SolarEdge Technologies, Inc. SEDG

Key data

The business

SolarEdge makes the electronics that sit between rooftop solar panels and the wall socket. Its signature product is the power optimizer, one small unit bolted behind each panel, paired with a central inverter that flips the panels' direct current into the alternating current a home or business runs on. It sells batteries too, and software on top. The customers are solar installers and the distributors who supply them, historically more in Europe than anywhere else, across homes and commercial rooftops. The edge is the installed base and the installer who already knows the platform and reaches for it again. It is real but shallow: this is hardware in a market that lives and dies on power prices and subsidy, squeezed between Enphase on one side and cheap string inverters on the other. What actually happened here is a warehouse story. After the 2022 energy panic, European distributors bought far more inverters than households installed. When buying stopped, SolarEdge was left writing off product nobody wanted, a $1.8 billion loss in 2024 and revenue cut by two thirds from the peak.

The numbers

The recovery reads cleanly across five quarters.

Revenue in the June quarter was the most in over a year, up 20% on Q2 2025, and the loss is a fraction of what it was. The real signal is the gross margin: 11% a year ago, 28% now, four quarters of steady climb as written-down inventory clears and higher-margin product flows. Operating loss shrank from $115M to $16M, within sight of breakeven. Q4 2025 is derived and lumpy, carrying a year-end charge. On a non-GAAP basis the company has beaten estimates four quarters running, printing positive $0.06 in June against a small expected loss.

The 2022 per-share figure is a filing-data artifact, the real number was several times higher. The arc is the point: $3.1 billion of revenue in 2022, $901 million in the 2024 trough, back to $1.18 billion in 2025 and about a $1.3 billion pace this year. At $34 and a $2.08 billion market value the stock is 1.6 times sales, not obviously cheap and not dear. The whole case is a swing from a $6.88 loss in 2025 to a $1.83 profit analysts expect in 2028. What this memo believes and the tape does not fully price is that the margin recovery is genuine and cash is rebuilding, but a real slice of that 28% gross margin is US manufacturing tax credits I cannot size from this filing, and gross margin excluding those credits is the single print that says whether 2028 is real or borrowed.

Cash has risen every quarter off the 2024 bottom, and capex is almost nothing, so the balance sheet is healing on its own rather than on a raise.

Management

The record here is defensive and mostly right. Buybacks stopped after 2024 ($50M spent that year, zero since), capex was cut from $170M in 2023 to under $4M in the March quarter, and the company sat on cash while the channel drained. The one place money still goes is new US assembly lines, about $13M committed, aimed squarely at domestic-content manufacturing credits, which is also the lever quietly lifting gross margin. Insiders give no signal worth much: no open-market buys in the last year, two small director sales of about $99K in May, plan status not disclosed. Nobody is buying at $34, but nobody senior is dumping either.

How it fails or surprises you

The 2029 swing arrives early (right tail). Consensus already models $3.04 of earnings in 2029, and the market is paying for 2028, not 2029. If European residential demand and US commercial both firm while margin holds, the loss-to-profit turn re-rates the stock faster than the number moves. First tell: a June-quarter-style print with GAAP operating income crossing zero, possibly at the next report.

The margin is rented, not owned. The jump from 11% to 28% gross margin leans on US production credits booked against cost of goods. A rollback or faster phase-out under the current cuts to those programs would strip the recovery back to its underlying rate. The print that reveals it is gross margin excluding credits, which the company discloses. I cannot size the credit from this filing, which is itself the risk.

Two years is a long time to burn. The fact the bull read glides past: the company still lost money at the operating line last quarter, and there is no consensus profit until 2028. That is eight quarters of execution with a convertible note in the stack and cash that actually dipped year on year. If installs stall, the $527M cash and the convert become the whole story.

Closing thoughts

The next few prints settle most of this, so watch one pair. Hold gross margin near 28% while revenue keeps climbing and the operating line crosses into profit, and the recovery is real and pays for itself. Slip back toward 20% on flat revenue and the destocking bounce was the entire move, with the tax credits carrying what is left. An ambiguous print, margin steady but revenue flat, tells you the channel refilled and demand has not yet followed, and the right response is to wait for the following quarter rather than pay up. The left tail is a credit rollback plus stalled demand hitting together, and that is the one that matters because it takes out the margin and the runway at once. The right tail is worth more than the drop is deep only if you believe demand is genuinely returning.

The bet is still that people in Europe and America keep putting solar panels on their roofs, and keep buying SolarEdge's box that turns that sunlight into power the house can use, and the mountain of unsold inverters that piled up in European warehouses stays gone. It breaks if demand stalls before 2028 or the credits propping up today's margin get pulled. The pair to watch is gross margin against sequential revenue: if margin holds while revenue climbs, the recovery is funding itself, and if margin fades while revenue flattens, the 2028 profit everyone is paying for does not arrive.