SolarEdge Technologies
It makes the small box bolted behind each rooftop solar panel, the inverter that turns the panels' power into the kind a building uses, and increasingly batteries.
Our viewUnattractive Medium conviction
The scores
The profile · 1-4
Growth3/4 ↑
Revenue grew 31.4% in FY2025 and 19.6% in Q2 2026, but off a base that fell 70%, narrowing to batteries while inverter units fell 28.5%.
Business quality2/4 →
The moat is rented from Congress and dated. The last downturn took 70% of revenue, and the share claim is management's alone.
Financial position3/4 ↑
Net cash of $264.6M, one maturity in July 2029, no financing need. Not a 4 because operations do not yet fund it.
Execution record2/4 ↑
Revenue fell 70% in a year, four businesses were sold or discontinued, Q3 2024 was restated, and the turnaround quarter was guided straight back down.
Per share capture2/4 ↓
Shares rose 1.9% in six months, $300M of notes convert at $34.32, and stock compensation is most of expected 2028 profit.
Potential2/4 ↑
Several independent things must go right and one is outside its control: NVIDIA's own documents say the solid-state architecture is still being studied.
Risk profile2/4 →
Treasury's final foreign-entity rules are one externally controlled ruling that decides credit eligibility, and the credit is the whole gross margin.
Valuation2/4 ↑
At 1.6 times sales the price needs the best operating margin of five years restored on less revenue, plus a data-centre business with no customer.
Evidence quality3/4 →
Central claims come from filings, a regulator, NVIDIA and a rival's 10-K. But the credit's size is inferred, not disclosed.
The bet · 2-3 years
Safe, little upside
No upside the evidence defends · if it fails you keep about 100%.
The Verdict
We rate SolarEdge unattractive. It makes the electronics behind rooftop solar panels, and sales are growing again after a collapse, but the business is mixed: without a US manufacturing tax credit it would have made a gross loss, and it never says how big that credit is. Even with it, operations still lose money. The price is full: it values a recovery without knowing how much of the profit is the taxpayer's. As a bet on its new AI data-centre product it is defensive: the solar business carries today's value, but with no customer yet we see little to win.
The long caseAI server racks are moving to higher-voltage power, which may need a new kind of transformer; SolarEdge would have to win customers and a standard it does not yet have.
The evidence under this - 7 claims
C03facthigh confidence
SolarEdge states in both its FY2025 10-K and its Q2 2026 10-Q that excluding the advanced manufacturing production tax credits it recognises under Section 45X, it would have reported a gross loss rather than a gross profit - and in the 10-Q, for both Q2 2026 and Q2 2025.
C05facthigh confidence
SolarEdge does not quantify the 45X credit anywhere in its FY2025 10-K or its Q2 2026 10-Q. The figure that decides whether the company has a gross profit at all appears as a sentence and never as a number.
C08facthigh confidence
In Q2 2026 SolarEdge reported revenue of $346.2M, gross profit of $95.2M or 27.5% of revenue, an operating loss of $16.0M, a net loss of $30.8M and diluted loss per share of $0.50.
C19facthigh confidence
At $36.75 SolarEdge's market capitalisation is $2.24B and its enterprise value $2.10B, 1.58 times trailing twelve-month revenue of $1.33B - the sum of the four quarters the company itself filed - against a prior five-year median enterprise value to sales of 3.46 times.
C20inferencemedium confidence
At this enterprise value, analysts' consensus adjusted earnings of $1.83 a share for 2028 require an operating margin SolarEdge has reached once in five years - FY2021's 10.5% - and it reached it on revenue a fifth higher than the 2028 consensus.
derivedConsensus 2028 earnings of $1.82783 on twelve analysts, against consensus 2028 revenue of $1.63B, both from the pack's analyst estimates. On about 62 million shares that is roughly $113M of net income, or a 6.9% net margin, which needs an operating margin above it once tax and the notes' 2.25% coupon are paid. The pack's annual series gives operating margins of 10.5% in FY2021, 5.3% in FY2022, 1.4% in FY2023, minus 189.5% in FY2024 and minus 25.5% in FY2025; FY2021 revenue was $1.96B against the $1.63B consensus for 2028.
against this
- The consensus figure is adjusted, not filed, and the comparison to FY2021's GAAP operating margin is therefore not like-for-like - an adjusted 2028 margin could clear the bar while the GAAP one does not, which is exactly the pattern every recent quarter shows.
- FY2021's margin was earned without any 45X credit, since SolarEdge only began United States manufacturing in the fourth quarter of 2023. So the credit is a reason the margin could be reached on lower revenue this time, and the honest statement is that the price needs the best operating margin of the last five years restored on a fifth less revenue, with a subsidy that was not there the first time and starts phasing out in 2030.
C22management claimhigh confidence
SolarEdge's 800-volt direct-current data-centre products are under development and validation and are not yet generally available; the three-phase medium-voltage to 800 VDC conversion stage is operating under load in the company's engineering laboratories.
against this
- The status is disclosed by the company itself in the same release that announces the progress, so it is a candid statement rather than one extracted from it.
- The engineering has moved: on the Q2 2026 call, five weeks earlier, management described demonstrating 'a working prototype' going from medium voltage to 800-volt regulated direct current, and said the proof-of-concept would reach full three-phase 34.5 kilovolt operation by the end of 2026. The Investor Day claim is a step along that stated path rather than a leap beyond it.
C23management claimhigh confidence
Management expects data-centre pilots in 2027 and meaningful revenue in 2028, and said at the 2026 Investor Day that it is in discussions with hyperscalers and neoclouds but holds no purchase order or agreement with a specific customer.
sourceSolarEdge Q2 2026 earnings call transcriptprimarySolarEdge Investor Day release: 'SolarEdge Advances Full DC…secondary
against this
- The 2027-pilot, 2028-revenue sequence tracks NVIDIA's own published schedule rather than running ahead of it, so the timeline is not obviously promotional.
- It is nonetheless a timeline management has already moved once inside two months: on the Q2 2026 call an executive said 'the next few months until the end of the year' for the full three-phase proof of concept, then corrected '2026 will be pilots' to '2027 will be pilots' mid-sentence. And no part of the sequence is contractual - the whole case rests on winning a customer that does not exist in any filing.
The company
A solar panel's output drops when it is shaded, dirty or old, and on a plain system one weak panel drags down its whole row. SolarEdge bolts a small converter, an optimizer, behind each panel so every panel works at its own best.
The pooled power then goes to one central inverter that turns it into the alternating current a building uses. Enphase, its main rival, names SolarEdge first among its competitors.
The evidence under this - 2 claims
C01facthigh confidence
SolarEdge sells a DC optimized inverter system: a power optimizer on each solar panel that holds that panel at its own best operating point, feeding one central inverter that converts the pooled direct current to grid alternating current.
C27facthigh confidence
Enphase Energy's FY2025 10-K names SolarEdge first in its list of inverter-market competitors and again in its storage-market competitor list, and states that its microinverters offer advantages relative to string inverter technology 'even when supplemented by DC-to-DC optimizers on the roof'.
How it makes money
It sells hardware: optimizers, inverters and, increasingly, batteries, which supplied $115.0M of the first half's $147.8M rise in sales while inverter units fell 28.5%. Most of it goes through distributors, and one, Consolidated Electrical Distributors, took 18.6% of 2025 revenue. Every dollar of gross profit, the company says, depends on a US credit for making the parts in America.
The evidence under this - 3 claims
C13facthigh confidence
In the six months ended June 30, 2026 SolarEdge's inverter units recognised as revenue fell 28.5% to 113.1 thousand while battery megawatt hours rose 96.1% to 757.0 and optimizer units rose 3.6% to 4.9 million. Of the $147.8M revenue increase, $115.0M was batteries.
C02facthigh confidence
In the year ended December 31, 2025 one customer, Consolidated Electrical Distributors, Inc, represented 18.6% of SolarEdge's revenues, and the top three customers together represented 35.8%. Three customers were 45.5% of trade receivables, net.
C03facthigh confidence
SolarEdge states in both its FY2025 10-K and its Q2 2026 10-Q that excluding the advanced manufacturing production tax credits it recognises under Section 45X, it would have reported a gross loss rather than a gross profit - and in the 10-Q, for both Q2 2026 and Q2 2025.
Where the money comes from, stream by stream
| Stream | Share of sales (Q2 2026) | How it earns it |
|---|---|---|
| Batteries | 36.4% | Home and commercial battery storage sold with its inverters, mostly to European and US installers through distributors. |
| Power optimizers | 33.7% | The converter fitted to each solar panel, the part that defines the system and earns the 45X microinverter rate. |
| Inverters | 20.6% | The central unit turning the panels' pooled direct current into grid power, residential and commercial. |
| Battery accessories and other | 9.4% | Backup interfaces, chargers, communication and software, and other items sold around the system. |
Who runs it
Shuki Nir, who once turned round SanDisk's consumer business, has run SolarEdge since December 2024; Maoz Sigron became its third finance chief in two years in May 2026. Insiders own about 1% of the company. The yearly bonus pays mostly on sales, profit, cash and new products; half the share awards pay only if the share price climbs past set levels, and the last two such awards paid nothing or are on course to.
The evidence under this - 3 claims
C53facthigh confidence
Shuki Nir has been SolarEdge's chief executive since December 4, 2024, after six months as chief marketing officer and earlier as general manager of SanDisk's consumer business; he succeeded Zvi Lando and an interim chief. Maoz Sigron, Perion's former CFO, became chief financial officer on May 31, 2026, the third since June 2024 after Ariel Porat and Asaf Alperovitz. Uri Bechor has been chief operating officer since 2019.
C51facthigh confidence
As of April 6, 2026 SolarEdge's ten directors and executive officers owned 633,605 shares, 1.04% of 60,817,929 outstanding; the largest insider, chair Avery More, held 297,945 and the CEO 103,267. BlackRock, at 11.3% on a filing reporting April 2024, was the only holder listed above 5%. There is one class of common stock.
C52facthigh confidence
SolarEdge's 2025 bonus paid nothing below $1B of revenue and then weighed revenue 30%, non-GAAP gross margin 10%, non-GAAP operating income 10%, free cash flow 10%, product quality 10%, new product introduction 20% and manufacturing KPIs 10%; it paid 86.2% of target. Half of each annual equity award is PSUs earned at share-price hurdles of $40, $70 and $100 within three years, half RSUs vesting over four. The 2023 PSUs paid nothing and the 2024 PSUs were tracking to nothing.
against this
- The non-GAAP gross margin goal (16.7% achieved against a 16% target) is a margin the 45X credit makes positive (C03), so a tenth of the bonus rides on the subsidy rather than on the product.
- New product introduction carried 20% and paid 25 points, the most of any measure, on goals the proxy withholds as commercially sensitive.
Why it wins, and who could beat it
Why customers come back
The advantage is a rule more than a product. SolarEdge builds in Utah, Florida and Texas, and US tax rules treat its panel-by-panel system like a microinverter: 11 cents of credit per watt, against 2 cents for an ordinary commercial inverter.
Rules on foreign ownership shut out rivals too. Management says it reached 50% of US commercial rooftop installations last quarter because the other two leaders, SMA and Chint, fail them. But the credit phases out by the end of 2031.
The evidence under this - 4 claims
C06facthigh confidence
Under 26 CFR 1.45X-3 a DC optimized inverter system qualifies as a microinverter and earns 11 cents per alternating-current watt, against 2 cents for a commercial inverter, 1.5 cents for a utility inverter and 0.25 cents for a central inverter.
C28management claimlow confidence
Management said on the Q2 2026 call that SolarEdge reached 50% of United States commercial and industrial rooftop installations in the quarter, and attributed it to its technology and to foreign-entity and domestic-content compliance that the other two leaders, SMA and Chint, do not meet.
against this
- The figure is management's own, unsourced, stated in answer to an analyst question rather than in a filing, and appears in neither the results release nor the 10-Q. No independent measure of United States commercial and industrial inverter share was found; no trade association or regulator publishes one that was reachable.
- The characterisation of SMA and Chint is qualified by management itself - 'to the best of our knowledge' - and its consequence is qualified twice more in the same answer: it is 'not yet clear' when the FCC ruling will affect competitors' new products, or whether their existing products can continue to be imported at all.
- It is also a share of a shrinking pool for SolarEdge: revenue from outside the United States rose to 55.3% of the total in the same quarter, so United States revenue fell year on year while this share was allegedly being won.
C29management claimlow confidence
Management said on the Q2 2026 call that SolarEdge complies with the FCC Covered List and sees no reason it will not continue to, noting that it is a Delaware company with the majority of its manufacturing in Utah, Florida and Texas.
against this
- The primary text could not be read: the FCC's public notice PDF did not render and the Covered List page returns HTTP 403, so the ruling's actual scope is carried here only in the words of a company that benefits from it.
- Management's own framing of the benefit is heavily hedged in the same answer: it is 'not yet clear' when the ruling will affect companies whose new products are not approved, and 'not yet clear whether their existing products can actually continue to be imported into the U.S. or not and when, if at all'. A restriction on new device models leaves an incumbent's approved catalogue selling.
C07management claimhigh confidence
SolarEdge expects the advanced manufacturing production credits to be phased out by the end of 2031, and states that reductions without an offsetting fall in manufacturing cost would adversely affect gross margin and 'may include transitioning into a gross loss'.
against this
H.R.1, enacted 2025-07-04, did not shorten the term of the 45X credits - the company says so itself - so the phase-out date is the one already in statute rather than a new political risk. The nearer risk runs the other way: the foreign-entity rules attached to 45X began applying from January 1, 2026 and the final rule had not been issued when the 10-K was filed, so eligibility can be lost years before the phase-out arrives.
The threat
In solar, Enphase puts a complete converter under every panel and argues in its own annual report that this beats optimizers. Cheaper string inverters skip panel electronics altogether.
In data centres, NVIDIA names Eaton, Schneider Electric and Vertiv as its power partners and SolarEdge nowhere, and it is still weighing traditional transformers against the solid-state kind SolarEdge is building.
The evidence under this - 4 claims
C27facthigh confidence
Enphase Energy's FY2025 10-K names SolarEdge first in its list of inverter-market competitors and again in its storage-market competitor list, and states that its microinverters offer advantages relative to string inverter technology 'even when supplemented by DC-to-DC optimizers on the roof'.
C01facthigh confidence
SolarEdge sells a DC optimized inverter system: a power optimizer on each solar panel that holds that panel at its own best operating point, feeding one central inverter that converts the pooled direct current to grid alternating current.
C24facthigh confidence
NVIDIA's published 800 VDC partner ecosystem names Eaton, Schneider Electric and Vertiv as data-centre power system partners, ten silicon providers and five power-system component makers. SolarEdge appears in none of the categories, on either of NVIDIA's two public pages.
sourceNVIDIA 800 V HVDC architecture technical blogprimaryNVIDIA 800 VDC architecture product pageprimary
against this
- Absence from a partner list is not exclusion from a market. SolarEdge and NVIDIA jointly published a white paper, '800 VDC Protection and Grounding for AI Data Centers', on 2026-09-10 - so there is a working technical relationship that the partner pages, which may not have been refreshed, do not reflect.
- The partner categories are drawn around the parts NVIDIA specifies - silicon, power shelves, facility power systems - and a medium-voltage conversion stage may simply not map to one of them. Against that, 'data centre power systems' is exactly where such a product would sit, and the three names in it are the incumbents SolarEdge would have to displace.
C25facthigh confidence
NVIDIA states that it and its partners are actively studying the capital cost, operating cost and safety implications of both traditional transformer-based and solid-state transformer approaches - so the architecture SolarEdge is building has not been chosen by the party setting the standard.
against this
- SolarEdge is not alone in betting on it: Infineon is collaborating with SolarEdge on solid-state protection for 800 VDC, and Infineon is itself on NVIDIA's named silicon list - so a named NVIDIA partner is putting engineering behind the same architecture.
- The joint white paper SolarEdge and NVIDIA published is described as a technology-neutral framework for protection and grounding, which is consistent with the architecture being open rather than with either side having settled it. Neutrality cuts both ways: it is evidence of engagement and evidence that nothing has been decided.
The rivals: ENPH, ETN, VRT, TSLA, and 1 more that is private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| ENPH | Enphase Energy microinvertersPriced against it: Same product | Solar and storage | On sale | Same product |
| Sungrow Power Supply and Huawei string invertersPriced against it: Listed abroad | Solar and storage | On sale | Listed abroad | |
| ETN · VRT | Conventional transformer and rectifier chains from Eaton, Schneider Electric and VertivPriced against it: One segment only | Data-centre 800 VDC solid-state transformer | On sale | One segment only |
| TSLA | TeslaPriced against it: Small part of it | - | On sale | Small part of it |
How big it could get
How big this could get
In solar, America had 59,506 MW of small rooftop-scale capacity at the end of 2025. The growth case is data centres: NVIDIA says racks will move to 1 megawatt from 2027, beyond what today's in-rack wiring can carry.
Management targets $2.4B of sales by 2029; analysts expect $1.84B. Nobody has sized what SolarEdge could sell into those racks.
The evidence under this - 3 claims
C35facthigh confidence
The United States had 59,506 MW of small-scale solar photovoltaic generation capacity at the end of 2025, generating about 93 billion kilowatt hours.
C26facthigh confidence
NVIDIA states it is leading the transition to 800 VDC data-centre power to support 1 megawatt racks starting in 2027, with full-scale production coinciding with its Kyber rack-scale systems in 2027, and that traditional 54-volt in-rack distribution begins to hit physical limits as racks exceed 200 kilowatts.
C21management claimmedium confidence
SolarEdge told its 2026 Investor Day it targets $2.4B of revenue by 2029. Analysts' consensus 2029 revenue is $1.84B - $563M, or 23%, below the target.
sourceSolarEdge Investor Day release: 'SolarEdge Advances Full DC…secondarySolarEdge research pack, frozen at…primary
against this
- The consensus of nine analysts for 2029 was set before or around the Investor Day and may not yet reflect the target, so the gap partly measures timing rather than disagreement.
- Management has a poor record on multi-year targets at this company: revenue fell from $2.98B in FY2023 to $901.5M in FY2024, and the businesses that were to carry earlier growth - energy storage cells, e-Mobility, PV Trackers and Automation Machines - have since been discontinued or sold.
The results
The numbers
Is it growing?Fast
Sales grew 46% over the twelve months to Jun 2026, faster than the 31% in the year to Dec 2025.
Is it profitable?Losing
Operating margin is −12%, down from 1.4% in the year to Dec 2023.
Does growth pay off?Pays off
Running costs take 35¢ of each sales dollar, up from 21¢ in the year to Dec 2021.
Does it earn good returns?Poor
Shareholders' money lost 58% a year, down from 14% in the year to Dec 2021.
How steady is it?Swingy
Yearly sales growth has ranged from −70% to +58% over the last 5 years, with sales falling in 2 of them.
Are shares being diluted?Creeping
The share count rose 9.1% in 4.5 years, and pay in shares costs 6.1% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-10-04. Fiscal years end in December.
| Dec 2021 | Dec 2022 | Dec 2023 | Dec 2024 | Dec 2025 | 12 mo to Jun 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $1.96B | $3.11B | $2.98B | $901M | $1.18B | $1.33B |
| Sales growth | +35% | +58% | −4.3% | −70% | +31% | +46% |
| Profit | ||||||
| Gross margin | 32% | 27% | 24% | −97% | 17% | 23% |
| Operating margin | 11% | 5.3% | 1.4% | −190% | −25% | −12% |
| Net margin | 8.6% | 14% | 1.2% | −200% | −34% | −20% |
| What growth costs, per sales dollar | ||||||
| Research and development | 11% | 9.3% | 11% | 31% | 19% | 16% |
| Sales and marketing | 6.1% | 5.1% | 5.5% | 16% | 9.9% | 8.4% |
| General and admin | 4.2% | 3.6% | 4.9% | 16% | 8.5% | 8.4% |
| Pay in shares | 5.2% | 4.7% | 5.0% | 15% | 7.8% | 6.1% |
| Capital spending | 7.6% | 5.4% | 5.7% | 12% | 2.0% | - |
| Cash | ||||||
| Cash from operations | $214M | $31.3M | −$180M | −$313M | $104M | - |
| Free cash flow | $64.9M | −$138M | −$351M | −$421M | $80.8M | - |
| Balance sheet | ||||||
| Cash and short-term investments | $698M | $1.02B | $860M | $586M | $493M | $547M |
| Returns and shares | ||||||
| Return on capital | 16% | 5.0% | - | - | - | - |
| Shares (diluted) | 56.0M | 58.1M | 57.2M | 57.1M | 59.0M | 61.0M |
What growth costs
Running costs barely move with sales, and they have been cut hard since 2023, so each extra sale now lifts the margin. But those costs still take more of every sales dollar than the company keeps after making its products, and what it keeps rests on a tax credit.
The evidence under this - 3 claims
C49facthigh confidence
Per sales dollar, research and development ran 10.8 cents in FY2023, 30.8 in FY2024, 18.7 in FY2025 and 15.7 in the four quarters to June 2026; sales and marketing 5.5, 16.3, 9.9 and 8.4; general and administrative 4.9, 16.4, 8.5 and 8.4; stock pay 5.0, 15.2, 7.8 and 6.1. The three operating lines cost $632.3M in FY2023, $571.6M in FY2024, $439.6M in FY2025 and $432.5M over the last four quarters.
C50inferencemedium confidence
SolarEdge's operating costs are mostly fixed in dollars, so growth widens the margin: the three operating lines fell from 63.4 cents per sales dollar in FY2024 to 32.5 over the last four quarters as revenue recovered and costs were cut. But 32.5 cents still exceeds the 23.3 cents of gross profit, which itself rests on the 45X credit.
derivedC49's cost lines and revenue, the trailing gross profit of $310.0M on $1,332.3M of revenue in the numbers card, and C03 on the credit
against this
Most of the fall since FY2024 is cost-cutting rather than growth: the three lines are $139M lower in dollars, and revenue over the last four quarters is still 55% below FY2023's, so the leverage has not yet been tested by growth on a stable cost base.
C03facthigh confidence
SolarEdge states in both its FY2025 10-K and its Q2 2026 10-Q that excluding the advanced manufacturing production tax credits it recognises under Section 45X, it would have reported a gross loss rather than a gross profit - and in the 10-Q, for both Q2 2026 and Q2 2025.
The last report (5 Aug)
Quarter ended June 2026, filed figures.
20%
Sales growth
against the same quarter a year earlier
11.5%
Growth from the quarter before
against -7.4% the quarter before that
-5¢
Operating profit
per dollar of sales
The second quarter of 2026 looked like a turn: sales up 19.6% to $346.2M and the first adjusted operating profit since 2023, $10.2M. But the reported accounts still lost $16.0M from operations, and the gross margin included $13.3M of one-off tariff refunds.
The growth was batteries and markets abroad; US sales fell about 16%. The next quarter is guided to $310M to $340M at a lower margin, and the shares fell 30.5% the next day.
What it did to the thesisIt looked like a turnaround but supported our negative view. The profit leaned on one-off tariff refunds, the business still lost money on the reported measure, and American sales fell. The next quarter is guided lower, and the data-centre option has no customer.
The quarter in detail
Sales growth from the previous quarter, filed figures. The company has restated its own history, so a quarter either side of the restatement may describe a different business; n/c marks one that does not compare.
Earnings per share $0.06 against $-0.02 expected; $-0.50 as filed, the gap being what the company adds back.
The evidence under this - 6 claims
C08facthigh confidence
In Q2 2026 SolarEdge reported revenue of $346.2M, gross profit of $95.2M or 27.5% of revenue, an operating loss of $16.0M, a net loss of $30.8M and diluted loss per share of $0.50.
C09facthigh confidence
On a non-GAAP basis SolarEdge reported Q2 2026 gross margin of 28.6%, operating income of $10.2M, net income of $3.6M and diluted earnings of $0.05 a share - its first non-GAAP operating profit since Q2 2023. The distance between the non-GAAP and the filed GAAP per-share figure was $0.56.
C10facthigh confidence
Q2 2026 gross margin, on both the GAAP and the non-GAAP basis, included a benefit of $13.3M related to IEEPA tariff matters, and a further $11.5M of IEEPA refunds was received in July 2026 and excluded from the Q3 2026 guidance.
C14inferencehigh confidence
Revenue from outside the United States was 55.3% of SolarEdge's total in Q2 2026 against 36.0% in Q2 2025, so United States revenue fell about 16% year on year while total revenue grew 19.6%.
derivedThe 10-Q states the two percentages directly. Applying them to the filed revenue of $346,245 thousand and $289,429 thousand gives United States revenue of about $154.8M against about $185.2M, a fall of about 16.4%. The arithmetic is mine; both inputs are the company's.
against this
- The 45X credit is earned on production in the United States and on sale to an unrelated person, not on the buyer being American - so falling United States revenue does not mechanically shrink the credit, and SolarEdge can in principle manufacture in Utah, Florida and Texas and ship to Europe while still earning it.
- The six-month figures are less extreme than the quarter's - 52.3% against 37.7% - so part of the Q2 gap is quarterly mix rather than a settled shift.
C11management claimhigh confidence
SolarEdge guided Q3 2026 to revenue of $310M to $340M and non-GAAP gross margin of 22% to 26%, against Q2 2026 revenue of $346.2M and non-GAAP gross margin of 28.6%. The chief financial officer said on the call the company expects 24%.
sourceSolarEdge Q2 2026 results release, Exhibit 99.1 to the 8-KprimarySolarEdge Q2 2026 earnings call transcriptprimary
against this
- Management attributes the sequential fall to scale against fixed costs in cost of goods and to European seasonality, and said that excluding the scale effect there is 'a small improvement quarter-over-quarter' - so the guide is not management conceding a deterioration in unit economics.
- Against that, the guide explicitly excludes IEEPA refunds while Q2's reported margin included $13.3M of them, so the two margins are not struck on the same basis and the underlying fall is smaller than 28.6% to 24% suggests. Management also said the midpoint implies a non-GAAP operating profit once the July refunds are included.
C12facthigh confidence
The Q2 2026 results printed on 2026-08-05 and the measured price reaction in the session that followed was minus 30.5%, the largest single-session fall in fourteen months.
The price, and where we disagree
In short. Full. It looks cheap only against a bigger, profitable company that no longer exists. The price pays for a recovery and an AI side-bet before the tax credit's share of profit is known.
Share price
$31.87
close,
Company value for each $1 of sales
1.6
its 6-year normal 3.5
The shares rose 14.8% in the twenty sessions into September's Investor Day, where management pitched the data-centre business, yet they sit 54.8% below their high. Short sellers hold 23.1% of the freely traded shares.
The next test is 4 November: whether the margin holds inside its guide without tariff refunds, and whether the credit is finally sized.
The evidence under this - 4 claims
C30facthigh confidence
Short interest in SolarEdge was 13,801,412 shares at the 2026-08-31 settlement, 23.1% of the float, up 9.9% on the prior settlement and up from 10,480,781 shares at the 2026-06-15 settlement - 4.69 days of average volume to cover.
C11management claimhigh confidence
SolarEdge guided Q3 2026 to revenue of $310M to $340M and non-GAAP gross margin of 22% to 26%, against Q2 2026 revenue of $346.2M and non-GAAP gross margin of 28.6%. The chief financial officer said on the call the company expects 24%.
sourceSolarEdge Q2 2026 results release, Exhibit 99.1 to the 8-KprimarySolarEdge Q2 2026 earnings call transcriptprimary
against this
- Management attributes the sequential fall to scale against fixed costs in cost of goods and to European seasonality, and said that excluding the scale effect there is 'a small improvement quarter-over-quarter' - so the guide is not management conceding a deterioration in unit economics.
- Against that, the guide explicitly excludes IEEPA refunds while Q2's reported margin included $13.3M of them, so the two margins are not struck on the same basis and the underlying fall is smaller than 28.6% to 24% suggests. Management also said the midpoint implies a non-GAAP operating profit once the July refunds are included.
C10facthigh confidence
Q2 2026 gross margin, on both the GAAP and the non-GAAP basis, included a benefit of $13.3M related to IEEPA tariff matters, and a further $11.5M of IEEPA refunds was received in July 2026 and excluded from the Q3 2026 guidance.
C05facthigh confidence
SolarEdge does not quantify the 45X credit anywhere in its FY2025 10-K or its Q2 2026 10-Q. The figure that decides whether the company has a gross profit at all appears as a sentence and never as a number.
The betSafe, little upside
If it works
no upside
none the evidence defends inside two to three years
If it fails
keep 100%
of today's market value: the rest of the company, less its debt
How we read the multiple
Counting its cash and debt, the company costs about $2.1B, about 1.6 times last year's sales against a normal of about 3.5. But the company behind that normal sold $2.98B a year. Today, without a US tax credit, it would sell at a loss.
Analysts expect $1.83 a share of adjusted profit in 2028:
- That needs a profit margin SolarEdge has reached once in five years.
- It leaves out pay in stock, about $1.28 a share a year.
- So most of it is a cost the official accounts charge.
Against its own history. Far cheaper than its own normal, but that normal belongs to a bigger and profitable company, so the gap measures the decline rather than a bargain.
The evidence under this - 4 claims
C19facthigh confidence
At $36.75 SolarEdge's market capitalisation is $2.24B and its enterprise value $2.10B, 1.58 times trailing twelve-month revenue of $1.33B - the sum of the four quarters the company itself filed - against a prior five-year median enterprise value to sales of 3.46 times.
C20inferencemedium confidence
At this enterprise value, analysts' consensus adjusted earnings of $1.83 a share for 2028 require an operating margin SolarEdge has reached once in five years - FY2021's 10.5% - and it reached it on revenue a fifth higher than the 2028 consensus.
derivedConsensus 2028 earnings of $1.82783 on twelve analysts, against consensus 2028 revenue of $1.63B, both from the pack's analyst estimates. On about 62 million shares that is roughly $113M of net income, or a 6.9% net margin, which needs an operating margin above it once tax and the notes' 2.25% coupon are paid. The pack's annual series gives operating margins of 10.5% in FY2021, 5.3% in FY2022, 1.4% in FY2023, minus 189.5% in FY2024 and minus 25.5% in FY2025; FY2021 revenue was $1.96B against the $1.63B consensus for 2028.
against this
- The consensus figure is adjusted, not filed, and the comparison to FY2021's GAAP operating margin is therefore not like-for-like - an adjusted 2028 margin could clear the bar while the GAAP one does not, which is exactly the pattern every recent quarter shows.
- FY2021's margin was earned without any 45X credit, since SolarEdge only began United States manufacturing in the fourth quarter of 2023. So the credit is a reason the margin could be reached on lower revenue this time, and the honest statement is that the price needs the best operating margin of the last five years restored on a fifth less revenue, with a subsidy that was not there the first time and starts phasing out in 2030.
C37estimatemedium confidence
Stock compensation ran $39.5M in the six months to June 30, 2026 - about $1.28 a share a year on 61.5 million shares - against the $1.83 of adjusted 2028 earnings analysts publish. Most of the expected adjusted profit is a cost the official accounts charge.
derivedStock-based compensation of $39,549 thousand in the H1 2026 cash-flow statement, annualised, over 61,512,619 shares outstanding at June 30, 2026, against consensus 2028 earnings of $1.82783 from the pack. The comparison assumes stock compensation stays at its current dollar run rate and the share count stays flat; both are stated assumptions, not forecasts.
against this
- Stock compensation has been falling, not flat: $50,687 thousand in H1 2025 against $39,549 thousand in H1 2026, a fall of 22%. Extended to 2028 the per-share burden would be materially lower than $1.28, and the company has cut headcount deliberately.
- Consensus itself may assume the add-backs shrink, in which case the comparison double-counts. What it cannot do is disappear: stock compensation is the largest single add-back in every quarter's reconciliation and is a recurring cost of employing the engineers the whole data-centre case depends on.
C03facthigh confidence
SolarEdge states in both its FY2025 10-K and its Q2 2026 10-Q that excluding the advanced manufacturing production tax credits it recognises under Section 45X, it would have reported a gross loss rather than a gross profit - and in the 10-Q, for both Q2 2026 and Q2 2025.
Priced against its rivals
Only Enphase Energy sells the same thing to the same buyers as a large part of what it does. Tesla does too, but as a small part of a much bigger company, so its price says little about this one. A price comparison needs at least two, so the price is judged against its own history instead.
Private or foreign rivals that cannot be priced: Sungrow Power Supply Co., Ltd., Huawei Technologies Co. Ltd., Sigenergy Technology Co., Ltd., SMA Solar Technology AG, Chint (Chint Power Systems).
The bet: what it pays if it works
This bet is on one business only: the power converter SolarEdge is building for AI data centres. The solar and battery business is counted separately, and at today's price it accounts for the whole company value, so the price asks nothing of the data-centre product.
We see no upside we can defend within two to three years: it has no customer or order yet, pilots come in 2027 and real sales only in 2028. If it never sells, shareholders keep the solar business.
The steps it has to take
- Done Medium-voltage to 800 VDC conversion stage running under load in SolarEdge's own laboratories
- In progress A named data-centre customer signs a purchase order due 2027-06
- Not started Data-centre pilots installed with a customer due 2027-12
- Not started NVIDIA names SolarEdge in its published 800 VDC partner ecosystem due 2027-12
- Not started Meaningful data-centre revenue reported due 2028-12
The detail behind the label
Rating
No real upside
the big win's payoff against the 2× and 5× lines
The option
free
the price pays nothing for Data-centre 800 VDC solid-state transformer
Market case
1.5×
an expensive bet if the market forecast holds
The evidence under this - 4 claims
C19facthigh confidence
At $36.75 SolarEdge's market capitalisation is $2.24B and its enterprise value $2.10B, 1.58 times trailing twelve-month revenue of $1.33B - the sum of the four quarters the company itself filed - against a prior five-year median enterprise value to sales of 3.46 times.
C20inferencemedium confidence
At this enterprise value, analysts' consensus adjusted earnings of $1.83 a share for 2028 require an operating margin SolarEdge has reached once in five years - FY2021's 10.5% - and it reached it on revenue a fifth higher than the 2028 consensus.
derivedConsensus 2028 earnings of $1.82783 on twelve analysts, against consensus 2028 revenue of $1.63B, both from the pack's analyst estimates. On about 62 million shares that is roughly $113M of net income, or a 6.9% net margin, which needs an operating margin above it once tax and the notes' 2.25% coupon are paid. The pack's annual series gives operating margins of 10.5% in FY2021, 5.3% in FY2022, 1.4% in FY2023, minus 189.5% in FY2024 and minus 25.5% in FY2025; FY2021 revenue was $1.96B against the $1.63B consensus for 2028.
against this
- The consensus figure is adjusted, not filed, and the comparison to FY2021's GAAP operating margin is therefore not like-for-like - an adjusted 2028 margin could clear the bar while the GAAP one does not, which is exactly the pattern every recent quarter shows.
- FY2021's margin was earned without any 45X credit, since SolarEdge only began United States manufacturing in the fourth quarter of 2023. So the credit is a reason the margin could be reached on lower revenue this time, and the honest statement is that the price needs the best operating margin of the last five years restored on a fifth less revenue, with a subsidy that was not there the first time and starts phasing out in 2030.
C22management claimhigh confidence
SolarEdge's 800-volt direct-current data-centre products are under development and validation and are not yet generally available; the three-phase medium-voltage to 800 VDC conversion stage is operating under load in the company's engineering laboratories.
against this
- The status is disclosed by the company itself in the same release that announces the progress, so it is a candid statement rather than one extracted from it.
- The engineering has moved: on the Q2 2026 call, five weeks earlier, management described demonstrating 'a working prototype' going from medium voltage to 800-volt regulated direct current, and said the proof-of-concept would reach full three-phase 34.5 kilovolt operation by the end of 2026. The Investor Day claim is a step along that stated path rather than a leap beyond it.
C23management claimhigh confidence
Management expects data-centre pilots in 2027 and meaningful revenue in 2028, and said at the 2026 Investor Day that it is in discussions with hyperscalers and neoclouds but holds no purchase order or agreement with a specific customer.
sourceSolarEdge Q2 2026 earnings call transcriptprimarySolarEdge Investor Day release: 'SolarEdge Advances Full DC…secondary
against this
- The 2027-pilot, 2028-revenue sequence tracks NVIDIA's own published schedule rather than running ahead of it, so the timeline is not obviously promotional.
- It is nonetheless a timeline management has already moved once inside two months: on the Q2 2026 call an executive said 'the next few months until the end of the year' for the full three-phase proof of concept, then corrected '2026 will be pilots' to '2027 will be pilots' mid-sentence. And no part of the sequence is contractual - the whole case rests on winning a customer that does not exist in any filing.
Where we disagree with the market
The market believes
The solar business is recovering, and the AI data-centre platform is a free option on top.
We believe
The gross profit is a subsidy, the operating loss is real, and the data-centre option has no customer and no chosen design.
The sentence that decides whether SolarEdge has any gross profit sits in the management discussion and is never turned into a number. NVIDIA's own partner list, which omits SolarEdge, is public.
We would be wrong on a named hyperscaler order for the solid-state transformer, or a quarter with gross margin above 26% and the credit sized and small.
What analysts say, and the evidence
The evidence under this - 5 claims
C03facthigh confidence
SolarEdge states in both its FY2025 10-K and its Q2 2026 10-Q that excluding the advanced manufacturing production tax credits it recognises under Section 45X, it would have reported a gross loss rather than a gross profit - and in the 10-Q, for both Q2 2026 and Q2 2025.
C05facthigh confidence
SolarEdge does not quantify the 45X credit anywhere in its FY2025 10-K or its Q2 2026 10-Q. The figure that decides whether the company has a gross profit at all appears as a sentence and never as a number.
C24facthigh confidence
NVIDIA's published 800 VDC partner ecosystem names Eaton, Schneider Electric and Vertiv as data-centre power system partners, ten silicon providers and five power-system component makers. SolarEdge appears in none of the categories, on either of NVIDIA's two public pages.
sourceNVIDIA 800 V HVDC architecture technical blogprimaryNVIDIA 800 VDC architecture product pageprimary
against this
- Absence from a partner list is not exclusion from a market. SolarEdge and NVIDIA jointly published a white paper, '800 VDC Protection and Grounding for AI Data Centers', on 2026-09-10 - so there is a working technical relationship that the partner pages, which may not have been refreshed, do not reflect.
- The partner categories are drawn around the parts NVIDIA specifies - silicon, power shelves, facility power systems - and a medium-voltage conversion stage may simply not map to one of them. Against that, 'data centre power systems' is exactly where such a product would sit, and the three names in it are the incumbents SolarEdge would have to displace.
C22management claimhigh confidence
SolarEdge's 800-volt direct-current data-centre products are under development and validation and are not yet generally available; the three-phase medium-voltage to 800 VDC conversion stage is operating under load in the company's engineering laboratories.
against this
- The status is disclosed by the company itself in the same release that announces the progress, so it is a candid statement rather than one extracted from it.
- The engineering has moved: on the Q2 2026 call, five weeks earlier, management described demonstrating 'a working prototype' going from medium voltage to 800-volt regulated direct current, and said the proof-of-concept would reach full three-phase 34.5 kilovolt operation by the end of 2026. The Investor Day claim is a step along that stated path rather than a leap beyond it.
C23management claimhigh confidence
Management expects data-centre pilots in 2027 and meaningful revenue in 2028, and said at the 2026 Investor Day that it is in discussions with hyperscalers and neoclouds but holds no purchase order or agreement with a specific customer.
sourceSolarEdge Q2 2026 earnings call transcriptprimarySolarEdge Investor Day release: 'SolarEdge Advances Full DC…secondary
against this
- The 2027-pilot, 2028-revenue sequence tracks NVIDIA's own published schedule rather than running ahead of it, so the timeline is not obviously promotional.
- It is nonetheless a timeline management has already moved once inside two months: on the Q2 2026 call an executive said 'the next few months until the end of the year' for the full three-phase proof of concept, then corrected '2026 will be pilots' to '2027 will be pilots' mid-sentence. And no part of the sequence is contractual - the whole case rests on winning a customer that does not exist in any filing.
What could go wrong, and how we would know
The risks
- Cash borrowed from suppliers. Without $188.3M of slower supplier payments, the half's operating cash flow would have been about minus $152M.
- More shares coming. $300M of convertible bonds can now be turned into shares, which would add about 14.2% to the count.
- Stock piling up. $599.8M of inventory is about 2.4 quarters of costs, and the quarter already carried write-downs.
- A subsidy with an end date. The company warns that if the credit fades without costs falling, it could return to a gross loss.
The evidence under this - 4 claims
C15facthigh confidence
SolarEdge's operating cash flow in the six months ended June 30, 2026 was $35.8M, of which trade payables contributed $188.3M. Capital expenditure was $12.0M. Without the payables movement, operating cash flow would have been about minus $152M.
C18facthigh confidence
SolarEdge's $300M of 2.25% convertible senior notes due July 1, 2029 convert at 29.1375 shares per $1,000, a conversion price of about $34.32. The sale-price condition was satisfied during Q2 2026, so holders may convert during Q3 2026; full conversion is about 8.74 million shares, 14.2% of the 61,512,619 outstanding.
against this
- SolarEdge bought capped call transactions for about $25.2M at issue, which the company says are expected generally to reduce the potential dilution on conversion or offset cash payments above principal - so the gross 14.2% overstates what a holder actually loses.
- On conversion the company may choose to settle in cash, shares or a combination, so the dilution is a company decision rather than an automatic share issue. Against that, settling in cash spends the net cash position the balance sheet's strength rests on.
C34facthigh confidence
At June 30, 2026 inventories were $599.8M against Q2 2026 cost of revenues of $251.1M - about 2.4 quarters of cost of sales on hand - and Q2 2026 carried an increase of $25.2M in indirect cost of revenues primarily from inventory write-down accruals.
C07management claimhigh confidence
SolarEdge expects the advanced manufacturing production credits to be phased out by the end of 2031, and states that reductions without an offsetting fall in manufacturing cost would adversely affect gross margin and 'may include transitioning into a gross loss'.
against this
H.R.1, enacted 2025-07-04, did not shorten the term of the 45X credits - the company says so itself - so the phase-out date is the one already in statute rather than a new political risk. The nearer risk runs the other way: the foreign-entity rules attached to 45X began applying from January 1, 2026 and the final rule had not been issued when the 10-K was filed, so eligibility can be lost years before the phase-out arrives.
What we still do not know
- How big the manufacturing credit is. SolarEdge has never given the number, only that gross profit would be a loss without it.
- Whether any data-centre buyer signs. Management says it is in talks but holds no order.
- Whether the 50% share of US commercial rooftops repeats, or was one quarter.
The evidence under this - 4 claims
C05facthigh confidence
SolarEdge does not quantify the 45X credit anywhere in its FY2025 10-K or its Q2 2026 10-Q. The figure that decides whether the company has a gross profit at all appears as a sentence and never as a number.
C03facthigh confidence
SolarEdge states in both its FY2025 10-K and its Q2 2026 10-Q that excluding the advanced manufacturing production tax credits it recognises under Section 45X, it would have reported a gross loss rather than a gross profit - and in the 10-Q, for both Q2 2026 and Q2 2025.
C23management claimhigh confidence
Management expects data-centre pilots in 2027 and meaningful revenue in 2028, and said at the 2026 Investor Day that it is in discussions with hyperscalers and neoclouds but holds no purchase order or agreement with a specific customer.
sourceSolarEdge Q2 2026 earnings call transcriptprimarySolarEdge Investor Day release: 'SolarEdge Advances Full DC…secondary
against this
- The 2027-pilot, 2028-revenue sequence tracks NVIDIA's own published schedule rather than running ahead of it, so the timeline is not obviously promotional.
- It is nonetheless a timeline management has already moved once inside two months: on the Q2 2026 call an executive said 'the next few months until the end of the year' for the full three-phase proof of concept, then corrected '2026 will be pilots' to '2027 will be pilots' mid-sentence. And no part of the sequence is contractual - the whole case rests on winning a customer that does not exist in any filing.
C28management claimlow confidence
Management said on the Q2 2026 call that SolarEdge reached 50% of United States commercial and industrial rooftop installations in the quarter, and attributed it to its technology and to foreign-entity and domestic-content compliance that the other two leaders, SMA and Chint, do not meet.
against this
- The figure is management's own, unsourced, stated in answer to an analyst question rather than in a filing, and appears in neither the results release nor the 10-Q. No independent measure of United States commercial and industrial inverter share was found; no trade association or regulator publishes one that was reachable.
- The characterisation of SMA and Chint is qualified by management itself - 'to the best of our knowledge' - and its consequence is qualified twice more in the same answer: it is 'not yet clear' when the FCC ruling will affect competitors' new products, or whether their existing products can continue to be imported at all.
- It is also a share of a shrinking pool for SolarEdge: revenue from outside the United States rose to 55.3% of the total in the same quarter, so United States revenue fell year on year while this share was allegedly being won.
What would change our view
- Unattractive→AttractiveThe subsidy is sized and the business earns without it
It needs all of:
- ○SolarEdge publishes the size of its manufacturing tax credit, and it is under half of gross profit.checked at the next results, due by 4 Nov
- ○Reported operating profit turns positive for two quarters running, with inverter sales growing again alongside batteries.checked at the next results, due by 4 Nov
- Reopens the researchThe data-centre story becomes a business
It needs any of:
- ○A named data-centre customer signs an order for the solid-state transformer.watched as it happens
- ○NVIDIA adds SolarEdge to its published list of 800-volt power partners.watched as it happens
What is coming
- 4 November 2026 Q3 2026 results: Gross margin against the 22%-26% guide without the July IEEPA refunds, whether the 45X credit is finally sized, and whether inverter units stop falling.
Every dated event ahead
- 4 November 2026 Q3 2026 results: Gross margin against the 22%-26% guide without the July IEEPA refunds, whether the 45X credit is finally sized, and whether inverter units stop falling.
- February 2027 Q4 2026 results and the FY2026 10-K: A full year's statement on gross profit without the credit, the cost base after a year without cuts to lean on, and the 2027 outlook.
- December 2027 First data-centre pilots, as management expects: Whether a named customer runs the solid-state transformer at all; no pilot by then leaves the AI-factory target with nobody behind it.
Since we wrote this
Written
11 Sep
Price the thesis read
$36.75
10 Sep
Latest close
$31.87
close, · -13.3% since
- 5 OctThesis reviewThe bet now names what it ridesThe owner asked that every bet name the macro trend behind it, after BWXT's bet was missing its chief executive's case for nuclear over gas.
- 4 OctThesis reviewOwner's checklist: revenue streams, cost of growth, management, catalystsThe owner's checklist found the method silent on who runs the company, what growth costs, where the money comes from and what is dated ahead; this pass fills those four blocks from the proxy and filings.
- 25 SepNewsOld class action closed, paid by insurers. View heldThe court approved the $55M settlement on 26 August, funded by insurers. It removes a cost from the balance sheet but touches neither the unsized credit nor the customerless data-centre product.
- 11 SepThesis writtenUnattractive, medium convictionThe price said AI data centres. The filings said that without a US manufacturing tax credit SolarEdge would report a gross loss, and the credit is never sized.
Also checked 5 times with nothing new to act on.
Sources 24
Where this came from
Sources - 24
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-02-25Item 1 business and competition, the 45X and H.R.1 discussion including the gross-loss sentence and the 2031 phase-out expectation, customer concentration at 18.6% and 35.8%, the single-segment statement, the FY2025 revenue, gross margin and net loss, and the government-grant accounting policybacksC01C02C03C05C07C32C33C48C49C54C55
- S02primarycompany filing2026-08-05the gross-loss-without-the-credit sentence repeated for Q2 2026 and Q2 2025, the Performance Measures unit table, the geographic revenue split, the inventory and prepaid-expense notes including the $234.6M government-authorities receivable, the convertible note terms and conversion trigger, and the legal-claims provisionbacksC03C05C08C13C16C17C18C19C34C36C48C49C54
- S03primarycompany release2026-08-05the GAAP and non-GAAP summary table, the $13.3M IEEPA benefit, Q3 2026 guidance, the condensed statements of loss, balance sheet and cash flows including the $188.3M payables movement, and the non-GAAP reconciliation with the FY2023 to FY2025 gross-profit historybacksC08C09C10C11C15C17
- S04primarytranscript2026-08-05SolarEdge Q2 2026 earnings call transcriptmanagement's answers on the Q3 gross margin, channel inventory, safe-harbour strategy, the FCC Covered List, the 50% commercial and industrial share claim and the solid-state transformer timeline; and six analysts' questions, which are a separate origin from the answers - the sequential-decline and guidance questions came from the analysts, not the company
Where we're going from here is we're going to spend the next few months until the end of the year to get the proof-of-concept prototype fully working at the full 3-phase 34.5 kilovolt voltage. And then 2026 will be pilots in -- at the data centers -- sorry, 2027 will be pilots at the data centers so that we could have a meaningful revenue in 2028. ... We have actually -- in the last report, we've actually gained share in the C&I market to the point of 50% of the rooftop C&I installations in that quarter. ... there have been 3 leaders in the C&I market: SMA, Chint and SolarEdge. The other 2, to the best of our knowledge, are not complying with both.
backsC11C23C28C29 - S05primaryvendor data2026-09-11SolarEdge research pack, frozen at content/SEDG/research-pack.jsonprice, market capitalisation and enterprise value, the volatility bucket and returns, the analyst split and estimates to 2030, the earnings history with filed GAAP earnings beside the adjusted actual, FINRA short interest across six settlements, insider filings, the measured news reactions, and the vendor annual and quarterly series with the cross-checks against the filingsbacksC09C12C19C21C30C31C33
- S06primaryregulator2024-10-2826 CFR 1.45X-3 - Eligible components (inverter credit amounts) US Department of the Treasury and Internal Revenue Servicethe statutory credit rate per alternating-current watt for each inverter type, and that a DC optimized inverter system qualifies as a microinverter where the inverter and the optimizers are sold together as a combined end product
the credit amount is equal to the product of 11 cents multiplied by the total rated capacity of the microinverter ... the credit amount is equal to the product of 2 cents multiplied by the total rated capacity of the commercial inverter ... the credit amount is equal to the product of 1.5 cents multiplied by the total rated capacity of the utility inverter ... the credit amount is equal to the product of 0.25 cents multiplied by the total rated capacity of the central inverter
backsC06 - S07primaryindustry data2026-09-11the 800 VDC timeline - 1 MW racks from 2027 and full-scale production coinciding with Kyber in 2027 - the 200 kilowatt limit on 54-volt distribution, the named partner categories which omit SolarEdge, and the statement that the solid-state transformer approach is still under study
NVIDIA and its partners are actively studying the CapEx and, OpEx, and safety implications of both traditional transformer-based and solid-state transformer (SST) approaches. ... NVIDIA is leading the transition to 800 VDC data center power infrastructure to support 1 MW IT racks and beyond, starting in 2027. ... Full-scale production of 800 VDC data centers will coincide with NVIDIA Kyber rack-scale systems in 2027. ... Traditional 54 V in-rack power distribution, designed for kilowatt (KW)-scale racks, isn't designed to support the megawatt (MW)-scale racks coming soon. ... As racks exceed 200 kilowatts, this approach begins to hit physical limits.
backsC24C25C26 - S08primarycompany filing2026-02-17the closest listed rival's own competition section, which names SolarEdge first in both the inverter and the storage markets and describes the substitution in its own words; and the absence of any megawatts-shipped disclosure, which is why the prize does not composebacksC27
- S09primarygovernment data2026-09-11EIA - Electricity generation, capacity, and sales in the United States US Energy Information Administrationthe installed base of United States distributed solar measured outside the industry - the base factor in the prize
at the end of 2025, the United States had 59,506 MW of small-scale solar PV generation capacity ... about 93 billion kWh were generated by small-scale PV systems
backsC35 - S10secondarycompany release2026-09-10the development status of the 800 VDC conversion stage, the absence of any named customer or order, the joint NVIDIA white paper described as a technology-neutral framework, and the $2.4B 2029 revenue target. Read through a reproduction because solaredge.com returns HTTP 403 on every path, so it is graded secondary rather than primary
These AI data center products are under development and validation, and are not yet generally available.
backsC21C22C23C42 - S11primaryindustry data2026-09-11a second NVIDIA page checked for SolarEdge's presence, which also does not name it, and confirmation that the standard is set jointly with Google, Microsoft and the Open Compute Project rather than by NVIDIA alone
With Google, Microsoft, and OCP, NVIDIA sets the 800 VDC standard to scale AI compute.
backsC24 - S12primarycompany release2026-02-18the full-year unit counts - 465.7 thousand inverters, 10.8 million optimizers and 928 MWh of batteries - which let the credit inference be cross-checked bottom-up; the FY2025 non-GAAP net loss of $140.3M and non-GAAP loss per share of $2.38 against $6.88 GAAP; the filed GAAP operating loss of $301.7M, which corroborates the filer's XBRL against the vendor; Q4 2025's $70.5M non-cash currency charge on the Korean liquidation; and the chief executive's framing of 2026 as a transformational year investing in AI data centre powerbacksC19C38C55
- S13primaryother2026-08-26In re SolarEdge Technologies, Inc. Securities Litigation, No. 1:23-cv-09748-GHW (S.D.N.Y.) - Final Order and Judgment, ECF No. 177, via CourtListener's RECAP archive United States District Court for the Southern District of New Yorkfinal approval of the $55M settlement on August 26, 2026 after the August 24 hearing, dismissal with prejudice, attorneys' fees of $13,750,000 and expenses of $750,000 paid from the settlement fund, and the objection of Avrohom Senderovic overruled as untimely; the docket index beside it records his notice of appeal from the separate order denying his objector service award (ECF 178).
Accordingly, the Settlement is hereby approved in all respects (including, without limitation: the amount of the Settlement; the releases provided for in the Stipulation; and the dismissal with prejudice of the claims asserted against Defendants)
backsC36 - S14secondarynews2026-09-28Latitude Media - Does the market finally have an opening for solid-state transformers? (reporting a Jefferies equity research note) Latitude Media, reporting Jefferiesa bank's forecast of the US data-centre SST market in 2030 and today, the price range per watt, the 2027-2029 adoption window, and the rival pilots (DG Matrix, GE Vernova, Amperesand, Enphase) with DG Matrix's 11 active units and $700-900 million of orders it says it is closing; the note itself is paywalled, so this is a reporter's account of it; read through ir_fetch
In a recent equity research note, Jefferies analysts wrote that the SST market for U.S. data centers could reach $4.3 billion in 2030, up from $37 million today. ... At the moment, SST prices vary from $30 cents to $1 per watt, which is the equivalent of $300,000 to $1 million per megawatt, according to Jefferies ... Analysts believe SST adoption at scale will start happening between 2027 and 2029
backsC40 - S15primaryindustry data2026-07-01MarketsandMarkets press release - Solid-state Transformers Market for US Data Centers worth $154.0 million by 2030 MarketsandMarketsa market-research firm's forecast of the same US data-centre SST market, about 28 times below Jefferies' for the same year, with renewable integration rather than 800 VDC racks its largest application; the publication date is the PR Newswire/GlobeNewswire release date; read through ir_fetch
The solid-state transformers market for US data centers is expected to reach USD 154.0 million by 2030 from USD 40.3 million in 2025, at a CAGR of 30.8% during the forecast period.
backsC41 - S16primarycompany filing2026-07-28a direct rival's solid-state transformer progress: RFI and RFP stages with a potential multi-gigawatt pipeline, a 15-module series stack at 4.16 kV, and a full-system demonstration later in 2026; with the Q1 2026 release (1.25 MW IQ SST, 342 power modules, native 800 V DC) found by EDGAR full-text searchbacksC43
- S17secondarynews2026-04-29pv magazine USA - Enphase announces distributed solid-state transformer for AI data centers, targets 2028 for volume deliveries pv magazine USAthe rival's own sizing of the US addressable market; FMP's transcript of the same Q1 2026 call reads 'exceed 11 gigawatts by 2021', an evident transcription error for 2031, which this report and the call's context settle; read through ir_fetch
Enphase said its engineers have built functioning samples of the power modules and that it intends to conduct full system demonstrations later this year, with customer pilots in 2027 and volume shipments in 2028. On the earnings call, Kothandaraman said he expects the addressable market for the IQ SST to reach 11 GW by 2031.
backsC43 - S18primarycompany filing2026-07-31Eaton's acquisition of Resilient Power Systems, a solid-state transformer developer, on 6 August 2025 for $86 million into Electrical Americas, so an incumbent on NVIDIA's partner list owns SST technologybacksC44
- S19primarycompany filing2026-07-27a power-semiconductor supplier's own timetable for 800 V content: SiC power supplies from 2H26, 800 V sidecars mid-2027, native 800 V trays mid/late 2027 to 2028, and solid-state transformers '2028+'backsC45
- S20secondarynews2026-09-13management's own Investor Day sizing of the AI-factory market and its 2029 revenue target for it, the split of the $2.4 billion 2029 target ($1.8 billion core, $600 million AI factory) that C21 did not carry; an automated account of the webcast, so graded secondary, and the company's own claim, so grouped with S10; read through ir_fetch
It projected the total addressable market for its AI-factory power products could increase from $200 million in 2027 to $4.1 billion in 2030. ... It forecast $600 million in AI-factory revenue in 2029, while acknowledging during the question-and-answer session that it does not yet have firm purchase orders or customer agreements for the products.
backsC42 - S21primaryvendor data2026-10-01FMP key metrics (TTM) and annual analyst revenue estimates for VRT, ETN, GEV, HUBB, POWL and ENPH, fetched 2026-10-01 Financial Modeling Prepeach peer's enterprise value and consensus 2028 revenue, giving forward EV/sales, read through scripts/api/fmp.pybacksC46
- S22primarycompany filing2026-04-23beneficial ownership as of April 6, 2026, director and executive officer biographies, the CEO and CFO succession, the 2025 bonus plan's measures, weights and payout, the PSU and RSU design, and the outcome of the 2023 and 2024 PSUsbacksC51C52C53
- S23primarycompany filing2026-05-11Asaf Alperovitz leaving the CFO role on May 31, 2026, Maoz Sigron's appointment from that date, his record at Perion, and his pay terms including PSUs on total shareholder return against the Solar IndexbacksC53
- S24primaryvendor data2026-10-04FMP earnings calendar for the roster, saved by scripts/earnings_calendar.py at content/calendar/2026-10-04.json Financial Modeling Prepthe date of the Q3 2026 print
"SEDG": {"archived": false, "next": {"date": "2026-11-04", "eps_actual": null, "eps_estimated": 0.02, "revenue_actual": null, "revenue_estimated": 330971400}
backsC55