SolarEdge's $2.4 Billion Plan Rests on an AI Data-Center Order It Doesn't Have Yet

SolarEdge wants to be a $2.4 billion company by 2029. On the surface that sounds like a comeback story from a fallen solar stock. Look closer and the honest score is this: the market is already giving the company credit for the comeback it has banked, and for the AI upside it has not. The stock is up about 27% year to date and rallied again around its September 10 Investor Day, where management laid out the number.

At its 2026 Investor Day, SolarEdge set three long-range targets for fiscal 2029: $2.4 billion in revenue, a 35% gross margin, and $360 million in EBIT. Run the arithmetic before you react. Trailing revenue across the last four reported quarters is roughly $1.33 billion, so the plan calls for roughly an 80% increase. The $360 million EBIT on $2.4 billion implies a 15% operating margin, up from a business that only just returned to non-GAAP operating profitability this spring.

The interesting part of that bridge is not the solar base. It is the new claim sitting on top of it.

The AI story is a narrative, not a contract

The reason $2.4 billion matters more than the last few growth quarters is the company's new AI data-center pitch. SolarEdge is building a direct-current power architecture meant to carry electricity from the medium-voltage grid all the way down to the compute rack, and it claims the medium-voltage-to-800 VDC conversion stage is now operating under load. Together with Nvidia, it published a joint white paper on 800 VDC protection and grounding for AI data centers. CEO Shuki Nir framed the moment plainly: the transition to 800 VDC will hinge on whether these systems can be protected, serviced, and trusted at scale.

Here is the gap that separates the headline from the economics. The company says it is in discussions with hyperscalers and neoclouds — and management acknowledges it has no concrete purchase order or agreement with a specific customer yet. A white paper co-authored with Nvidia is a proof-of-capability document, not a demand signal. As of today, the revenue line that would justify an AI-powered rerating has exactly zero booked orders behind it.

That is the scoreboard the Investor Day set up. The core solar turnaround is real — six consecutive quarters of revenue and margin growth, with fiscal Q2 revenue up 19.6% year over year to $346.2 million and $10.2 million in non-GAAP operating income, the first positive quarter on that measure since Q2 2023. That is a genuine expectation reset for a company that came out of the 2023–2024 residential-solar bust. But the multiple the market is now paying embeds a second story on top of it, and that second story has not produced a customer.

Credibility you can bank, upside you cannot

SolarEdge's recent history is the reason to take the turnaround seriously rather than dismiss the whole pitch. The company was a marquee residential-solar name until the industry hit an inventory glut and high-rate demand collapse, forcing layoffs and, by August 2024, the departure of CEO Zvi Lando. Revenue bottomed near $196 million in the fourth quarter of 2024 before recovering in a straight line through last year.

Against that baseline, the current run rate and the first non-GAAP operating profit in two years are earned, not borrowed. The market has noticed: the stock trades near $37 with a market value around $2.3 billion, and an enterprise value of roughly $2.0 billion against about $1.33 billion in trailing revenue — roughly 1.5 times sales for a company that only just crossed into non-GAAP operating profitability. A solar-equipment maker at that multiple is being valued for conversion it has not yet demonstrated.

The deeper question the Investor Day leaves open is whether SolarEdge can convert the AI infrastructure pitch into something that belongs on a backlog. Rerating in this market happens when installed capability connects to a named buyer, a schedule, and a contract. SolarEdge has the capability stage in the lab and the white paper, but no buyer yet. Until a hyperscaler or neocloud signs, the AI leg of the $2.4 billion plan is a roadmap, not a roadmap with revenue.

What to watch

We only need to watch one thing to know whether the story is advancing: a named customer or a purchase order in the 800 VDC data-center power business. That single event would convert the AI narrative into the beginning of a backlog and justify treating the $2.4 billion target as an emerging revenue path rather than a marketing horizon. The signal that breaks the current narrative is the opposite — a repeated Investor Day where the flagship AI product still has no booked demand, at which point the stock is paying an AI multiple for a solar company.

Until then, hold the two numbers separate. The solar recovery is real and worth monitoring on ordinary operating metrics like gross margin stepping toward that 35% target. The AI rerating is a bet on conversations turning into contracts. The stock has already absorbed a good chunk of that upside; the next leg depends on the order book, not the white paper.