Alectrona
Manufacturers we specifySolarEdge takes the maximum power point tracking out of the inverter and puts a power optimiser on each panel, which recovers yield on shaded and multi-pitch roofs that a string design gives away. It is also the only maker here whose accounts are filed publicly, and those accounts show very large recent losses. We specify it on the architecture and we read the filings before we do.
The only architecture in the line-up that can save a badly shaded roof, from the only maker whose losses you can look up.
SolarEdge built its commercial range around one engineering decision taken in the mid 2000s: move the maximum power point tracking out of the inverter and put it on every module. Twenty years on, that is still the thing that separates it from every other maker on our list, and it is still the only reason to pay for it.
It is also a company in visible financial difficulty, and because it is publicly listed the difficulty is a matter of record rather than rumour. This page treats the architecture and the balance sheet as two separate questions, answers both, and does not let a good answer to one stand in for the other.
Every roof gets pitched module-level electronics by somebody. The honest answer is that the yield they recover is entirely a function of how much shade the array actually sees, across the day and across the year, and that is a measurement rather than an opinion. So the SolarEdge decision is made from the survey model, not from a preference for the architecture.
Our in-house drone survey produces a three-dimensional model of the roof and everything standing on it: plant, flues, parapets, lift overruns, the neighbouring block that only matters in December. That model drives a half-hourly PV*SOL simulation of the array with and without per-module tracking, so the comparison is a number rather than an argument. On a roof where a plant deck throws a moving shadow across three rows for four hours a day, the optimisers usually win comfortably. On a clean single pitch they usually do not, and we say so and specify a plain three-phase string inverter instead.
The same model settles the other half of the question, which is device count. Optimisers are an active component on every panel, at height, for twenty-five years. Where the shading case is marginal, that hardware is a maintenance liability rather than a yield gain, and it is better to know that from the model than from the second year of an operations contract.
In a conventional string design, the modules on one tracker are electrically bound together. The weakest panel in the string sets a ceiling for the rest of it, so a soiled panel, a cracked cell or a shadow from a flue drags on every module wired behind it. Grouping strings carefully by orientation limits the damage, and a high tracker count limits it further, but the coupling is still there.
A power optimiser breaks it. Each module gets its own device performing its own maximum power point tracking and reporting its own output, so a panel that is underperforming loses its own production and nobody else's. Two operational consequences follow, and they matter more on a commercial roof than the yield arithmetic does. The first is diagnosis: monitoring tells you which module is down rather than which string is low, which turns an afternoon of testing into a single visit to a known position. The second is safety. Because each optimiser can be commanded off, the array can be brought down to a safe voltage at module level, which is exactly the DC de-energisation that RC62 fire-safety guidance looks for and a real advantage for maintenance teams and for anyone attending an incident.
The cost of all that is hardware and complexity in the place hardest to reach. Whether the trade is worth making is set out in our comparison of string against central architectures, and it is decided on your roof rather than in general.
A merged SolarEdge site is covered by three separate warranties that expire in three different decades. That is worth planning for at the point of purchase rather than discovering in year eleven.
SolarEdge does not scale its commercial inverters by making one machine bigger. It changes architecture, and the change lands right in the middle of the band a roof above 50 kWp cares about.
A single three-phase inverter with the DC Safety Unit built in, giving AC and DC segregation, MC4 connectors and a DC surge protection device in one enclosure. Peak efficiency sits around 98.3%.
The usual answer up to about 33 kW per inverter, and often the right one above it as well.
The Synergy platform builds roughly 90 to 100 kW from multiple Synergy units reporting to a single Synergy Manager, so the largest ratings are a coordinated set rather than one large machine.
More parts in the plant room, and a different commissioning and spares picture from a single unit.
Which side of that line a scheme sits on comes out of the PV*SOL model and the connection, and it changes the cable, switchgear and plant-room design as much as it changes the inverter.
Thearchitectureissound.Thebalancesheetneedsreading.Thosearetwoquestions.
Weanswerbothbeforewespecify,andwewritethesecondonedown.
Ifasettledmanufacturersuitsyourroofjustaswell,thatiswhatwewillputforward.
Each module carries its own power optimiser, so tracking, monitoring and shutdown all happen at panel level and a shaded or soiled module drops only itself. Above about 33 kW the platform stops being one machine and becomes several Synergy units reporting to one manager.
Every module gains an active component. A 500 kWp roof therefore carries several hundred more devices than the equivalent string design, each one something that can fail and something that has to be reached at height.
The battery inverter works from the AC busbar, so a store can be added to an array that already works without touching the solar inverters. This is also the clearest illustration on the site of how a commercial battery is actually bought: the same 50 kW rating is supplied against one cabinet or two, so the power stays where it is and the duration doubles. Power covers the height of the peak, duration covers its length, and they are ordered separately.
The storage cover runs ten years while the inverter path runs to twenty, so on a merged SolarEdge site the battery and the inverter do not reach the end of their warranted lives together. The cabinet variant designation also has to be confirmed against the current order guide before an order is placed, because the family carries more than one two-letter variant code and they differ by a single character.
Datasheet figures only. We publish no price, and the exact model is confirmed against the design and the current datasheet before order.
SolarEdge is listed on Nasdaq, which makes it the unusual case in this line-up: you do not have to take a bankability view on trust, because the filings are public and you can read them yourself. That transparency is a genuine advantage over a privately held maker whose scale figures come from its own marketing. It is also how everybody knows how bad the last two years were.
The figures are serious. SolarEdge reported a GAAP net loss of about US$1.81bn for full-year 2024, driven by roughly US$1.17bn of write-downs and impairments, with revenue falling from about US$2.98bn in 2023 to about US$901m in 2024, and it posted a further loss in 2025. So the question a finance committee should ask is not whether the technology works, because it does, but whether the entity behind a 12-year inverter warranty and a 10-year storage warranty will be in a position to honour them. We confirm in writing how the warranty is backed and how UK support and spares are arranged before we specify it, we hold the monitoring access and the as-built records ourselves so the site is not dependent on a manufacturer call centre, and where a more settled maker suits the roof equally well we recommend that instead. On the storage side the certification is confirmed for the exact product against IEC 62619 at cell level and the IEC 62933 series that frames the storage system and its grid integration, because a battery has no financeability scheme of its own and the standards carry the weight.
SolarEdge is specified on the survey, never on preference. The architecture is worth real money on some roofs and is simply more hardware on others, and the model decides which of those a building is.
Not automatically, but it changes the diligence. SolarEdge reported a GAAP net loss of about US$1.81bn for 2024 on roughly US$1.17bn of write-downs and impairments, with revenue falling from about US$2.98bn in 2023 to about US$901m, and it posted a further loss in 2025. Before we specify it we confirm in writing how the warranty is backed and how UK support and spares are arranged, and we hold the monitoring access and the as-built records ourselves so the site is not dependent on the manufacturer for day-to-day operation. Where a more financially settled maker suits your roof just as well, we will tell you plainly and put that forward instead.
When the roof genuinely loses energy to shade or mismatch, and the model can show it. Plant decks, flues, parapets, lift overruns, several pitches under one connection: those are the cases where per-module tracking recovers what a string design gives away, and where per-module monitoring shortens every future fault visit. On a clean, unshaded single pitch the yield gain narrows sharply while the hardware count does not, so a conventional three-phase string inverter is usually the better-value answer. We model both from the drone survey rather than deciding by brand.
Because they are different products with different degradation. The commercial three-phase inverters carry a 12-year manufacturer warranty with a purchased path to 20 years, the power optimisers carry a 25-year term, and the Commercial Storage System carries 10 years or 6,000 cycles to 70% state of health, whichever comes first. So a merged SolarEdge site has three clocks running at once. That is worth planning around at purchase: the battery is the part most likely to need a commercial decision partway through the array's life, and we set out that expectation at design rather than leaving it to be found.
No. The 50 kW PCS050 battery inverter is AC-coupled, so it sits on the AC side of the supply and the store can be retrofitted to an existing array without changing the solar inverters. Where the array already runs SolarEdge optimisers the integration argument is genuinely strong, because the generation and the storage report into one platform and the charge logic sees panel-level data. Where it does not, the AC coupling still works cleanly, and the choice is then decided on the merits against the other makers rather than on platform loyalty.
The related question is how much store sits behind that 50 kW, because the answer is not fixed. The two-hour build gives about 102 kWh, which holds roughly two hours at full output; the four-hour build gives roughly 204 kWh at the same 50 kW. A site whose peak is short and sharp wants the first, and a site holding a load down across a whole afternoon rate wants the second. We settle that from the half-hourly data on our battery sizing page rather than from the power rating alone.
Not in China, which is unusual in this market and occasionally decides a procurement question on its own. SolarEdge builds through contract manufacturing partners rather than its own plants, and that footprint has been consolidated onto Vietnam and the United States since 2024, with other sites closed. The head office and the engineering are in Israel. Because the arrangement has changed repeatedly, we confirm the build origin for the exact model with the distributor at the point of specification rather than quoting it from a datasheet that may predate the last move.
The equipment is rarely the pacing item. The Commercial Storage System, the battery inverter and the three-phase units are ordered to the surveyed configuration with stock confirmed in writing before contract. Running alongside that, the G99 submission to your distribution network operator usually sets the programme, because its assessment timescale and any connection conditions sit outside our control. We start both clocks together and give a project-specific schedule once the survey and the connection position are known rather than a generic lead time.
Last updated August 2026
Tell us about the roof and the load. We survey the site, model the array in PV*SOL against your half-hourly data, and specify what actually fits, with the manufacturer, the series and the warranty named. SolarEdge goes on the roof where it is the right answer and not because it is on a list.
