Electronics Guide

Why Cost per Kilowatt-Hour Is the Wrong Metric

The levelized cost of energy (LCOE) expresses the lifetime cost of an energy system divided by the total energy it delivers, and it is the standard yardstick for grid-scale generation. Applied naively to milliwatt- and microwatt-class harvesters, LCOE produces alarming numbers, because the denominator, total energy delivered, is tiny while the harvester, power-management circuit, and storage still carry real fixed cost. A device that averages one hundred microwatts delivers roughly nine-tenths of a watt-hour per year, under ten watt-hours across a ten-year life. Against electricity priced in cents per kilowatt-hour, no plausible bill of materials can compete on that basis.

The metric is not wrong; it answers the wrong question for this domain. The decision a designer actually faces is not whether harvested energy is cheaper than the grid, but whether a harvester is cheaper than the alternative way of powering the same low-power device, typically a primary battery that must be replaced or a wired connection that must be installed and maintained. Reframing the economics around the cost displaced rather than the energy produced is the single most important shift in evaluating an energy harvesting opportunity.

Avoided Cost and Total Cost of Ownership

For a deployed sensor, the dominant lifetime expense is rarely the device itself; it is the recurring cost of keeping it powered. Replacing a coin cell in a sensor mounted on a rotating machine, sealed inside a wall, or installed on a remote tower can cost far more in technician time, travel, and production downtime than the battery or even the sensor is worth. In a hazardous or classified area, the same task may require a work permit, a second person, or a process shutdown, and the labor cost then dwarfs the hardware entirely.

Scale compounds the effect. A four-hundred-room hotel with four wireless sensors per room carries sixteen hundred cells; at a three-year replacement interval that is on the order of five hundred service actions every year, indefinitely. When a fleet numbers in the thousands, these visits dominate total cost of ownership, and they must be scheduled, tracked, and verified. An energy harvester that extends maintenance intervals from months to the life of the product can pay for itself many times over, even when its component cost exceeds that of the battery it replaces.

Battery lifetime claims deserve scrutiny in this analysis. Manufacturers commonly quote five to ten years for a primary lithium cell in a low-duty-cycle sensor, but that figure assumes a specific reporting interval, radio range, and temperature. Raising the transmission rate, lengthening the link, or operating at temperature extremes can cut the figure by a large factor. Lithium thionyl chloride cells offer very low self-discharge and a wide temperature range, yet they form a passivation layer during long storage that causes voltage delay on the first high-current pulse, a failure mode that surfaces in the field rather than on the bench.

A sound business case therefore models the full ownership cost of each alternative across the deployment lifetime: hardware, installation, energy, scheduled maintenance, unscheduled failures, and end-of-life disposal. Battery-based designs also carry costs that are easy to overlook, including inventory and logistics for spares, shipping restrictions and hazardous-material handling for lithium cells, and the warranty exposure of devices that fail when a battery is exhausted. Harvesting shifts spending from recurring operating expense toward up-front capital expense, a trade that favors long-lived, hard-to-service, or safety-critical installations, and one that may also need to clear an internal capital-approval threshold that routine maintenance spending avoids.

Metrics That Decision Makers Use

Different audiences evaluate the same design with different instruments, and a proposal is far more persuasive when it speaks in the buyer's units.

• Incremental bill-of-materials cost. The added cost of the transducer, power-management integrated circuit, storage element, and any enclosure change, minus the cost of the battery and holder removed. This is the number a hardware team controls directly.
• Payback period. The time for accumulated avoided cost to equal that incremental cost. Industrial buyers frequently look for payback inside one to three years, and a design that pays back only over a ten-year horizon faces a much harder sale.
• Total cost of ownership. The sum of all costs over the deployment life, discounted where the horizon is long. This is the metric that most often favors harvesting, because it captures the service visits that a purchase-price comparison hides.
• Return on investment and net present value. Finance functions compare a harvesting program against other uses of the same capital, so the case must survive a discount rate rather than merely showing a positive total.
• Non-recurring engineering. Design effort, tooling, qualification, and certification are one-time costs amortized across volume. They can decide viability on their own: a fixed cost that is trivial across a million units may exceed the entire margin on a run of two thousand.

Cost Reduction Through Volume and Learning

Like most electronics, harvesting components follow learning-curve dynamics, generalized as Wright's law: unit cost falls by a roughly constant percentage with each doubling of cumulative production volume, as designs mature, yields improve, and supply chains scale. Photovoltaics are the canonical example. The pattern known as Swanson's law describes module prices falling by roughly twenty percent for every doubling of cumulative shipments, a rate that has held over decades and produced a reduction of several orders of magnitude in module price.

Thermoelectric generators, piezoelectric and electromagnetic transducers, and the dedicated power-management integrated circuits that condition their output sit much earlier on this curve, at cumulative volumes far below those of photovoltaics. They therefore stand to benefit substantially as Internet-of-Things volumes grow, though the rate of improvement differs by technology and depends on whether the dominant cost is materials, precision assembly, or silicon area. Anticipating these reductions is central to roadmapping, because a system that is uneconomic at pilot volumes may become clearly viable at scale. The corresponding risk is planning a product around a cost point that arrives later than forecast, or not at all, if the volume that drives the learning never materializes.

Market Size and Growth

Published estimates of the energy harvesting market vary widely. Recent analyst forecasts place 2025 revenue anywhere from roughly six hundred million to more than four billion US dollars, with 2030 projections ranging from under one billion to several billion, and reported compound annual growth rates clustering in the high single digits to the low double digits. The spread reflects genuine differences in scope rather than disagreement about direction: analysts variously include or exclude large solar installations, wireless power transfer, and the supporting power-management semiconductors, and they draw the boundary of "energy harvesting" differently.

The consistent finding across sources is sustained growth, driven by demand for battery-free Internet-of-Things devices and the spread of ultra-low-power electronics. Because definitions differ by an order of magnitude, a specific forecast figure is close to meaningless without the market boundary the analyst used, and any figure quoted in a business plan should carry that boundary with it.

Adoption Curves and Application Segments

Energy harvesting does not advance uniformly; it succeeds first where the avoided cost is highest. Commercial building automation is among the most mature segments, where self-powered light switches and room sensors avoid both the conduit of a wired installation and the service calls of a battery-powered one. Industrial condition monitoring follows closely, because vibration and temperature sensors on rotating machinery are numerous, long-lived, and expensive to reach. Asset tracking, cold-chain logistics, utility metering, and electronic shelf labels form a further tier where fleet size makes even a modest per-visit cost significant in aggregate.

Consumer products adopt more slowly. There the harvester competes directly against an inexpensive battery that the owner replaces at no cost to the manufacturer, so the avoided-service argument largely disappears and the case must rest on convenience, form factor, or sustainability positioning. Passive radio-frequency identification and near-field communication occupy a distinct and long-established position. The tag draws all of its operating power from the reader's field rather than from an ambient source, which places these systems closer to wireless power transfer than to ambient harvesting, but their economics are instructive: adoption settled decades ago because the reader supplies the energy on demand and the tag itself costs almost nothing.

Mapping where a technology sits on its adoption curve, from early demonstration to mainstream deployment, helps distinguish markets that are ready now from those that depend on further cost reduction or efficiency gains. The distinction matters commercially, because a product aimed at a segment that is still two cost reductions away will consume capital long before it earns any.

Competitive Landscape

The industry spans transducer specialists, power-management semiconductor vendors, energy-storage suppliers, and system integrators who combine these into finished products. Competition turns less on raw conversion efficiency than on delivered value: integration effort, reliability over a product's life, and total installed cost. A transducer that is ten percent more efficient but requires a custom power-management design will usually lose to one that arrives with a qualified reference design and a supported software stack.

Standardized modules, reference designs, and radio ecosystems lower the barrier to adoption and shift competition toward ecosystem completeness. Interoperability profiles for self-powered building controls, and the broader low-power radio ecosystems around Bluetooth Low Energy and long-range networks, let an integrator treat harvesting as a component choice rather than a research project. The result is that the firms capturing the most value are often those that reduce a customer's design risk rather than those with the highest-performing single component.

Quantifying the Value Proposition

Translating technical capability into a purchasing decision requires expressing benefits in the language of the buyer. For an operations manager this means service visits avoided and uptime preserved; for a product manager it means new features or markets enabled by removing the battery; for a sustainability officer it means electronic waste and disposed batteries eliminated. For a facilities owner it may mean placing a sensor where no conduit can run and no ladder can reach.

The strongest cases pair a clear functional benefit, maintenance-free operation in a location where service is costly, with a defensible quantitative estimate of the cost it displaces over the deployment lifetime. Estimates drawn from the customer's own maintenance records carry far more weight than industry averages, and a pilot deployment that produces such records is often worth more as a sales instrument than as a technical validation.

Risk, Uncertainty, and Design Margin

Harvested power is variable, and variability has a price. A luminaire may be switched off for a week, a machine may idle over a holiday, and an outdoor panel may be shaded or soiled. Designing for the worst credible case rather than the average means a larger transducer and a larger storage element, both of which erode the cost advantage that justified harvesting in the first place. Much of the engineering work in this field is the search for the smallest margin that still meets the availability requirement.

That margin decision carries commercial consequences. A device that stops reporting during a low-energy period generates support calls and may breach a service-level agreement, so the cost of under-provisioning is not merely technical. Warranty terms must reflect what the environment can actually supply, which is why credible vendors specify minimum illuminance, temperature difference, or vibration amplitude as a condition of the performance claim. Storage elements also age: supercapacitors and rechargeable cells degrade with cycling and temperature, so harvesting reduces the maintenance burden rather than always eliminating it, and the business case should say which of the two it is claiming.

Policy, Regulation, and Supply Chain

Economic outcomes are also shaped by forces outside any single design. Environmental regulation is the clearest example. In the European Union, Regulation (EU) 2023/1542 on batteries and waste batteries requires that portable batteries in products placed on the market be readily removable and replaceable by the end user, with those provisions applying from February 18, 2027, and with spare batteries to remain available for at least five years after the last unit is placed on the market. Meeting that requirement constrains enclosure design, sealing, and ingress protection. A product that stores its energy in a capacitor and carries no battery at all falls outside those portable-battery obligations, which is a genuine, and often underweighted, commercial argument for harvesting.

Other compliance costs are fixed rather than variable, and therefore weigh most heavily on low-volume products. Restrictions on hazardous substances, waste electrical and electronic equipment obligations, and product-safety assessment all apply regardless of unit count. Any device with a radio must also clear spectrum certification, under Part 15 of the United States Federal Communications Commission rules or the European Union Radio Equipment Directive, and both add time to market as well as money.

Supply-chain factors determine whether projected cost reductions are actually realized. Thermoelectric modules depend on tellurium and bismuth, whose supply is concentrated and whose prices are volatile; magnetic harvesters depend on rare-earth magnets subject to the same concern; and the specialized power-management integrated circuits at the heart of most designs are produced by a small number of vendors, which creates single-source risk for a product expected to ship for a decade. Decisions about where to manufacture, and about qualifying a second source, belong in the business case rather than in a later engineering-change cycle. A realistic analysis treats these as integral variables rather than externalities.

Common Pitfalls in Harvesting Business Cases

Several errors recur often enough to be worth naming.

• Costing the transducer alone. The relevant comparison includes power management, storage, protection, and any enclosure change, not merely the solar cell or the piezoelectric element.
• Comparing against the wrong baseline. The alternative is often not a battery but a wired supply, or the status quo of not measuring at all, and each has a different economic profile.
• Using laboratory energy figures. Indoor illuminance is a small fraction of full sun, and a specification derived from bench conditions will overstate available energy by a wide margin.
• Omitting non-recurring costs. Qualification, certification, and tooling can exceed the entire savings on a small production run.
• Claiming maintenance-free operation too broadly. If the storage element has a finite cycle life, the honest claim is a longer service interval, not the elimination of service.