Battery Arbitrage vs Self-Consumption: Loss Math

Last updated: November 2026

Battery self-consumption and battery arbitrage describe different uses of stored energy. Self-consumption keeps solar energy on site for later loads. Arbitrage shifts charging and discharge between differently priced periods. A system can combine them, but the same stored energy cannot earn both benefits independently. Compare the customer’s bill under complete operating scenarios rather than adding two headline savings figures.

This guide covers the economic comparison. Equipment selection, installation, grid charging and export permissions require the current manufacturer, utility and qualified project review. The battery sizing guide addresses the separate capacity and backup decision.

What changes between the strategies?

Self-consumption need not be passive: a controller may forecast loads or prices while deciding when to use stored solar. Arbitrage need not involve selling electricity. Discharging to a household load can avoid a more expensive import without any export transaction.

See the glossary entries on round-trip efficiency and battery storage for definitions. The DOE storage overview distinguishes energy capacity from power capacity and explains that storage conversion loses energy. A battery with enough kWh for a scenario may still lack the kW needed at the relevant time.

Compare values at the same energy boundary

Choose a consistent boundary. The following simplified equations use one AC kWh delivered from the battery. Assume η is the applicable AC-to-AC round-trip efficiency for the charging path, between zero and one. Use rates in the same currency per kWh, including the marginal charges or credits actually affected by the transaction.

For permitted grid charging followed by discharge to household load:

Grid energy needed = delivered energy / η
Grid-to-load operating margin per delivered kWh
  = avoided import rate at discharge
    − charging import rate / η
    − incremental operating or wear allowance

The charging rate must be divided by efficiency. Multiplying the difference between peak and off-peak prices by efficiency does not account correctly for buying more energy than the battery delivers.

For stored solar that would otherwise be exported:

Solar-to-load operating margin per delivered kWh
  = avoided import rate at discharge
    − forgone export value at charging time / η
    − incremental operating or wear allowance

That second equation assumes the input solar energy could have been exported and received the stated marginal value. If it would have been curtailed, its export opportunity cost may be zero. If credits have caps, carryover restrictions or limited bill use, a nominal export rate may not equal the household’s realized marginal value. For a DC-coupled system, use a model with the correct alternative solar conversion path rather than transferring this AC worksheet unchanged.

For permitted grid-to-grid dispatch, replace the avoided import rate with the eligible export value at discharge. Never use a peak retail import rate to value exported electricity unless the actual settlement rules support it.

Worked comparison: one hypothetical discharge event

Assume these illustrative inputs, not current rates or an equipment specification:

• 9 AC kWh delivered to household load during a higher-price period.
• 90% AC-to-AC round-trip efficiency, requiring 10 kWh of charging energy.
• Avoided import price of $0.35/kWh at discharge.
• Grid charging price of $0.10/kWh.
• Solar export opportunity value of $0.06/kWh at charging time.
• An assumed incremental wear allowance of $0.02 per delivered kWh.

Solar charging has the higher margin in this event if the 10 kWh of surplus solar is actually available and both scenarios serve the same 9 kWh of later load. This is not an annual saving, payback forecast or proof that one operating mode always wins.

On the delivered-energy basis, the hypothetical grid-to-load margin is about $0.2189/kWh; the solar-to-load margin is about $0.2633/kWh. Without the wear allowance, the grid-to-load operating break-even discharge price is $0.10 / 0.90 ≈ $0.1111/kWh. With the assumed allowance it is about $0.1311/kWh. There is no universal $0.15/kWh spread threshold: the charging price, efficiency, other marginal costs and actual discharge opportunity all matter.

If import and export rates are both $0.35/kWh, storing exportable solar at 90% efficiency to avoid an import at the same rate produces a negative operating margin before wear. Buying and delivering equal-value energy through a lossy battery is not a free saving. Backup or another service may still justify storage, but it belongs in a separate decision.

Why annual totals cannot settle a combined strategy

A household with annual solar generation below annual consumption can still export at midday and import at night. Conversely, a large annual surplus does not guarantee enough energy in every charging window. Evaluate time-aligned PV, load and tariff records.

A useful dispatch comparison tracks:

1. PV generation and customer load at the relevant interval and meter boundary.
2. Direct PV use before storage.
3. Available charging power, discharge power and usable energy.
4. Charging and discharge losses with consistent boundaries.
5. State of charge carried from one interval to the next.
6. Minimum reserve, equipment constraints and permitted operating modes.
7. Actual imports and eligible exports under each schedule.
8. The bill calculated under the applicable tariff and settlement rules.

Grid charging overnight can fill capacity that would otherwise accept solar the next day. Preserving reserve can reduce energy available for daily price shifting. A virtual power plant event can change both the household’s imports and its later backup availability. These are interacting schedules, not independent benefits to stack at their maxima.

Compare at least PV-only, solar-first storage and a permitted tariff-aware storage schedule using the same customer data. Include forecast errors and unavailable intervals where appropriate. A perfect-foresight model is a planning upper bound, not a controller’s guaranteed outcome.

Check tariff and controller scope before modeling

The current CPUC net metering and net billing overview distinguishes the California investor-owned utilities’ net billing framework from earlier NEM tariffs. Its comparison identifies export credits based on Avoided Cost Calculator values rather than assuming the retail import rate. That is a framework description, not one statewide fixed export price or a rule for every California utility.

Record the actual utility, service tariff, enrollment date, import schedule, export schedule, credit settlement and storage interconnection conditions. Confirm permitted grid charging and export separately. Do not infer permission from a battery app offering a setting.

Tesla’s current Savings mode documentation describes tariff-aware charging and discharge for Powerwall. This is vendor-reported functionality for its product, not comparative proof of the best controller or a savings result for another system. Check the supported model, region, firmware, tariff configuration, reserve behavior, connectivity and current agreement.

For any controller, request evidence of:

A monitoring device, tariff service and dispatch controller are different roles. Confirm what actually commands the battery; do not treat an energy-display product as a controller without evidence.

Separate operating margin from investment return

A positive dispatch margin does not show that buying the battery is worthwhile. Evaluate installed cost, recurring fees, maintenance, financing, degradation, expected replacement and the customer’s chosen analysis period. A cycle-life rating does not guarantee that every rated cycle will be available under the modeled temperature, calendar life and duty cycle.

For a project investment model:

NPV = − incremental installed cost
      + sum of annual net cash flows / (1 + discount rate)^year
      + discounted residual value, if justified

Annual net cash flows should come from the whole-bill comparison and defensible additional payments, less operating costs. Use one consistent treatment of capital and degradation. Do not subtract a full capital-recovery LCOS charge per kWh and also subtract the same capital cost again without explaining the accounting. A marginal wear allowance used to choose dispatch is different from a complete lifecycle cost estimate.

Show a sensitivity table for tariff changes, lower throughput, efficiency, reserve requirements, equipment downtime and installed cost. Keep resilience preferences separate from money unless the customer supplies a documented valuation method.

A practical decision tree

• Is bill management the objective? Obtain the current load, tariff and settlement records; if backup is the priority, define critical loads and reserve first.
• Is grid charging or export unverified? Keep that scenario out of the available option set until the relevant approval and product conditions are confirmed.
• Is there surplus solar at useful times? Test solar-to-load shifting against the export alternative, including losses.
• Are lower-price charging periods available? Test permitted grid-to-load shifting against the actual later load and rate, not just the day’s highest price.
• Do the two schedules compete? Run one combined time-series schedule rather than summing separately maximized results.
• Does the lifecycle case remain acceptable under sensitivity? Present its assumptions and limitations alongside PV-only and storage alternatives.

For related inputs, see the self-consumption rate guide, TOU scheduling guide and storage financial modeling guide. The generation and financial tool covers energy-yield and financial modeling. This article makes no claim about native battery dispatch optimization or live tariff integration, so ask for those functions specifically in a SurgePV demo.

Frequently Asked Questions

What is the difference between battery arbitrage and self-consumption?

Self-consumption uses solar energy on site, including surplus stored for later loads. Arbitrage shifts charging and discharge between differently priced periods. Discharge may avoid a household import or create a permitted export, and those have different values. A combined schedule must account for shared capacity, losses, load timing and operating constraints.

Is battery arbitrage worth it?

It is worth it only when the avoided import or eligible export value exceeds the charging rate divided by round-trip efficiency, plus wear, and the yearly margin repays installed cost. A positive operating margin alone says nothing about lifecycle payback. Model your actual tariff, efficiency, throughput and installed cost before you decide.

What price spread makes battery arbitrage worthwhile?

No universal spread applies. On a delivered-energy basis, compare the avoided import or eligible export value with the charging rate divided by round-trip efficiency, plus relevant marginal costs. A positive operating margin still leaves lifecycle payback open until installed cost, fees, degradation and actual throughput are included.

How do round-trip losses change battery arbitrage math?

Round-trip efficiency means you buy more energy than the battery delivers. At 90 percent, delivering 9 kWh needs 10 kWh of charging energy, so the charging price must be divided by 0.90 before you compare it with the discharge value. In the worked example, the break-even discharge price for a $0.10 charging rate is about $0.111 per kWh before wear.

Can a battery combine arbitrage and solar self-consumption?

It may, if the equipment, controller and applicable utility terms support the scenario. Model one feasible schedule, because grid charging can occupy capacity needed for later solar and backup reserves reduce available energy. Do not add the maximum value of separate strategies or count the same discharge twice.

Does arbitrage affect the battery warranty?

It can, so check the exact model and territory warranty, permitted charging sources, operating conditions and throughput limits. Cycle ratings alone do not prove warranty eligibility or usable lifetime. Do not transfer one manufacturer’s terms to another product or region.

How should California net billing be modeled?

Use the customer’s actual utility tariff, enrollment framework, import rates, export values and settlement rules. The CPUC distinguishes investor-owned utility net billing from earlier NEM tariffs, and that is not one statewide fixed price or a storage permission. Confirm grid charging and export conditions for the actual installation.