Bazan 2025: Refining Recovered, but the Bottleneck Moved to Polymers and Cash Flexibility

Getting To Know The Company

Bazan is not just a refinery. It is an integrated industrial chain in which refining, aromatics, and polymers sit on the same operating backbone: the refinery produces feedstocks, Carmel Olefins and Gadiv pull them through the system, and intermediate streams flow back into the chain. That is why anyone reading 2025 only through the refining margin misses the core story. The strongest engine recovered, but the bottleneck moved downstream, mainly to polymers and to the question of how much of the cash being generated is actually left after restoration spending, investment, debt service, and distributions.

What is working now? The refining segment finished 2025 with $412 million of adjusted EBITDA, up from $363 million in 2024, and the group as a whole rose to $446 million from $396 million. Even in a year that included major turnarounds, a direct hit on the energy center, and only a gradual return to operations, the adjusted refining margin improved to $11 per barrel from $10.2 per barrel in 2024. That is a real improvement, not cosmetic.

What is still not clean? First, that adjusted EBITDA already includes $160 million of business interruption insurance income. It is still a useful metric for stripping out inventory and derivative noise, but it is not a metric that says the war is fully out of the system. At the same time, the polymers segment moved to negative EBITDA of $29 million after positive EBITDA of $9 million in 2024, and in the fourth quarter alone it posted negative EBITDA of $30 million. Anyone summarizing the year as a return to normal is getting there too early.

That is exactly what a superficial read may miss: Bazan no longer looks like a balance-sheet survival story, but rather like a gap between a refining business that is back to generating money and a downstream chain that is still not producing a comfortable return on the capital and cash it consumes. Cantium, acquired in August 2025 and adding $39 million of adjusted EBITDA in the period, is not a magic solution either. It also brings debt, hedging requirements, distribution limits, and a $350 million abandonment and restoration obligation at the partnership level.

In trading terms, this is not a trapped stock story. As of April 6, 2026, the shares traded about NIS 13.9 million in daily value, and short interest stood at 0.67% of float with an SIR of 1.93 days. The debate here is therefore not about illiquidity or an unusually crowded short. It is about whether 2026 becomes a proof year for a clean earnings recovery, or just a transition year in which refining carries the story while polymers, Cantium, and the restoration effort continue to absorb a large part of the oxygen.

Events And Triggers

The hit on the energy center changed the year, not just the quarter

The first trigger: June 2025 split the story in two. The direct hit on the energy center, alongside additional damage to pipelines and transport infrastructure, shut down all refining facilities in the second quarter and pushed the third and fourth quarters onto a gradual return-to-operations path. The company has already restored steam facilities that allow full ongoing operations with sufficient redundancy, but full steam redundancy is only expected to be completed during 2026.

The numbers matter here, but the gap between them matters even more. The company estimates direct damage at $150 million to $200 million, received an advance of NIS 160 million from the compensation fund, and recognized $160 million of business interruption insurance income. By the time the report was approved, $155 million had been collected from insurers, yet $58 million of the claim was still not recognized in the accounts and there is no certainty it will be received in full. In other words, the report already contains meaningful compensation, but the file is still not fully closed.

Cantium opened a third leg, but not a shorter route to shareholders

The second trigger: In August 2025 Bazan expanded beyond Haifa Bay and invested $100 million in Cantium through a partnership in which it holds roughly 52% of the equity rights. At the group level this is a real addition: $39 million of adjusted EBITDA in five months, $16 million of share in profits of equity-accounted associates, and a new layer of exposure to WTI rather than only to refining and polymer spreads.

But that trigger has to be read in both directions. At the Cantium level there was only $2.8 million of cash at year-end, debt of $64.5 million as of December 31, 2025 that rose to $85 million by March 23, 2026, and a $350.1 million abandonment and restoration obligation alongside a dedicated $70.6 million escrow account. So Cantium improves Bazan's EBITDA picture, but it is not equivalent, dollar for dollar, to cash that is immediately accessible to Bazan shareholders.

The dividend says management is more comfortable, and also marks the boundary of that comfort

The third trigger: The company paid $50 million of dividends during 2025, and after the balance-sheet date approved another $35 million distribution. The message is clear: management is not reading the situation as a liquidity crisis. That also fits with a net debt to adjusted EBITDA ratio of only 1.7x under covenant calculations, and with solo cash plus committed credit lines of more than $1 billion.

But the distribution also has a cost. The same year included $259 million of fixed-asset purchases, $103 million of investment in associates, and a $134 million decline in cash and cash equivalents. So the 2025 dividend does not reflect a deep excess-cash position. It reflects a management choice that there is enough room. That is an important choice, but not a free one.

After the balance sheet, the energy market got louder and operations still sent a reminder

The fourth trigger: Between the balance-sheet date and report approval, Brent traded in a very wide range of $62 to $119 per barrel, and near report approval stood around $105. At the same time, diesel and gasoline spreads moved sharply higher on the back of regional supply disruptions. That is clear near-term support for refining.

On the other hand, in March 2026 the group site and a critical external infrastructure asset owned by a third party were hit again in a separate event. According to the company, the damage from that incident was not material, but it is still a reminder that restoration is not simply a past-tense event. Even if the system is back to work, the operating fragility of a single-site complex in Haifa Bay remains higher than the market is likely to assume comfortably.

Efficiency, Profitability And Competition

Refining recovered, and the key number is not revenue

Group revenue fell to $5.84 billion from $7.54 billion, but that is almost secondary. At Bazan, value is created by margin and utilization, not by top-line volume. On that front, 2025 does look stronger: adjusted refining margin improved to $11 per barrel from $10.2, and pro forma adjusted refining margin rose to $12.2 per barrel from $9.6 in 2024. The reference margin also improved to $9.9 from $7.9.

That means Bazan did not only benefit from a better environment. It also maintained a positive spread above the reference margin. That matters especially because the year also included an $8 million loss on the realization of refining-margin hedges. Refining did not get an easy paper year here and still improved.

Even at the utilization level, the reported headline is weaker than the economic reality

Reported refinery utilization fell to 68% from 76%, and total throughput declined to 7.24 million tons from 8.81 million tons. But the company estimates that 2025 pro forma utilization, excluding the direct hit and the turnarounds, stood at 87% versus 83% in 2024. Even in the fourth quarter, while the return to operations was still not fully complete, pro forma utilization was 89%.

The implication is that the refinery did not lose its economic production capability. It simply went through a year with much more downtime. Anyone trying to extrapolate 2026 only from the reported 2025 utilization will therefore get to a picture that is too conservative.

Polymers are not the second engine right now, they are the main drag on the thesis

The polymers segment moved from positive EBITDA of $9 million in 2024 to negative EBITDA of $29 million in 2025. This is not just an operational disruption story. Yes, Carmel Olefins operated only partially after the hit and Ducor was shut for scheduled turnarounds, so polymer output fell to 471 thousand tons from 639 thousand tons. But the margin environment also remained difficult: the polypropylene margin over naphtha fell to $544 per ton from $608 per ton.

There is a more important point hiding underneath that. Carmel Olefins is the only Israeli producer of polypropylene and low-density polyethylene, but that local position does not cancel what actually determines the economics: feedstock prices, distance from target markets, and relatively limited scale in global terms. A domestic monopoly does not automatically translate into a comfortable margin.

In addition, year-end 2025 included inventory write-downs of $9 million on Carmel polymer products and another $5 million on aromatics products. This is not an existential event, but it is a signal that pressure in polymers had already moved beyond current EBITDA and into inventory values as well.

At Ducor, management already shifted into defensive mode

An $8 million impairment was recorded at Ducor already on June 30, 2025. No additional impairment was recorded at year-end, but that was not because the environment became clearly better. It was because the company decided to change the way the asset is being used, lowering the operating rate of its production lines to the level required for contractual customer commitments while reducing spot-market exposure.

That is what really matters: before any visible recovery in margins, one of the downstream assets already moved from trying to maximize volume to trying to defend margin. It is a rational reaction, but it also says the group itself is reading the environment much more cautiously.

The bottom line does not capture refining, but it is also not lying about the burden

Reported EBITDA fell to $360 million from $403 million, and net income dropped to $47 million from $113 million. Two mistakes need to be avoided here. One is to read the year only through net income and miss the operating recovery in refining. The other is to get excited about adjusted EBITDA without remembering that it includes $160 million of business interruption insurance.

The right read sits in between: the core economics of the refinery improved, but 2025 is still not a normal year. It is a year in which refining proved it came back while polymers, insurance, and restoration were still mixed into the result.

Cash Flow, Debt And Capital Structure

I am using an all-in cash flexibility frame here, not normalized cash generation. The reason is that the main question at Bazan right now is not how much refining could generate in a sterile world. It is how much cash is actually left after investment, restoration, debt service, distributions, and the other real cash uses of the year.

Operating cash flow was solid, but the year still consumed cash

Cash from operations reached $297 million in 2025. That is a good number, and it already includes $123 million received from insurers for business interruption. The problem is that the line did not stay available. Investing activity consumed $302 million, mainly $259 million of fixed-asset purchases and another $103 million of investment in associates. Financing activity consumed another $129 million. The result was that cash and cash equivalents fell from $752 million to $615 million.

Put simply, 2025 was not a year of cash cushion building. The operating business generated cash, but almost all of those dollars went back out through CAPEX, Cantium, debt service, leases, and dividends.

2025 CAPEX was not normal CAPEX

To avoid confusing cash-generating power with the specific year, the investment line has to be unpacked. Additions to fixed assets were $236 million, including $68 million tied to restoration of war damage and about $62 million of direct turnaround costs in the Midan, hydrogen production, MAG 3, and related facilities. This is not normal maintenance CAPEX, but it is also not money that simply vanished. Part of it was the price of getting back to fuller operations.

Still, that is only partial comfort for investors. As long as the steam system has not reached full redundancy and the polymers layer has not returned to positive EBITDA, even one-off CAPEX continues to feel like recurring cash use.

Debt looks manageable, with wide covenant headroom

On capital structure, Bazan looks much less stressed at the end of 2025 than the legacy narrative would suggest. Net financial debt, including the company's share of Cantium, stood at $667 million, and leverage was 1.5x. At the covenant level the picture is even looser: adjusted equity of $1.78 billion versus a required minimum of $750 million, adjusted equity to net assets of 45.1% versus a 20% minimum, net financial debt to adjusted EBITDA of 1.7x versus a 5.0x ceiling, and debt service coverage of 3.2x versus a 1.1x minimum.

That is the core balance-sheet point. There is no current scenario in which Bazan is operating with its back against the covenant wall. There is room. But precision still matters here too: the company states that $123 million received from insurers during the reporting period was included in adjusted EBITDA for covenant calculations. So the headroom is real, but part of it also sits on an unusual insurance year.

The financing layer improved in real terms as well

During 2025 the company took new floating-rate bank loans of $140 million with final maturities in 2033 to 2034. It also extended the tenor of a loan originally taken in 2023, pushing its final maturity to 2034 from 2028, and made an early repayment of a $26 million fixed-rate loan. The weighted average interest rate on bank debt fell to 6.8% from 7.5%.

So debt did not merely stay under control. Its quality improved as well. There is a smaller near-term maturity wall, a longer spread of obligations, and a lower weighted financing cost. That does not remove funding costs, but it does reduce refinancing pressure.

Cantium adds EBITDA above the line, but the cash there is still tied to the asset's own economics

This is the easiest point to smooth over. At the Bazan level, Cantium looks like $39 million of adjusted EBITDA in five months. At the Cantium level, the picture is more complex: $64.5 million of debt at December 31, 2025, $85 million by March 23, 2026, a hedging requirement covering 40% to 50% of monthly PDP production through September 30, 2027, a maximum leverage covenant of 3.0x, a minimum current ratio of 1.0x, and distribution limits.

In addition, Cantium ended the period with a $350.1 million abandonment and restoration obligation, $264 million of surety bonds, and $70.6 million in a dedicated escrow account for that commitment. The value created there may be real, but not all of it is immediately accessible to Bazan.

The company does note that a roughly $100 million project-finance loan signed near report approval could allow a withdrawal of $25 million to $30 million, Bazan's share, through a distribution or shareholder loan repayment. That is an important potential source of flexibility, but it is still subject to general partner decisions, covenant compliance, and a cash-sweep mechanism. It therefore has to be treated as an option, not as money already sitting at the parent.

Outlook

Before getting into 2026, there are four non-obvious findings that need to stay in view.

• First finding: refining is already back, but 2025 still does not represent it cleanly. The $446 million of adjusted EBITDA includes $160 million of insurance income, while also carrying a $275 million estimated lost-profit impact from partial operations and another $56 million from turnarounds. That means both the conservative and the optimistic reading can find support in the report.
• Second finding: the Carmel Olefins impairment test ended without a write-down, but it did not provide a large comfort stamp either. The activity was valued at $652.3 million against a carrying amount of $574.6 million at December 31, 2025. That is a cushion of about $77.7 million, not one of several hundred million.
• Third finding: inside that same valuation work, Carmel Olefins projects negative EBITDA of $24.0 million in 2026, followed only then by $7.3 million in 2027 and $58.4 million in 2028. That does not look like a breakout year. It looks like a transition year with a delayed recovery profile.
• Fourth finding: Cantium may improve Bazan's cash flexibility over time, but for now it still looks more like an asset with its own project-level economics than like a cash box opened to Bazan.

For refining, 2026 looks like an operating proof year

On the positive side, Bazan enters 2026 with a refinery that already proved it can generate EBITDA even after a highly unusual year. The company restored steam facilities that allow full operations with sufficient redundancy, and full redundancy is expected to be completed during 2026. If that happens without another meaningful operational event, the next report will arrive on a cleaner base.

The market backdrop is not bad either. The report was written against strong diesel and gasoline spreads, sanctions on Russian refined products, regional supply disruptions, and Brent moving sharply higher after the balance-sheet date. Bazan does not need wartime peak spreads to stay forever. It only needs the refining environment to remain good enough to hold profitability while polymers try to stabilize.

For polymers, 2026 looks like a difficult transition year

Here the picture is less comfortable. In the Carmel Olefins impairment review, 2026 EBITDA is projected to remain negative because of a sharp rise in butane and propane prices, which feeds into propylene costs, with about 60% of propylene purchased from Bazan and about 40% from Ashdod Refinery. That is exactly the kind of detail a reader may miss if they look only at the fact that no impairment was booked.

In other words, no impairment does not mean 2026 is comfortable. It means the company still sees a path back to profitability, but a path that runs through a very weak year before the recovery only starts to appear from 2027 onward. The sensitivity analysis says something important as well: a $15 per ton reduction in terminal margin or a 0.5% increase in the discount rate still keeps value above book, but not by a large gap. The cushion exists, but it is not generous.

The valuation work also says something between the lines

There is an even more delicate detail. The same exercise assumes that if the government's Haifa Bay redevelopment and petrochemical shutdown plan is implemented, the company would receive full compensation equal to the activity's value, and therefore the operating assumption taken was that the plan would not affect value. That does not mean the plan is not a risk. It means the valuation neutralized it through a compensation assumption.

That is the key analytical point: Carmel Olefins' impairment test is mainly testing margins and operations, and much less the regulatory risk of continuing in Haifa Bay itself. So the fact that no impairment was recorded is not evidence that the strategic risk disappeared. It is evidence that the accounting model did not let that risk determine the outcome.

Cantium is an important option, but 2026 also has to prove value accessibility

Cantium enters 2026 with 55.8 million BOE of 2P reserves, including 42.6 million BOE of proved 1P reserves, and with EBITDA that depends on WTI and on execution of its workover and drilling plans. The investment can be genuinely interesting precisely because it gives Bazan an additional engine that is not dependent only on Haifa Bay.

But for the market to give full credit, it needs to see not only EBITDA but also capital discipline. That means covenant compliance, hedging that protects downside without giving up too much upside, and proof that cash can move upstream to Bazan without undermining the asset's stability. Until then, Cantium is mostly a value layer added to the story, not a cash layer already opened.

2026 therefore looks like a transition year, not a breakout year

The right label for the coming year is a transition year with clear proof points. Bazan already passed the stage of asking whether the refinery can get back to work. It has not yet passed the stage of asking whether the group can translate that recovery into net income, available cash, and polymers that no longer dilute the story. If those three edges line up, the market read on the company can change quickly. If not, 2025 will be remembered as a partial recovery year, not as a reset.

Risks

Insurance and restoration are still not fully closed

The first risk is that the gap between what the company estimates it should receive and what has already been recognized and collected remains open for longer, or closes at a lower number. A total of $58 million of the business interruption claim is still not recognized in the accounts, and the company itself says there is no certainty that this amount will be received in full. At the same time, final direct damage is still estimated in a wide range of $150 million to $200 million.

Polymers are closer to the edge than the "no impairment" headline implies

No impairment was recorded at Carmel Olefins, but Ducor has already been impaired, polymer inventory has already been written down, and Carmel Olefins' 2026 forecast is still negative. So the risk here is not an immediate write-down. It is another year in which refining works while polymers fail to carry the downstream layer, continuing to pressure net income, CAPEX, and the group's strategic read.

Cantium adds diversification, but also project-level risk of its own

Cantium has meaningful customer concentration: 60% of oil and gas revenue in the five months of 2025 came from one customer and 98% from the top two. In addition, debt there is floating-rate, hurricane season can disrupt production, and a significant part of the economics depends on sustaining production and replacing reserves. This is not a risk that immediately threatens Bazan, but it does mean the new asset is still a long way from being a simple cash machine.

Regulatory and environmental risk did not disappear

The group's emission permits for Bazan, Carmel Olefins, and Gadiv are valid through 2031, and the company itself says that compliance with permit requirements, land restoration requirements, and ESG pressure from lenders and customers could require additional investment and weigh on financing. This is not a one-quarter trigger, but it is exactly the kind of risk layer that can keep a discount on a company that, even when refining is strong, still operates in the middle of a highly sensitive environmental and planning area.

Conclusions

Bazan finishes 2025 as a company that looks better operationally than its net income suggests, but less clean than its adjusted EBITDA suggests. Refining is back, covenants are distant, and the balance sheet no longer looks stressed in the old sense. The main bottleneck has shifted to polymers, to the quality of cash generation, and to the question of how much of the new value inside Cantium can actually move upstream in a reasonable timeframe.

Current thesis: the refinery is already back to generating money, but the Bazan thesis only gets materially stronger if 2026 proves that polymers, restoration, and value access from Cantium stop diluting the picture.

What changed versus the earlier read: the center of risk has moved away from an immediate operating shock and leverage anxiety toward the question of whether the downstream layer can carry the refining recovery without consuming the cash it creates.

Counter-thesis: one could argue that the difficult year is already behind the company, that refining has fully recovered in a supportive environment, that polymers will emerge from the trough in 2027, and that Cantium will start returning cash, making 2025 look in hindsight like only a temporary earnings low.

What can change the market read in the short to medium term: the pace of completion of the steam redundancy work in 2026, the development of the gap between the business interruption claim and the amount ultimately recognized, first signs of improvement in polymers EBITDA, and Cantium's ability to generate a distribution or shareholder-loan repayment without compromising funding discipline.

Why this matters: Bazan is no longer being judged on simple survival. It is being judged on the quality of the conversion from refining recovery into cash and into value that is actually accessible to shareholders. This is no longer a question of whether the company can continue. It is a question of what really remains after all the friction layers.

What must happen in the next 2 to 4 quarters: polymers have to stop being EBITDA negative, the energy system has to reach full redundancy, the insurance gap has to close on reasonable terms, and Cantium has to start showing not just EBITDA but a practical route to cash upstreaming. What would weaken the thesis is another operational disruption, another restoration delay, further pressure in polymer margins, or a situation in which Cantium remains trapped behind its own debt and abandonment obligations.