Ashdod Refinery 2025: Q4 Brought The Refinery Back, but 2026 Will Test Paz, Exports, and Power

Company Overview

Ashdod Refinery looks stronger if you focus on the fourth quarter, and riskier if you focus on 2026. The company ended 2025 with adjusted EBITDA of $166 million, adjusted operating profit of $97 million, cash of $289 million, and equity of $477 million. After an unusual defective feedstock event in the third quarter, the fourth quarter already returned to reported EBITDA of $173 million and adjusted EBITDA of $120 million. That means the operating recovery is real.

This was still not a normal run rate year. The company recognized a $47 million indemnity asset from the supplier, $40 million of insurance income, and an $85 million provision on defective inventory. At the same time, the 2026 discussion has already moved beyond plant recovery. The real question is what happens when Paz falls from 48% of sales volumes in 2025 to about 30% of forecast total volumes in 2026.

What is working now? Refinery utilization recovered to 84% in the fourth quarter, the power plants remained highly stable, and refined product margins, especially diesel and jet, improved. What is still not clean? The customer base is concentrated, exports absorb logistics costs, the larger power strategy is not yet cash generating, and the cash flow story still rests not only on earnings but also on supplier credit, bank funding, and covenant definitions that were tailored to the incident.

Ashdod Refinery is still, above all, a single asset company. It has one site in Ashdod, nominal refining capacity of about 5.4 million tons per year, and two cogeneration power plants producing about 109 MW of electricity together, with roughly 30% consumed internally and the balance sold externally. This is a model in which refining margin is the main engine, but not the only one. Power sales, dispatching, crude mix flexibility, and the ability to place volumes in the domestic market or in exports are all part of the economics.

This is also a relatively young standalone company after the separation from Paz in 2023. That makes 2025 the year in which the market starts to test whether the refinery can operate not only with a large anchor customer, but with a broader commercial base. That is the point a superficial reader may miss if they stop at the fourth quarter rebound.

The Economic Map

At the end of 2025 the company had 443 employees. Against revenue of $3.028 billion, that comes to roughly $6.8 million of revenue per employee. Market cap stood at roughly NIS 1.1 billion in early April 2026. This is not a tiny speculative story, but it is also not yet a broad energy platform. One site, a handful of large contracts, and heavy working capital needs still move the entire picture.

This chart highlights an important point. The decline in 2025 revenue did not come from a collapse in the domestic market. Domestic sales were almost flat, while the main weakness came from exports. That matters because the 2026 debate is not only about how much volume the company can sell. It is about which channel absorbs that volume.

This is the heart of the early read. Anyone reading Ashdod Refinery as a pure refining margin story misses that a large part of the real business question is customer concentration. Two customers represented 62% of revenue. That provides commercial anchoring, but it also means that the 2026 Paz reset is not a side topic. It is part of the thesis itself.

Events And Triggers

The first trigger: the defective feedstock event was much more than a temporary operating fault. During the third quarter of 2025 the company received defective feedstock from a large international supplier, and refinery operations stopped running normally. The company gradually returned to producing standard products, had to shut down part of the facilities again in early August, and only from October returned to near full output. In January 2026 it completed an additional regeneration process that was required partly because of the incident, but part of the damage will only be handled during the 2027 turnaround, which was pulled forward.

The implication is that the fourth quarter really does reflect recovery, but 2025 is still not fully closed as a historical event. The tail of the incident continues into 2026 through defective inventory, recovery claims against the supplier and insurers, and the earlier turnaround schedule.

This chart explains why the fourth quarter matters, but also why it is not enough. The third quarter drop was sharp, and the fourth quarter return was almost full both in throughput and in utilization. That supports the argument that the plant is back. It still does not answer what a clean refining earnings year looks like through 2026.

The second trigger: the renewal of credit facilities at the end of December 2025 and start of January 2026 gave the company important breathing room, but also showed how material the incident had been. Bank facilities were renewed to NIS 1.925 billion for 12 months, and receivables discounting frameworks were renewed to NIS 495 million. That is positive, because it shows the banking system stayed with the company after the incident. But the banks also agreed that, for covenant testing, losses on sales of defective inventory and all related insurance proceeds would be excluded, and in practice the supplier indemnity asset was also neutralized. Even the lenders were not reading 2025 as a normal year.

The third trigger: the debt market did not treat the year with the same calm. On November 18, 2025, Midroog downgraded bond series 2 and 3 to Baa1.il with a stable outlook. As a result, the coupon on the outstanding principal of both series rose by 0.25% on each series. This is not a balance sheet crisis, but it is a clear sign that the market saw the incident as a real deterioration in risk profile.

The fourth trigger: the new 2026 Paz agreement changes the direction of the story. Paz represented 49% of revenue and 48% of sales volumes in 2025, but forecast Paz volume for 2026 falls to about 30% of total forecast volume. The company has already signed part of the missing volume with other domestic customers and is working to place the rest in the domestic market and exports. That reduces dependence on an anchor customer, but it does not necessarily preserve the same economics because export volumes carry logistics costs.

Efficiency, Profitability And Competition

The improvement in 2025 profitability is real, but it rests on three different layers: better market margins, a relatively fast return to normal operations, and unusual accounting recognition around the incident. Anyone who compresses those three layers into one clean picture will read 2025 as stronger than it really was.

What Actually Drove Profit

Net sales fell by $188 million to $3.028 billion. Management's explanation is straightforward: average selling prices fell by about 10%, partly offset by about 4% higher sales volumes, including discounted sales of non standard products. Against that, cost of sales fell by $216 million to $2.966 billion, partly because average crude prices were lower, but also because the supplier indemnity asset of $47 million was recorded as a reduction of cost of sales.

That is how gross profit rose to $62 million from $34 million in 2024, and operating profit rose to $73 million from only $5 million. But that operating improvement also includes $35 million of other income, mainly the insurance compensation. Anyone trying to understand actual earning power therefore needs to look at the adjusted figures as well.

This chart matters because it prevents two opposite mistakes. The first mistake is to say everything was accounting noise. That is wrong, because 2025 is much stronger than 2024 even on an adjusted basis. The second mistake is to say the year was clean. That is also wrong, because the gap between reported and adjusted figures, together with the fact that the banks themselves used a special covenant EBITDA definition, shows that 2025 is still far from a simple baseline year.

The fourth quarter reinforces that point. Gross profit jumped to $127 million from $16 million in the comparable quarter, and operating profit to $156 million from $7 million. But the headline number is still too generous on its own because it includes insurance and supplier related recognition. Adjusted fourth quarter EBITDA was $120 million versus $16 million a year earlier, which is still a very strong result. In other words, the fourth quarter does prove a real operating recovery, just not to the full extent implied by the reported headline alone.

Price, Volume, and Mix

The company ultimately lives on refining margins, but not on a generic margin concept. In 2025, Mediterranean benchmark margins for diesel, gasoline, and jet all improved versus 2024, with the stronger move in middle distillates: diesel margin rose to $24.7 per barrel from $22.0, and jet to $23.1 from $20.9. In the fourth quarter that environment strengthened further, with $30.7 in diesel and $29.6 in jet.

What worked for the company was also the refinery's own configuration. Ashdod Refinery competes in the local market against Bazan and imports, and estimates its 2025 market share at about 35% to 40% of Israel's refined products market. The advantages it cites are real: geographic proximity to demand centers, proximity to crude import infrastructure and marine connectors, and a gasoline biased configuration that fits the local market.

But not every barrel is equal. The company explicitly says that in export sales it absorbs the relevant logistics costs. That is why the shift from 80% domestic sales in 2024 to 85% in 2025 is not only a geographic data point. It is also a quality improvement. The problem is that 2026 may push part of the missing Paz volume back toward exports precisely because of the new contract.

Another important layer is power. The two power plants together generate about 109 MW, of which only roughly 30% is consumed internally. The rest is sold to private customers and to the system operator. That supports the total refining margin, but it cannot be read as a standalone power business. All of the steam produced by the plants is used by the refinery, so the power activity still depends on refinery operations.

Cash Flow, Debt And Capital Structure

The right way to read Ashdod Refinery in 2025 is through an all in cash flexibility lens, not through net profit or through "net financial debt" alone. This is a company that managed to remain cash stable after a major operating event, but it did so inside a heavy working capital funding system.

Cash Flow

Operating cash flow rose to $130 million from $109 million in 2024. That is a good number, especially after the year the company had. But it is not enough to stop there. After $23 million of capex, $39 million of interest paid, $22 million of bond repayment, $4 million of long term loan repayment, $8 million of lease principal repayment, and $11 million of net repayment of bank loans and credit, cash increased by only $23 million.

That is the number that really matters. It is not negative. On the contrary, it shows the company got through a difficult year and still finished with more cash. But it also shows that the $130 million operating cash flow should not be read as fully spare cash. That is precisely why an all in cash flexibility frame is more useful here.

Debt, Covenants, and What Really Funds Working Capital

At the end of 2025 the company had $289 million of cash, $36 million of current bond and loan maturities, $228 million of long term bonds, and $26 million of long term bank loans. Equity stood at $477 million, and working capital at $116 million.

In the presentation, the company shows net financial debt of only $1 million. That is correct under the narrow definition used there, bonds plus loans minus cash. At the same time, the board report shows net financial debt of about $47 million under the bank financing agreements. These are not contradictory numbers. They are two different bridges. But neither one, on its own, tells the full funding story because working capital also rests on large supplier credit.

The numbers here are large. Average supplier credit during 2025 stood at about $406 million, and year end interest bearing supplier balances were about $337 million. So even if narrow financial leverage looks low, Ashdod Refinery is still a business that funds working capital through suppliers, banks, and receivables discounting. That is normal for the sector, but it means the balance sheet should not be read as if the company is economically debt free.

Covenant headroom, however, is very wide. Against the banks, equity stood at $477 million versus a $232 million minimum, equity to assets at 34% versus 20%, adjusted EBITDA at $166 million versus a $58 million minimum, and net debt to adjusted EBITDA at 0.3 versus a ceiling of 4.8. The bond covenants are also wide, with net debt to adjusted EBITDA at 0.0 versus a 5.5 ceiling.

There is still one yellow flag here. Covenant comfort did not come only from a stable balance sheet. It also relied on a specific adjustment: the banks agreed to exclude losses on sales of defective inventory and all related insurance proceeds, and in practice the supplier indemnity asset was also neutralized. That does not mean the company is weak. It does mean the lenders understood that 2025 was unusual and hard to measure through ordinary covenant tools.

Outlook

Finding one: 2026 is no longer an operating recovery year. It is a commercial proof year.

Finding two: the fourth quarter proves the plant is back, but it does not prove that 2025 net profit is the right earnings base for 2026.

Finding three: the power strategy can change the angle of the story, but for now most of the value is still created in refining, not in a new power plant.

Finding four: the balance sheet buys the company time, but it does not remove the test. Exports, customers, and working capital will decide whether 2026 is truly strong.

The First Test: Replacing Paz Volume Without Buying Volume at a Bad Price

The drop in Paz from about 48% of sales volumes in 2025 to roughly 30% of forecast total volume in 2026 is the most important issue for the coming year. Not because Paz is disappearing. It is not. The real issue is that this gap forces Ashdod Refinery to prove it can replace volume in the domestic market or in exports without damaging margin quality.

The 2026 order backlog stands at about $1.5 billion versus $2.2 billion a year earlier, and the company says explicitly that the decline is driven mainly by the lower quantities under the Paz agreement. The company has already sold part of the gap to other domestic customers and is working to sell the rest in the domestic market and exports.

This chart sharpens the core issue. The decline in backlog is not mainly a generic market problem. It is a concrete commercial reset with an anchor customer. If Ashdod Refinery fills most of the gap in the domestic market, the picture remains reasonable. If a large share has to move to exports, readers should expect some pressure on margin quality because of logistics costs.

The Second Test: How Fast 2025 Cleans Up After the Incident

Even after a strong fourth quarter, 2026 will continue to carry the tail of the defective feedstock event. In January 2026 the company completed a successful regeneration process, but part of the damage will only be addressed in the 2027 turnaround. In the first quarter of 2026 it had already sold part of the defective feedstock inventory. That means the event will keep flowing through the accounts even after core operations recovered.

Another non obvious point is that the inventory debate is not fully settled only at management level. The auditor identified inventory valuation as a key audit matter. Inventory stood at about $230 million after a provision of about $99 million, mainly related to damaged crude and product inventory. At the same time, the board report discusses an $85 million provision on defective inventory. The right reading is not to force a dramatic accounting contradiction. It is to understand that damaged inventory remained a material judgment issue even at year end.

That is exactly why 2026 will not be measured only in EBITDA. It will also be measured by whether defective inventory sales, cash recoveries from the supplier and insurers, and the effect on adjusted EBITDA are closed at a reasonable scale. The company itself says it cannot currently estimate the future scope of recoveries it may receive.

The Third Test: Does Power Become an Engine, or Remain an Option

The most interesting strategic option at Ashdod Refinery is not another barrel. It is power. The company is at an advanced stage in developing a high voltage electricity storage project of about 55 MWh together with a 5 MW steam turbine. Estimated cost is about $30 million, mostly expected in 2026. The company expects the storage facility to begin operating during 2026 and the turbine to be integrated in the first half of 2027.

At the same time, the company is promoting a framework plan for a third power plant on site with capacity of about 800 MW. That is a much bigger move, and a much less mature one. It depends on planning, approvals, and authorization, even if the company has real location and infrastructure advantages.

This is where created value needs to be separated from accessible value. The value being created today sits in the existing power activity, which already contributes to the total margin. The value that is still not accessible sits in the expansion plan. Until the storage project and the regulatory path of the larger plant are actually advanced, this remains more strategic direction than proven profit engine.

The Fourth Test: Margin Support Is Real, but Hedging Is Partial

There is also good news heading into 2026. The company says it placed partial hedges on refined product margins for 2026, and increased the hedge level in the first quarter of 2026 after a meaningful strengthening in diesel and jet margins. As of the report approval date, the main hedges for 2026 covered about 20% of estimated annual diesel and jet output, and about 12% of total estimated annual production.

That helps, but it does not lock in the year. Most of production remains exposed. That is why 2026 looks like a proof year, not a breakout year. If the strong middle distillate environment holds, and if the company fills the Paz gap without buying volume at a bad price, the picture can improve quickly. If margins cool and exports absorb more volume than planned, the fourth quarter may end up looking like a peak quarter rather than a new normal.

Risks

Customer Concentration and Volume Replacement

The first risk is commercial. Paz accounted for 49% of revenue and 48% of volume in 2025, and another material customer for 13% of revenue. Even after the planned reduction in Paz volumes in 2026, the customer base remains concentrated. In addition, the company makes clear that it has no binding backlog beyond the current calendar year, except for power sales and propylene sales to CAOL. That means visibility is inherently short.

External Infrastructure Bottlenecks

The second risk is external and operational. The company depends on infrastructure owned or operated by KAA, Tasha, KMD, and IEC, both for product flows and for crude storage and delivery. In particular, the temporary business license for IEC's marine connector ran, according to the report, until May 1, 2026. The company states explicitly that only if approval is granted for continued operation beyond that date, or if a replacement connector is built, will it be able to export excess fuel oil through that infrastructure. Without an alternative, operating continuity could be affected.

Single Site Risk, Incident Tail, and the Earlier Turnaround

The third risk is single asset exposure. The 2025 incident showed how exposed a single refinery site is to shutdowns or partial throughput. Even after the return to near full operation, part of the damage will only be addressed in the 2027 turnaround. That means 2026 begins from a cleaner operating point, but with a pulled forward turnaround and with the possibility that additional implications still appear along the way.

Working Capital Funding, Rating, and Gas Dependence

The fourth risk is financial. The bond rating was cut to Baa1.il, bank lines were renewed but remain short dated, and the company also relies on large supplier credit. On top of that, it depends on two main natural gas suppliers, Tamar and Leviathan, to run several facilities and the power plants in particular. There is no immediate sign of stress here, but it does mean balance sheet flexibility must be judged together with supply chain access and funding cost, not in isolation.

Conclusions

Ashdod Refinery finished 2025 in much better shape than the market could reasonably have feared in mid year. The fourth quarter showed a real plant recovery, the balance sheet remained comfortable, and the covenants are not close. The main bottleneck has shifted from restoring the facilities to the quality of commercialization: how missing Paz volume is replaced, how much export is required to do it, and what happens to margin and cash flow in the process.

Over the short to medium term the market is likely to focus on three things: how clean the first 2026 reports look without incident noise, whether the Paz gap is filled mainly in the domestic market or through exports, and whether the power projects on the table start turning into practical milestones. The story is no longer only "the refinery is back." The story is whether that recovery can be turned into a cleaner year.

Current thesis in one line: Ashdod Refinery exited 2025 as a more stable operating and financial company, but 2026 will determine whether that recovery becomes a cleaner earnings engine, or only a bridge year managed through compensation, exports, and working capital funding.

What has changed versus the mid 2025 read is sharp. Then the debate was about whether the refinery could return to normal production. Today the debate has moved to customer mix, earnings quality after the incident, and whether the power layer can start generating broader cash value. In other words, risk has moved from the plant itself to the commercial and capital engine around it.

The strongest counter thesis is that the cautious read may still be too harsh. On that view, the fourth quarter already proved the damage was temporary, contractual leverage is low, cash is high, short interest is negligible, and the decline in Paz could ultimately improve the company's independence and open a healthier customer base. If that is correct, 2025 will not be a bridge year but the start of normalization.

What can change market interpretation over the short to medium term is a combination of three confirmations: early 2026 reports without fresh surprises from defective inventory, evidence that the Paz gap is being filled without a sharp deterioration in margin quality, and practical progress on the storage project. This matters because for a refinery it is not enough to show that the plant is working. It also has to show that the right volume is being sold through the right channel, and that real cash remains after all funding needs.

Over the next two to four quarters, the thesis will strengthen if the company sustains high utilization, replaces the Paz gap mainly in the domestic market, closes a meaningful part of the accounting and economic tail of the incident, and advances the storage project without stretching the balance sheet. It will weaken if exports become the main absorption route, defective inventory keeps creating volatility, or the power plans remain mostly conceptual.