Delek Group: 2025 looks great, but the real 2026 test is whether value actually reaches the top

Company Overview

The easy mistake in reading Delek Group is to think 2025 proves the transformation is already complete. On the screen, almost everything looks right: attributable net profit of 3.256 billion ILS, the completed Isracard acquisition, a sharp production jump at Ithaca, real strategic progress at Leviathan, an A2.il rating, more than 1 billion ILS of dividends, and a fresh buyback plan. The market has largely responded in kind. As of April 9, 2026, market capitalization stood at about 18.5 billion ILS, and short interest was down to just 0.12% of float. This is no longer a stock the market reads as a distressed leveraged energy name.

But that is still a partial reading. A large part of 2025 profit did not come from recurring operating improvement. It came from a roughly 2.678 billion ILS gain recorded when Delek stopped consolidating Ithaca. Once management strips out that item, together with Isracard-related credit-loss and transaction costs and the deferred-tax adjustment in the North Sea, attributable net profit falls to 1.003 billion ILS. That is still a respectable earnings base, but it tells a very different story: less of a harvest year, more of a transition year from a leveraged energy company into a holding company that now has to prove that value created inside the assets can actually move upward.

What is working now is clear enough. Leviathan still generates cash and dividends even after operational interruptions, Ithaca sharply increased production and EBITDAX after the ENI combination, Isracard gives the group a real financial engine, and Mehadrin is building a real-estate layer that could become a secondary growth driver over time. The problem is that these engines sit at different layers of the structure, in different currencies, and under different risk regimes. That is why Delek's current bottleneck is no longer just leverage. It is value accessibility: how quickly dividends, cash and financing capacity can actually move up to the parent, and how much of the value remains trapped below as minorities, reinvestment needs, or accounting gains that do not translate into cash.

That also explains why 2026 looks like a holding-company proof year. For the reading to improve, three things need to happen together: Leviathan has to return to normal operations and use the 14 BCM step-up, Ithaca has to sustain large dividends even after Delek's lower ownership share, and Isracard has to prove that it adds real earnings power rather than just financial and regulatory complexity. The market is already giving credit for the strategic direction. What it wants now is evidence.

A quick economic map helps frame where the story really sits:

• Delek now operates through three main engines: Israel and East Med gas and oil, North Sea oil and gas, and finance through Isracard. Beyond that, it has a smaller layer of holdings in Mehadrin, Delek Israel and Delek Assets.
• As of the report date, the company held 54.18% of NewMed participation units, 50.47% of Ithaca, 40.05% of Isracard and 61.36% of Mehadrin.
• On a segment basis, 2025 activity reached 16.129 billion ILS, but the shape of that activity changed materially because Ithaca left consolidation while Isracard entered it.
• At the parent level, the funding picture is much cleaner than before: solo working capital of 423 million ILS, 500 million ILS of undrawn bank lines, a solo equity-to-balance-sheet ratio of about 64%, and an A2.il rating with stable outlook.
• The practical constraint is no longer liquidity in the shares or heavy short pressure. The constraint is how much value can really move from the subsidiaries to the parent while the group is still investing, refinancing and returning capital.

Events And Triggers

The first trigger: 2025 was a structural-change year, not just a result year. From September 30, 2025, Delek stopped consolidating Ithaca and began presenting it as an associate. That created a gain of roughly 2.678 billion ILS, but it also changed the right way to read the group. Delek now looks less like a fully consolidated energy operator and more like a parent company living off dividends, equity-method earnings and capital allocation across several engines.

That bridge is the heart of the story. Without separating the one-off gain from the recurring earnings base, it is easy to walk away believing the group suddenly changed its underlying earnings power. That would be the wrong read. 2025 was strong, but it was not clean.

The second trigger: Isracard closed on July 24, 2025 and turned Delek into a holding company with a real financial leg. At the subsidiary level this is a large business, with 4.75 million active cards, more than 3.7 million private customers, more than 100 thousand business customers and an 11.8 billion ILS loan book. At the consolidated Delek level, the first year looks much less flattering: finance contributed 1.823 billion ILS of revenue in 2025 but still ended with a 171 million ILS net loss because of purchase-accounting charges and other first-year effects. That is exactly the difference between owning a good business and already showing clean value creation to Delek shareholders.

The third trigger: Leviathan went through both operating disruption and strategic acceleration in the same year. The June 2025 shutdown cost the partnership roughly 33 million USD of lost revenue and about 20 million USD of net profit. After year-end there was another shutdown, beginning on February 28, 2026, and as of the report approval date there was still no update on the restart date. At the same time, the third pipeline was completed, enabling a rise to 14 BCM per year, and on January 15, 2026 a final investment decision was approved for the expansion to 21 BCM, with a budget of about 2.36 billion USD and first gas targeted for the second half of 2029. The strategy is moving forward. The path to that value is still exposed to real operating and geopolitical risk.

The fourth trigger: Bulgaria reminded investors that exploration optionality is not a free upside call. At the start of 2026, NewMed and OMV brought in BEH for a 10% stake in the license, reducing NewMed Balkan's drilling commitment from 100 million EUR for two wells to 90 million EUR. That is the positive side. The negative side came immediately afterward: the Vinekh well was assessed as a dry hole, and drilling, plugging and abandonment costs were estimated at 86 million EUR, implying about 76 million USD of loss for NewMed Balkan, already recognized in the 2025 statements. This is not existential, but it is a useful reminder that not every option inside the portfolio is built like Leviathan.

The fifth trigger: At the parent level, 2025 and early 2026 show a management team that is much more comfortable with the capital structure. The group raised about 3.3 billion ILS through new bonds and expansions, used roughly 2.8 billion ILS for early redemptions, moved its rating up to A2.il, and after year-end launched a new 100 million ILS buyback program. In March 2026 Midroog also rated a framework of up to 500 million ILS for a possible issuance of new series. This is a positive signal on market access. It also means the parent is still being managed actively, not as a structure whose balance-sheet story is fully over.

Efficiency, Profitability And Competition

The confusing part of 2025 is that Delek presents two revenue numbers that are both correct but mean very different things. On a segment basis, activity revenue rose to 16.129 billion ILS. On the consolidated P&L, revenue was 4.148 billion ILS in general activities plus 1.823 billion ILS in finance. Anyone who looks only at the 16.1 billion ILS headline can walk away thinking the group suddenly became a giant consolidated revenue machine. In reality, that figure describes the full segment activity base, while the accounting perimeter already changed through Ithaca's exit from consolidation and Isracard's entry into it.

That chart sharpens two points. First, the center of gravity moved even further toward the North Sea. Second, the addition of finance changes the group even before it fully changes the earnings line available to Delek shareholders.

Leviathan And NewMed

In Israel and the surrounding gas activity, 2025 was weaker than the broader narrative suggests. Segment revenue fell to 3.033 billion ILS from 3.661 billion ILS, and the contribution of this activity to attributable net profit fell to 766 million ILS from 974 million ILS. NewMed itself tells the same story: revenue declined to 880 million USD from 989 million USD, and EBITDAX fell to 695 million USD from 814 million USD. That does not reflect a damaged asset base. Leviathan still sold 10.9 BCM in 2025, including 6.2 BCM to Egypt, 2.9 BCM to Jordan and 1.8 BCM to the domestic market. But the June shutdown, the planned interruption tied to the third pipeline work, and the Vinekh dry-hole loss in Bulgaria all hit the quality of the year.

That is the difference between asset quality and year quality. Leviathan remains a very strong asset, with a new 130 BCM export agreement to Egypt, a completed third pipeline and a final investment decision that points toward 21 BCM. But the 2025 numbers still show that the path to that value is not smooth. The market does not just need the big story. It needs operating continuity.

Ithaca

Ithaca is the place where it is easiest to confuse a real operating improvement with a weaker capture rate at the parent. Operationally, 2025 was very strong: average daily production jumped to about 119 thousand boe per day from about 80.2 thousand in 2024, and Adjusted EBITDAX rose to 2.025 billion USD from 1.405 billion USD. That is a real step-up and it explains why management keeps presenting Ithaca as the portfolio asset that offsets some of the East Med geopolitical risk.

But here comes the less intuitive part. Delek no longer owns Ithaca the way it did in 2024. Its weighted average holding in 2025 was about 52%, versus about 79.5% in 2024. The report even gives a useful bridge: if Delek had owned only 52% in 2024, Ithaca's contribution to net profit before one-offs would have been about 602 million ILS, versus about 537 million ILS in 2025. In other words, even inside a year of obvious operational improvement, the part that actually remains with Delek did not rise nearly as much as a superficial read of Ithaca's standalone operating numbers would suggest. That is the essence of the Delek holding-company story: the assets can improve faster than the listed parent layer.

Isracard

At Isracard, the picture is almost the mirror image. At the subsidiary level, 2025 looks good: revenue of 3.522 billion ILS versus 3.244 billion ILS, adjusted net profit of 305 million ILS versus 264 million ILS, card volume of 256 billion ILS, and a loan book of 11.8 billion ILS. That is a meaningful platform, with 4.75 million active cards and 1.378 million non-bank cards.

But at Delek level, 2025 was mainly a year of absorption. The finance segment still reported a 171 million ILS net loss, largely because of acquisition-related accounting and other initial costs. So anyone who looks at Isracard only through Delek's finance line can miss that there is already a real business engine here, while anyone who looks at it only through the presentation can miss that Delek has not yet proven how much of that engine becomes clean earnings power at the parent.

What matters even more is that Delek is not presenting Isracard as a side holding. It is presenting it as a financial platform. The non-binding ESH bank MOU signed in March 2026, together with the purchase of 25% of the technology company serving that bank, signals that the goal is not merely to own a credit-card company but to build a broader financial engine. That could create value. It could also consume capital, management bandwidth and regulatory patience.

Cash Flow, Debt And Capital Structure

When analyzing Delek in 2025, the right cash framework is all-in cash flexibility. The key question is not only how much cash the businesses generated before investment. The real question is how much room remained after acquisitions, exploration, dividends and debt service.

At the consolidated level, operating activities generated 2.713 billion ILS of cash in 2025. That is a strong number, but it did not translate automatically into free balance-sheet room. Investing activities used 3.717 billion ILS, including 834 million ILS for acquisitions of control in companies and activities and 1.039 billion ILS for oil and gas investment. Financing activities used 604 million ILS net, after 3.298 billion ILS of bond issuance and 4.228 billion ILS of bond repayments. On top of that, shareholders received 1.035 billion ILS in dividends. Put differently, 2025 was not a year in which accounting profit simply accumulated into cash. It was a year of active capital recycling.

At the parent level, the picture is cleaner. Subsidiaries distributed 2.6 billion ILS in 2025, of which Delek's share was about 1.37 billion ILS. Against that, Delek distributed about 1.035 billion ILS to its own shareholders and used only about 34 million ILS out of a 105 million ILS buyback plan during 2025. That supports two readings at once. On the positive side, management has fairly high confidence in the continuity of upstream cash flows. On the other side, the pace of distributions at the parent is already high, so any slowdown at NewMed or Ithaca would be felt faster at the top.

The positive part is that the parent no longer sits close to the wall. At year-end it had 500 million ILS of undrawn bank lines secured by NewMed units. Solo working capital was 423 million ILS, and the solo equity-to-balance-sheet ratio was about 64%. That is far above the 22% threshold required for dividends under the bond deeds and comfortably above the 18% covenant level tied to financial metrics. Distributable profits were also abundant at 9.579 billion ILS at year-end. So the problem today is not tight covenant headroom at the parent. The problem is capital-allocation judgment: how much to return and how much to retain as real balance-sheet room.

There is also a useful external signal that supports this reading. In March 2026, Midroog affirmed Delek at A2.il with a stable outlook and also rated a framework of up to 500 million ILS for potential new bond issuance. At NewMed level the financing picture is even wider: asset value to net financial debt stood at 4.71 as of September 30, 2025 against a minimum threshold of 1.5 in the relevant debt bucket, solo liquidity stood at 469.6 million USD against a 20 million USD minimum, and total financial debt was about 1.4 billion USD against a 3.0 billion USD cap. In other words, the financing structure around the core gas asset is currently far from a squeeze.

But there is a counterpoint here too. When a parent company refinances debt, upgrades its rating, distributes 1 billion ILS, launches a new buyback and still keeps optionality for new issuance, the message is not just "the balance sheet is clean." It is also that the group continues to run as an active capital-allocation machine. As long as dividends from NewMed and Ithaca keep flowing, that works. If one of the engines weakens, the same posture will look much less comfortable.

Outlook And Forward View

Four points should frame 2026:

• Point one: 2026 is not a maximum-accounting-profit year. It is a proof year for value accessibility at the parent.
• Point two: Leviathan is running on two clocks at once, a near-term operational restart and a long-dated expansion to 21 BCM only in the second half of 2029.
• Point three: Ithaca can continue to offset East Med risk, but Delek's share is smaller than before, so dividends matter more than gross operating statistics.
• Point four: Isracard is already a real business engine, but the ESH transaction is still not a fact. Until there is a binding agreement, it is strategic upside, not a forecast base.

Leviathan And NewMed

The market's first real test will be simple: a normal restart at Leviathan. The third pipeline is already complete, so once the field returns to regular operations it can move to 14 BCM per year. Above that sits the longer-duration engine, the final investment decision for the move to 21 BCM, with a budget of about 2.36 billion USD and a first-gas target in the second half of 2029. That matters a lot, but it is not 2026 cash. What determines the near-term read is whether Leviathan returns, serves its Egypt, Jordan and domestic customers without extended interruptions, and continues to move cash upward.

Post-balance-sheet dividends support that near-term bridge. On March 15, 2026, NewMed declared an additional 70 million USD dividend, of which Delek's share was about 38 million USD. That is useful short-term support at the parent. It is still not a substitute for multi-year proof.

Ithaca

Ithaca enters 2026 from a relatively strong position. In the investor call, management pointed to average production of 120 to 130 thousand boe per day and annual dividends in a 472 to 520 million USD range. After year-end Ithaca already declared an additional 200 million USD dividend, of which Delek's share was about 101 million USD. If the company meets those targets, Ithaca should remain a genuine cash offset to the geopolitical risk around Leviathan.

The critical point is that the market now has to shift from "Ithaca is strong" to "Ithaca keeps sending cash to Delek." Those are not the same thing. After deconsolidation, dividends matter more than full consolidation optics.

Isracard And ESH

In finance, 2026 is an even clearer proof year. Isracard itself already sits on a broad operating base, but Delek now has to show that this contribution becomes stable group-level earnings rather than just a strategically attractive asset. The non-binding MOU signed on March 17, 2026 values ESH bank at 400 million ILS. Under the outline, 250 million ILS would be paid in Isracard shares at closing and another 150 million ILS would remain in escrow subject to milestones. In parallel, Isracard would acquire 25% of the technology company serving the bank for 40 million USD, with the possibility of up to 100 million ILS of additional cash consideration over five years if defined events occur.

This has strategic logic because it aims to broaden Isracard from a payments-and-credit platform into a wider financial-services platform. But it is still non-binding, subject to definitive agreements, diligence, approvals and uncertain timing. So in 2026 the ESH thread can add upside to the thesis, but it cannot be treated as if it already exists.

What Has To Happen For The Read To Improve

In short, 2026 looks less like an automatic breakout year and more like a year in which Delek has to prove that each engine can create value and coexist inside one capital structure.

Risks

The first risk is geopolitical and operational, and it is not theoretical. Leviathan was already shut in during June 2025 at a cost of roughly 33 million USD in revenue and about 20 million USD in net profit, and after year-end it was shut again with no known restart date as of report approval. The entire improvement case for 2026 starts with whether the field returns to normal.

The second risk is the gap between value created and value accessible at the parent. Delek owns very good assets, but not all at 100%, and not all through full consolidation. A meaningful part of value moves through dividends, equity-method accounting, minority interests and reinvestment decisions at the subsidiary level. That is exactly why Delek can show strong assets and still remain dependent on steady upstream cash flow.

The third risk is energy prices and currency. Even though 2025 looks strong on net profit, total comprehensive income was only 1.968 billion ILS. The main reason was a 1.822 billion ILS loss in other comprehensive income, including 1.389 billion ILS of foreign-operation translation effects. This is a reminder that the group lives deep inside dollar exposure and FX translation, and that equity can move sharply even when headline profit looks strong.

The fourth risk is regulatory and capital intensity in finance. Isracard itself operates in a competitive market and inside a dense regulatory environment. If Delek advances the ESH deal, it adds banking-license complexity and a more demanding capital-allocation question. That can be a value-creating move. It can also slow down value realization.

The fifth risk is execution across the growth investments. Leviathan still has to move through expansion, Bulgaria has already taken its toll, and Mehadrin is trying to build a broader real-estate engine through Pi Engineering and additional post-balance-sheet land wins. Any of these moves can work. None of them is free of timing, capital and execution risk.

Short Interest View

In the equity itself, there is currently no sign that the market is holding a highly aggressive short-side counter-read. Short interest as a percentage of float fell from 0.50% in late December to 0.12% on April 10, 2026, and SIR dropped to 0.25 versus a sector average of 1.606. That is far from a picture of strong market dissonance against the fundamentals.

That does not mean the shares are protected. It means the near-term battle over Delek is less likely to come from technical short pressure and more likely to come from how investors interpret operating execution. If Leviathan restarts, Ithaca keeps paying and Isracard stabilizes, there is currently no large short base waiting to fight that move. If one of those pieces breaks, the pressure is more likely to come from long holders resetting expectations than from a crowded short trade.

Conclusions

Delek ends 2025 in a much stronger place than it occupied a few years ago. The parent balance sheet is cleaner, the core assets are better, and the group is no longer dependent on only one energy story. But this report does not prove the puzzle is solved. It proves the question has changed: less "is there value here," more "how does that value move upward and how much of it really remains for common shareholders."

Current thesis: Delek already looks like a multi-engine holding company with high-quality core assets, but 2026 will test whether the value created at Leviathan, Ithaca and Isracard actually becomes accessible cash and capital freedom at the parent.

What changed is not just the numbers. It is the nature of the company itself. Ithaca moved from consolidated subsidiary to associate, Isracard entered as a financial engine, and the discussion shifted from crude leverage and debt stress to capital allocation and value accessibility. That is a dramatic improvement. It still needs proof.

The strongest counter-thesis is that the market is already right and the proof is effectively already visible: Ithaca improved materially, Leviathan keeps advancing, NewMed and Ithaca continue to distribute cash, and Isracard adds a new leg, so even if 2025 was distorted by accounting, the underlying cash stream moving upward will keep strengthening. That is an intelligent objection. The problem is that it still depends too heavily on three engines that have not yet delivered a single clean year together.

What could change the market's interpretation in the short and medium term is fairly clear: Leviathan's restart, Ithaca's ability to meet its 2026 targets, continued dividends from the subsidiaries, and real clarity on whether ESH moves from a non-binding process to an actual transaction. If these happen together, the read on Delek improves quickly. If one of them breaks, especially Leviathan or the upstream dividend stream, the market will quickly remember that multi-layer holding companies are ultimately judged by cash that reaches the top.

Why does this matter? Because Delek's move from a leveraged energy company to a diversified holding company can justify a very different reading of business quality. But that shift is worth a lot only if shareholders actually benefit from it, rather than merely reading about it through accounting profit or asset values.

What has to happen over the next 2-4 quarters for the thesis to strengthen is straightforward: Leviathan must return to normal operation, Ithaca must keep paying, Isracard must show a cleaner contribution, and the parent must preserve capital discipline while still investing and distributing. What would undermine the thesis is just as straightforward: more Leviathan shutdowns, weaker dividend visibility from below, or an overly aggressive financial expansion before clean earnings are visible.