Enlight Paid $20 Million to Unsign Two PPAs — That's the Tell, Not the Beat

A record Q3, a guidance raise built partly on Israeli power trading, and a data-center pitch that just got noticeably more guarded.

ENLT.TA · Earnings Call · 2026-11-03

The beat was real; the mix is where the work is

Enlight's third quarter was genuinely strong: revenue and income of $231 million, up 40%; adjusted EBITDA $154 million, up 38%; operating cash flow near $104 million, up 26%. “Total revenues and income were $231 million up 40% year over year. While adjusted EBITDA increased 38% to $154 million” — Adi Leviatan, CEO · 2026-11-03 Management raised and narrowed 2026 guidance — revenue to $810–830 million from $790–820 million, adjusted EBITDA to $575–590 million from $565–585 million. But look at the shape of the raise. The revenue midpoint moved roughly $15 million while the EBITDA midpoint moved roughly $7.5 million. That asymmetry is the identical fingerprint to the prior quarter's raise, when Itay Banayan explained the gap directly: “we increased the guidance for the year for the revenues more than the EBITDA, and this is part of the reason” — Itay Banayan, Chief Corporate Development Officer · 2026-08-04 — the low-margin, low-capital Israeli electricity trading book. Same lever, pulled again, only bigger. That business is now a real line item, not a footnote. Adi Leviatan described it as a genuine extension of the platform: “we have also started to trade on electricity that is manufactured by other producers in Israel. This has been going. It has exceeded our expectations” — Adi Leviatan, CEO · 2026-11-03 The electricity trading keyword enters the company's top-ranked themes this quarter alongside trading activity and activity in Israel. Trading contributed $34 million of revenue in the quarter, itself up $7 million year over year, with essentially no G&A attached. That also re-frames the inevitable Q4 deceleration question. Justin Clare of ROTH asked straight out about the implied slowdown in Q4 growth. Banayan's answer was pure timing: “there is no slowdown. it is just a matter of specific quarter timing of COD of projects.” — Itay Banayan, Executive (likely CFO or similar) · 2026-11-03 Fair — but two-thirds of the answer is that the trading book is seasonal, and management said so: “it is expected to be lower than Q3” — Itay Banayan, Executive (likely CFO or similar) · 2026-11-03 because Israeli demand peaks on summer cooling, not winter heating. Investors should model the trading revenue as weather-and-arbitrage, not annuity.

The most under-read line in the release: paying to uncontract

Here is what actually changed this quarter. Enlight walked away from two contracted projects — Gemstone and Rustic Hills, 39 factored megawatts combined — and paid for the privilege.

we have chosen to go back to the market with our gemstone and Rustic Hills projects... This decision was made following rounds of negotiation with our previous contract holders. Noting market conditions for both input costs and energy pricing have greatly changed since the legacy contracts were originally signed. This decision required $20 million in termination fees. Which we believe we will more than recover with fresh contracts in the current high demand environment.
Jared McKee, Executive (likely Head of Development or similar) · 2026-11-03

For a company whose entire prior-call rhetoric was built on contracted baseload revenue and "healthy spreads between project returns and our cost of capital," voluntarily burning contracts is a statement of market view — and a confident one. The PPA termination charge is the price of admitting the legacy paper was struck in a different world. Site control is fully retained ("we own the land"), so optionality is intact. The test is whether those two projects re-price before or after year-end; management guided to a disclosure window of Q4 into early next year, and flagged advanced conversations with counterparties. The same logic underwrites their tariff posture. Jared McKee argued Section 32 cost inflation flows through: “We have seen there be a shift in the PPA prices... So most if not all, of this impact will be born in those increased prices.” — Jared McKee, Executive (likely Head of Development or similar) · 2026-11-03 Roughly 18 factored gigawatts are safe harbored, and the 2027–2028 COD tranche reported no cost creep. That is a credible moat — but it is a moat of procurement, not of pricing power, and the market has historically paid for the former only grudgingly.

Large load: swagger gives way to disclosure discipline

Last quarter, Adi Leviatan was unambiguous about wanting to be the data center, not just its landlord: “this electricity can be used for accreditation for large loads for data centers, and we want to utilize that ourselves” — Adi Leviatan, CEO · 2026-08-04 This quarter, the tone tightens to something closer to a regulated utility's: “we are not at a point in time where we can talk specifically about the offtake for blackwater.” — Adi Leviatan, CEO · 2026-11-03 The project itself progressed — local permitting cleared, COD targeted 2030, construction from 2027, 500 MW AC of solar with storage expansion available under the interconnection, and safe harbored. The keyword map tells the story of the pivot cleanly: Blackwater project surfaces for the first time as a top-tier company theme, and large load takes the number-one slot — displacing data center, which had held the top spot the prior quarter. The company is now describing itself in the language of capacity accreditation and bring-your-own-generation rules, not in the language of hyperscaler logos. Two honest notes crept in. First, local opposition is now acknowledged rather than glossed: management conceded pushback "in certain counties in The US" while arguing the binding constraint is power, not permits — “The primary reason why projects may not move forward is power.” — Jared McKee, Executive (likely Head of Development or similar) · 2026-11-03 Second, the framing with which Leviatan opened the call — infrastructure spending by "the large technology company" now being questioned on "scale timing, and returns" — reads as pre-emptive positioning. He is right to pre-empt. Across the market tape, AI data center is one of the most punished theme clusters of the last thirty days, with fifty-five tickers in the negative column against ten positive. Meanwhile large load is now boilerplate across the utility sector — PGE, Talen, Pinnacle West and MasTec all leaned on it this reporting cycle. Enlight's new number-one keyword is therefore consensus, not company-unique, and it arrives exactly as the market stops paying up for the adjacency.

What management stopped saying

Three prior obsessions went quiet, and that matters more than what got louder. Safe harbor was the centerpiece of both the February and May 2026 calls — the 14–17 factored gigawatt target, the July 2026 deadline, the deliberate project-by-project selection. Adi's framing then: “We do have the opportunity to safe harbor an additional 2 to 4 factored gigawatt” — Adi Leviatan, CEO · 2026-05-05 This quarter it is a single line in the prepared remarks (~18 factored GW), absent from the keyword set. That is what a completed task looks like — fine, but it removes the quarterly narrative crutch. Second, asset sell-downs vanished. The Sunlight cluster partial sale powered 2025 and was explicitly contemplated in early-2026 guidance; by August, Banayan was saying no further sell-downs were in the numbers. Now the funding story is entirely organic: roughly 70% of project financing secured and 60% of required equity already invested on the 8.7 factored gigawatt construction pipeline. Cleaner earnings quality, one fewer lever. Third, the European storage buildout — demand for storage and ancillary services, both prominent last quarter — graduated from narrative to numbers: two more Finnish projects for 858 MWh, a mature European portfolio now at roughly $630 million of annual revenue run rate, doubled in a year. Similarly, the $70 million of tax benefits embedded in Q3 revenue shows tax benefit has become an accounting reality rather than a talking point. The defensible read: Enlight is executing a large, well-funded 2028 plan (over $2.2 billion revenue run rate, 12 factored gigawatts, roughly 41% compounded growth) at ~13% unlevered and over 18% levered returns, funded from a $5.6 billion debt stack that is 80% non-recourse and close to 90% fixed-rate. The risk is less about delivery and more about narrative: the growth engine that the tape currently loves to hate is the one Enlight just made its headline. Watch the two re-priced PPAs and Q4 COD timing — those, not the 40% revenue print, are the next real data points.