Stem and the Art of Selling Less

Consider the peculiar ambition of an energy company that wants to sell fewer batteries. A battery is a satisfying thing to sell: substantial, expensive, photographable. Software is less obliging. Try posing beside a subscription. Yet this is the choice Stem made when it reconsidered how to earn money from the energy transition. The question was whether the impressive invoice was buying it a good business.

The invoice that stopped being impressive

In 2023, Stem reported $461.5 million in revenue. In 2024, that figure was $144.6 million. A company losing almost seven-tenths of its top line has little reason to order celebratory stationery. Stem also reported an $854 million net loss in 2024, a figure affected by substantial impairment and other charges. These were company-level results, not a measurement of whether every customer's battery worked.

One revealing problem sat in the contracts. Legacy parent-company guarantees exposed Stem to reductions in hardware revenue. Its third-quarter 2024 filing described $38.7 million of revenue reductions over nine months and a $104.1 million receivables write-off in the third quarter associated with hardware subject to those guarantees. Stem had stopped putting these guarantees into new contracts in July 2023. The commercial arrangements could hurt even after equipment had been delivered.

On October 1, 2024, the company announced a new direction following a board-led strategy review. Lead with software and consultative services. Improve recurring revenue. Offer procurement advice, while limiting hardware procurement to deals meeting tighter profitability and working-capital rules. This was an attempt to change the quality of the business, with fewer large equipment transactions doing the heavy lifting.

A battery needs a timetable

Stem began in California in 2009, examining electricity use and the way businesses were charged for it. The original attraction was straightforward: stored energy could help a business manage electricity costs without asking everyone inside the building to change their routine. Forecasting and automated decisions made the battery useful. The metal cabinet alone could not read a tariff.

That distinction remains the center of the product. A solar asset owner needs to know whether a site is producing as expected. An operator needs alarms and controls. A developer needs to understand a project's economics before committing capital. A battery owner needs to decide when to charge, discharge or reserve capacity for another obligation. These people may share a project while spending their days looking at different screens.

PowerTrack Software supplies a common portfolio view, reporting and diagnostics. Its equipment compatibility is a selling point: Stem advertises more than 8,000 drivers and point maps. Buyers still need to verify the particular devices on their site. The practical promise is that assembling a portfolio from several manufacturers should not oblige the owner to assemble an equally sprawling collection of operating systems.

PowerTrack EMS coordinates solar and storage at the site. PPC, the power plant controller, handles operating modes and grid requirements. SCADA provides monitoring, alarms and local or remote control. Logger gathers equipment data. The names sound like attendees at a particularly dull conference; the division of labor matters. Portfolio reporting and fast electrical control are different jobs.

PowerTrack Optimizer, formerly Athena, connects forecasts, dispatch and financial optimization with managed services. Consider a simplified choice: discharge a battery now to reduce a bill, or retain energy for a potentially more valuable event later. An owner cannot spend the same stored electricity twice. Optimization helps weigh those competing uses against equipment and market constraints.

Stem also sells the expertise around the software: revenue modeling, integration, commissioning and round-the-clock remote operations. Its named customers include NineDot Energy, Regis Energy Partners and Madison Energy Investments. The latter's Ben Hunter describes the monitoring job neatly: PowerTrack serves as our eyes and ears throughout our portfolio to provide us with that clarity.

An owner cannot spend the same stored electricity twice.An elementary constraint with an unusually complicated timetable.

The expensive education

Stem went public through its Star Peak merger in 2021. It then acquired AlsoEnergy in February 2022, bringing solar monitoring and controls together with storage optimization. The deal announcement put the value at $695 million. The later filing recorded $652 million in total consideration after adjustments. Whichever document one reads, this was an expensive route to becoming a broader software company.

The operating reset had a human cost. In April 2025, Stem announced an approximately 27% workforce reduction; its annual report recorded $6.1 million in associated restructuring costs. The careers page now emphasizes resourcefulness, craft and safety. Those are the company's stated values. The layoffs are a reminder that a tidier strategy can arrive by a decidedly untidy process.

By the second quarter of 2026, the mix looked different. Revenue was $33.7 million, down 12% from a year earlier, while GAAP gross margin rose from 33% to 41%. Annual recurring revenue reached $62.4 million. Stem reported its fifth consecutive quarter of positive adjusted EBITDA, alongside a $14.4 million net loss. Improved operating measures and a continuing bottom-line loss can coexist quite comfortably on the same page.

The business model now combines SaaS subscriptions and service contracts with project work and edge devices. That leaves room for recurring relationships after installation. It also makes renewals, customer support and useful product improvements important. A software invoice may be smaller than a battery invoice, but it can return next year.

Old solar, new tricks

The clearest recent examples involve existing solar plants gaining batteries. In Chile, Copec Flux is deploying PowerTrack EMS at Granja Solar, a 135 MW solar project being retrofitted with 420 MWh of storage. Stem supplies the master controls and monitoring architecture; Copec Flux supplies local delivery and operating expertise. The purpose is to make electricity available at more valuable times.

In August 2026, Solarmarkt Group and EPC partner Pannonwatt selected Stem for two Hungarian retrofits. Each existing 60 MWp solar site is to gain a 40 MW / 80 MWh battery system. A five-year agreement expands the PowerTrack controls already supplied for the solar facilities. The announcement expected hybrid commercial operation by September. That timetable was a plan, rather than evidence of completed commissioning.

These projects explain the attraction of an integrated product. An owner upgrading a functioning plant would prefer to preserve useful infrastructure and familiar controls. The software must nevertheless accommodate a battery's behavior, grid constraints and market instructions. In Hungary, the initial commercial strategy targets automatic frequency restoration reserve, a service for helping balance the electricity system.

Stem is also adding AI interfaces and services. PowerTrack Sage is available within the monitoring application to help teams interpret site data and diagnose issues. AIONA, a dedicated AI services offering, was announced in June 2026. The useful test for a buyer is wonderfully prosaic: does it shorten the journey from a puzzling alarm to a defensible action?

What to steal from Stem

Stem occupies the space between the renewable equipment and the people responsible for its performance. It has company. Power Factors offers asset management and plant controls; Tesla's Autobidder provides battery trading and dispatch. The appropriate comparison depends on the task. Stem's pitch joins hardware-independent monitoring, controls and energy expertise, rather than relying on a single feature to settle every purchase.

For an asset owner, the lesson is to start with a concrete operating problem. Inventory the equipment. Check the available data and integration requirements. Test the tariff or market opportunity against realistic forecasts. Decide who responds to alarms and who has authority to dispatch. Ask for a scoped proposal: the reviewed product pages provide no standard public price list.

The economics will be less attractive where price differences are small, market participation is restricted or equipment integration is costly. Better forecasts cannot manufacture an interconnection approval or make a weak battery reliable. These are practical deductions from the work the products perform, not a promise of any particular customer's result.

For another company, the transferable idea is to examine what each sale demands after the revenue is booked: cash, guarantees, support and managerial attention. Stem's experience makes that accounting hard to ignore. Clean energy needs more equipment. The interesting question for Stem is how much of the accompanying responsibility customers will pay it to handle.