two should be kept in separate calculations rather than merged into a single blended figure. A mining-pool profit calculator, such as ViaBTC's Profit Calculator, estimates theoretical daily mining yield based on selected hashrate, difficulty, and recent transaction-fee assumptions under the pool's PPS+ payout model. ViaBTC's documentation describes the calculator result as theoretical PPS+ mining yield based on the set difficulty and average transaction fees over the previous day (ViaBTC Help Center).

Electricity, hosting, and other operating costs should be modeled separately over the same period as the estimated revenue. In practice, this means an operator should first obtain an estimated daily or monthly revenue figure from the pool's calculator or dashboard, and separately calculate total facility electricity cost using the full IT load (or ASIC load under the assumption above), the site's measured or disclosed PUE, operating hours, and the applicable electricity rate. Combining the two—gross revenue minus total facility electricity cost minus any additional hosting fees not already reflected in the rate—produces a cost picture without double-counting cooling or infrastructure charges that may already be embedded in a hosting agreement.

Conclusion: Lower PUE Reduces the Electricity Needed to Support the Same ASIC Load

There is no single PUE value that applies uniformly to every mining farm; appropriate comparisons depend on climate, facility design, redundancy, and IT utilization, among other factors. What the metric reliably shows is that, for a fixed ASIC load, a lower PUE requires less total facility electricity, and therefore less billed electricity, than a higher PUE under the same rate and operating hours. Readers evaluating a mining site or hosting agreement should compare PUE figures only across equivalent facility boundaries and measurement periods, confirm how the figure interacts with the billing structure, and treat it as a complement to—not a substitute for—ASIC efficiency, pool-side hashrate, and mining-revenue estimates.

FAQ

Is a lower PUE always better for a mining farm?

A lower PUE means less non-IT facility energy is needed to support the same ASIC load, which reduces electricity cost when that overhead is billed to the operator. However, the specific PUE achievable at a site depends on climate, cooling design, and redundancy requirements, so comparisons are only meaningful between sites with similar measurement boundaries and operating conditions.

Does PUE measure cooling energy specifically?

No. PUE measures total facility energy divided by IT-equipment energy. PUE − 1 expresses non-IT facility energy relative to IT-equipment energy. That overhead includes external cooling, ventilation, pumps, power-distribution losses, and other facility loads. Cooling energy must be separately metered to isolate its contribution.

Can I calculate an ASIC's J/TH using PUE and pool-side hashrate?

No. ASIC J/TH is a hardware-efficiency metric based on measured power and hashrate at the device or fleet level. Pool-side hashrate is a share-based estimate and is not a substitute for a measured power reading, so combining it with PUE does not produce a valid J/TH figure.

How do I avoid double-counting cooling costs when a hosting agreement already includes an all-in rate?

Confirm with the hosting provider whether the quoted rate already reflects total facility energy, including cooling and infrastructure overhead. If it does, a separate PUE adjustment should not be applied on top of that rate; if cooling is billed as a distinct line item, it should be added once, not folded into both the base rate and a separate PUE multiplier.

References

• ISO. "ISO/IEC 30134-2:2026 — Information technology — Data centres — Key performance indicators — Part 2: Power usage effectiveness (PUE)." Source
• The Green Grid. "PUE: A Comprehensive Examination of the Metric," White Paper #49 (2012).
• Soluna Holdings. Form 10-K, fiscal year ended December 31, 2024. U.S. Securities and Exchange Commission. Source
• ViaBTC Help Center. "How are profits calculated?" Source
• ViaBTC Blog. "Pool Hashrate and Mining Statistics Explained." Source