Partner In 2025: Operations Improved, But 2026 Will Test TV, Cash And Capital Discipline

Getting To Know The Company

Partner enters 2026 from a very different place than the one investors usually associate with a local telecom operator. This is no longer a balance-sheet rescue story, and it is not just a one-year rebound either. It is a company with two large operating engines, mobile and fixed-line, and it ends 2025 with real improvement in the core business: postpaid subscribers grew by 52 thousand, fiber subscribers grew by 33 thousand, mobile ARPU excluding interconnect rose to NIS 43, internet ARPU rose to NIS 94, and adjusted EBITDA increased to NIS 1.224 billion.

The easy mistake is to stop at the 4% decline in reported revenue and conclude that the company was basically flat. That is the wrong read. The decline came mainly from the collapse in interconnect revenue, which fell to NIS 72 million from NIS 230 million after the regulatory change, while service revenue excluding interconnect actually rose to NIS 2.562 billion. In other words, the underlying economics improved, but the reported top line looks weaker because of an item that says very little about real pricing power or demand.

What is working now? Mobile is back to cleaner growth in postpaid subscribers, 5G paying subscribers reached 839 thousand, and the fixed-line business kept expanding its fiber base so that about 95% of internet subscribers are already on fiber. What is still not clean? Fixed-line has entered a new phase with the launch of Partner TV+, a more complex cost structure, heavier regulatory dependence in wholesale broadband, and an open question over whether the commercial upgrade will really translate into steadier profitability rather than just a better customer proposition.

Partner's active bottleneck is no longer debt. Net financial debt fell to NIS 128 million, net debt to adjusted EBITDA stood at 0.1 at year-end, equity reached NIS 2.322 billion, and S&P Maalot reaffirmed the ilAA- rating with a stable outlook in May 2025. Precisely because the balance sheet is no longer forcing the company into a corner, the story has become tougher in a different way. The test has moved to capital allocation. The NIS 465 million dividend approved in March 2026, and even more so the board's decision to examine an additional distribution of up to NIS 500 million not out of profits, shift the central question from whether the business can improve to how management chooses to use excess liquidity.

Partner's economic map is straightforward:

At the end of 2025 Partner employed 2,346 people, down from 2,405 a year earlier, which implies annual revenue per employee of about NIS 1.36 million. The company's main operations and assets are in Israel, so this is less a geographic diversification story and more a debate about network quality, regulation, and capital-allocation discipline in the local market.

Events And Triggers

TV+ is a commercial upgrade, but it is also a real expense

The most important fixed-line event of the year was the launch of Partner TV+ in September 2025 under the agreement with yes. During 2025 the migration of customers to the new service was completed, and the company ended up with a richer product, including a dedicated interface, dozens of linear channels, VOD, Catch Up, and exclusive content. From a commercial perspective, the move makes sense: the TV subscriber base, which had declined during the year, returned to growth in the fourth quarter and rose by 6 thousand subscribers versus the previous quarter.

But the offset matters just as much. The agreement did not only improve the product. It also introduced a new expense base. The company's annual expense under the agreement will not be lower than about NIS 64 million plus VAT, and in 2025 it already reduced operating profit by roughly NIS 17 million through accelerated depreciation, employee retirement costs, and one-off content expenses. That is exactly what the market will need to test in 2026: whether TV+ is mainly a retention tool, or whether it can also create better economics for the fixed-line business.

The government tender and 5G rollout support the mobile side

On the mobile side, Partner enters 2026 with two supportive developments. The first is its selection as the second awarded supplier in the government cellular tender, where it will provide 40% of the services and equipment, with an estimated scope of at least 40 thousand subscribers plus tens of thousands of data subscribers, for an initial 42-month term with options that can extend the total term to 114 months. The company did not quantify the profit contribution, so there is no reason to overstate it, but it is clearly an important commercial asset in the enterprise and institutional segment.

The second is continued 5G rollout. The company already covers about 83% of Israel's population with 5G and added 260 thousand paying 5G subscribers during 2025. This is not just a marketing datapoint. It indicates that Partner is still moving part of its base toward higher-value packages and heavier data usage.

Shutting down 2G and 3G is economically right, but operationally noisy

The shutdown of the 2G and 3G networks should free spectrum and improve network efficiency. Economically, that is the right move. But it also carries near-term friction: as of the report date, around 87 thousand customers still needed a device replacement or settings update to continue receiving full service after the shutdown. If that base is not migrated quickly enough, the company could face higher customer-service load, some deterioration in service experience, and possibly a degree of subscriber erosion. In other words, an operationally rational move can still create commercial noise in the short term.

The Bezeq IRU expands reach today, but reduces some flexibility later

The agreement with Bezeq to acquire an indefeasible right of use, or IRU, in connected fiber lines is strategically important. By the end of 2025 Partner had already exercised the option for 36 thousand additional lines out of 48 thousand, and during 2025 it paid about NIS 140 million under the agreement. The outstanding IRU principal stood at about NIS 461 million at year-end. The benefit is easy to see: broader reach, faster access to households, and less dependence on slower organic buildout.

But this is only half the story. The IRU increases deferred infrastructure costs and amortization, and it now connects directly to the Ministry of Communications' wholesale-market decisions. Under the ministry's February 2026 decisions, the lines included in the Bezeq IRU will be deducted from Partner's regulated BSA allocation, and the company already holds BSA line volumes above what the formula would otherwise allow because of that agreement. The practical result is that any additional Bezeq purchases from September 2026 will require ministry approval. The company says this is not expected to be material, but analytically it does change the reading: the IRU widens Partner's fiber reach while also reducing part of its future procurement flexibility within the regulated path.

Efficiency, Profitability And Competition

Partner's 2025 operating story is a real improvement story, but not a symmetrical one. Mobile improved commercially faster than it improved operating leverage. Fixed-line looked better in the annual view than it did in the fourth quarter. And at the consolidated level the company benefited both from better business momentum and from a few supporting items that are not necessarily a durable 2026 base.

Mobile delivered commercial growth, but only limited operating leverage

In mobile, service revenue excluding interconnect rose 3% to NIS 1.373 billion. That came from higher revenue from mobile packages and roaming, while the postpaid base rose to 2.514 million and prepaid subscribers declined slightly to 164 thousand. In other words, the company did not grow through weaker customers. It grew through a relatively higher-quality recurring base.

Mobile ARPU excluding interconnect also moved in the right direction, to NIS 43 from NIS 42 in 2024. That is not dramatic, but it matters. It suggests the company is not visibly buying growth through headline price erosion. At the same time, equipment revenue rose to NIS 527 million, mainly because of a higher average selling price per device and, in the fourth quarter, also because of the timing of flagship-device launches.

Even so, mobile adjusted EBITDA fell 2% to NIS 716 million, and operating profit rose only 1% to NIS 329 million. This is the heart of the story. Revenue trends improved, but the flow-through into profit was constrained mainly by higher spectrum-fee expense after the temporary discount ended in September 2024. Partner can grow, but at this stage it still is not converting every commercial improvement into parallel margin expansion.

Fixed-line improved, but the improvement is less clean than the headline

The most interesting part of the report sits in fixed-line. On one hand, adjusted EBITDA rose 11% to NIS 508 million and operating profit jumped to NIS 104 million from NIS 75 million in 2024. On the other hand, service revenue excluding interconnect actually fell 1% to NIS 1.272 billion.

How can both be true? First, the revenue base needs to be unpacked. The decline was partly driven by the absence of a prior-year one-off fiber lease deal with a business customer, and by lower revenue from paid works and Hubbing. Those were partly offset by growth in internet and business-data services. So the underlying fixed-line core improved, but the reported comparison was affected by a higher base.

Second, 2025 also benefited from the closure of the VAT provision related to Netflix customers and from lower wage-related expenses. Those items support profitability, but they do not by themselves prove a permanently better business quality. That is why a flat reading of the operating-profit jump would be too generous.

The third point, and probably the most important one, is the fourth quarter. Fixed-line operating profit fell to only NIS 13 million, from NIS 25 million in the comparable quarter, and adjusted EBITDA declined to NIS 115 million from NIS 122 million. The company explains this by one-off content costs tied to the yes agreement, higher system-maintenance expenses, and lower profit from paid works. That is a strong reminder that the move to TV+ may have improved the product, but in the short term it is also pressing on the cost base.

At the same time, the structural quality of the fixed-line base does look better. Internet subscribers rose to 492 thousand, fiber subscribers rose to 468 thousand, and internet ARPU increased to NIS 94. With about 95% of the internet base already on fiber, Partner is building a better fixed-line business than it had before. The open question is whether the translation into profitability can become consistent, or whether TV+ and the new infrastructure agreements will make the segment heavier before they make it more profitable.

The consolidated report is better than the headline, and worse than the headline

That sounds contradictory, but both statements are true. The report is better than the headline because the revenue decline is distorted by interconnect. It is also worse than the headline because part of the profit improvement leaned on provision closure, on comparison against a prior-year one-off gain, and on cost timing between periods.

That is why the right way to read Partner today is neither as a company in distress nor as one that has already cleaned up the story. It is a telecom operator delivering solid operating execution, but one that still needs to prove that this improvement rests on a stable enough base to carry a better TV product, higher spectrum costs, and larger capital distributions at the same time.

Cash Flow, Debt And Capital Structure

Cash flow is the section where it is easiest to read only the headline and miss the thing that matters most. The headline says adjusted free cash flow rose 35% to NIS 386 million. That is true. But once the full cash picture is rebuilt, the story becomes more complicated.

Adjusted free cash flow improved, but all-in cash flexibility is much less generous

Cash flow from operations increased to NIS 998 million from NIS 971 million in 2024. CAPEX declined to NIS 454 million from NIS 527 million, and that reduction together with working-capital support pushed adjusted free cash flow higher. Even after financial-debt interest, the company still shows adjusted free cash flow of NIS 354 million.

But that is not the same thing as cash that truly remains available. If the all-in cash flexibility picture is rebuilt, meaning how much cash is left after actual cash uses during the year, the result looks different: NIS 998 million from operations, less NIS 454 million of CAPEX, less NIS 158 million of lease cash principal and interest, less NIS 32 million of bond and bank interest, less NIS 155 million of financial-debt principal repayment, and less NIS 250 million of the dividend paid in March 2025. The result is about negative NIS 51 million.

That does not mean the company has a liquidity problem. It clearly does not. It does mean that 2025 capital returns were not financed only by the cash created during the year, but also by the existing cushion. That distinction matters, because it is exactly where the 2026 debate is heading.

Even the liquidity cushion is less flashy than it first appears

Another easy point to miss is that cash and cash equivalents rose to NIS 662 million from NIS 481 million. At first glance that looks like a healthy increase in the cash pile. But in 2024 the company also held NIS 211 million of short-term deposits, which were released during 2025. On a combined cash-plus-short-term-deposits basis, liquidity actually declined from NIS 692 million to NIS 662 million.

This is not a dramatic deterioration, but it tells a more accurate story: Partner did not end the year with a fatter cash position. It ended the year with slightly lower liquid resources while paying a dividend, reducing debt, and continuing to invest. That is still a strong position. It is simply less generous than the NIS 662 million cash line suggests on its own.

Debt is no longer the problem, but fixed obligations are still real

At the end of 2025 the company had NIS 670 million of bonds, NIS 120 million of bank loans, and NIS 558 million of lease liabilities. The financial-debt mix itself is relatively comfortable: Bond Series Z carries a fixed 4% coupon, Bond Series H carries a fixed 2.08% coupon, and the bank loan carries a fixed 2.55% rate. That matters because it limits the direct hit from higher rates on existing debt.

Covenant headroom is very wide. Net debt to adjusted EBITDA at 0.1 is far below the ceiling of 5, and equity of NIS 2.322 billion is far above the NIS 600 million to NIS 700 million thresholds. That is why the discussion today is not about debt service capacity. It has returned to the question of whether it makes sense to pull cash out of the company this quickly.

Still, the absence of financial stress should not be confused with the absence of fixed commitments. Alongside financial debt, Partner also carries a meaningful lease-payment stream, the Bezeq IRU with NIS 461 million of outstanding principal, and 2026 CAPEX guidance of NIS 450 million to NIS 500 million, basically close to the 2025 level. This is not a light infrastructure company. It is simply a more stable one.

Outlook

Before getting into the details, five non-obvious findings from the evidence set need to be stated clearly:

• The operating engine improved more than the top-line headline suggests. Service revenue excluding interconnect rose, both in mobile and in the core of fixed-line.
• The fixed-line improvement is real, but not clean. It benefited from provision closure and comparison effects, while the fourth quarter already showed what TV+ costs.
• Adjusted free cash flow is not the end of the story. Once all actual cash uses are included, 2025 looks materially less generous.
• 2026 does not get relief from CAPEX. Guidance of NIS 450 million to NIS 500 million means the next step up has to come from better conversion of revenue into margin and cash.
• Once debt stopped being the pressure point, capital allocation became the risk. That is the key shift in how Partner should now be read.

2026 looks like a proof year, not a breakout year

The central message from the report is that 2026 does not look like a breakout year. It looks like a proof year. Partner no longer needs to prove that it can keep the business stable. It needs to prove that the combination of fiber, TV, 5G, and large distributions still produces clean economics.

On the mobile side, the company needs to keep lifting ARPU and adding postpaid subscribers without giving up too much profitability in exchange for that growth. That matters even more because spectrum-fee expense has already stepped up, and that headwind is not about to disappear. In December 2025 the company even prepaid roughly NIS 100 million of spectrum fees for 2026, so the market will want to see in the coming quarters that this does not overly hurt cash-flow visibility.

On the fixed-line side, the test is sharper. Fiber growth and internet ARPU growth are good, but they are already known. What is still missing is proof that the new TV product does not merely reduce churn and improve the commercial bundle, but also avoids creating too heavy a cost base. If fixed-line profitability recovers from the fourth-quarter 2025 dip in the first two quarters of 2026, the market will likely treat TV+ as a value-building move. If content, maintenance, and bundle costs remain heavy, the thesis will look less clean.

The IRU and regulation will keep running under the surface

There is a deeper layer here that the market could easily miss at first glance. The Bezeq IRU gave Partner a real strategic acceleration in fiber. But once the ministry's wholesale decisions are folded in, the IRU is no longer only a growth engine. It also helps determine how much room remains within the regulated BSA path.

In the short term that may not hurt. In the medium term it does change the company's room to maneuver. If Partner needs more flexibility in access to Bezeq lines, it can no longer simply assume that the regulated BSA channel remains open in the same way. This is not a 2025 problem. It is a strategic 2026-and-beyond question.

The capital-markets debate will focus on the dividend, but that is not the only event

The issue most likely to attract immediate attention is the NIS 465 million dividend, and even more so the possibility of an additional distribution of up to NIS 500 million not out of profits. That is understandable. It is the easiest headline to read. But anyone who stops there will miss the more important point: Partner has already proved that it has a strong balance sheet. What now needs to be tested is whether it makes sense to pull capital out at that pace while fixed-line economics are still in proof mode.

That means the market could change its interpretation quite quickly, not only if the company beats operating expectations, but also if it signals that it will keep disciplined capital allocation. In a company with leverage this low, management signaling matters almost as much as the reported number itself.

Risks

The first risk is no longer financial. It is behavioral

Partner's main near-term risk is not a credit squeeze or a covenant issue. It is the possible tendency to treat the liquidity cushion as excess cash that can be extracted too easily. A NIS 465 million distribution is possible within the current balance sheet. Even examining an additional distribution does not by itself prove irresponsibility. But if the company pushes too far, the story will shift from capital policy that respects infrastructure stability to capital policy that assumes operations will keep rebuilding cash at a high enough pace without a comfortable buffer.

The second risk is regulation and dependence on the Ministry of Communications

Partner operates in a sector where the regulator can change business economics without changing the number of customers. That already happened in interconnect, and it also shows up in spectrum fees, which reached NIS 99 million in 2025. It shows up again in wholesale broadband, and in spectrum sharing with Palestinian operators, which from the second half of 2026 could impair 2600MHz service in parts of Judea and Samaria and within up to 8 kilometers of the Green Line. The company is trying to limit the impact, but this is not an area where it has full control.

The third risk is fixed-line execution

As fiber continues to grow, the market will be prepared to give the fixed-line business more credit. But that credit will remain limited if TV keeps requiring one-off expenses, if bundles hit ARPU more deeply, or if growth in internet and business data is not enough to offset weakness in other activities such as Hubbing and paid works.

The fourth risk is external shocks and roaming revenue

The company estimated that the 2025 pre-tax hit to roaming profit versus the pre-war period was about NIS 25 million. That is a reminder that even a domestic and apparently steady company is still exposed to security conditions and to the aviation market. If regional tension remains high, roaming revenue could stay below a more normal level.

The fifth risk is tax

At the end of 2025 the company received best-judgment tax assessments for 2020 through 2023 that imply additional tax of about NIS 185 million including linkage and interest. It paid NIS 45 million on account and filed an objection in March 2026. The company disputes the tax authority's position, and at this stage the final outcome cannot be assessed. This is not an existential threat, but it is absolutely an item that can create cash-flow and accounting noise if it develops in an unfavorable direction.

Conclusions

Partner ends 2025 as a better telecom company than it was a year earlier. Mobile is back to growth, the fiber base kept strengthening, adjusted free cash flow improved, and the balance sheet is far from stressed. Precisely for that reason, the story will no longer be decided by a survival question. It will be decided by a quality question: whether Partner uses its excess stability to build a stronger fixed-line business, or to extract capital too quickly before that move has really been proved.

Current thesis: Partner enters 2026 with a stronger core business, but the test has shifted from whether operations are improving to whether TV+, the Bezeq IRU, and large capital distributions can coexist without eroding financial discipline.

What changed from the prior reading: A year ago Partner could still be read mainly through the balance sheet and the recovery story. Today the balance sheet is no longer the constraint, which means the more important risk has moved to capital allocation and to whether the new fixed-line setup is profitable enough.

Counter-thesis: One could argue that the concern is overstated because net debt is very low, the rating is stable, fiber and ARPU keep growing, and the company is simply returning part of its excess cash to shareholders.

What could change the market's interpretation in the short to medium term: Two consecutive quarterly reports showing fixed-line margin recovery together with disciplined capital returns would improve the read materially. On the other hand, if TV costs stay heavy and the board pushes for another aggressive distribution, the market may start to treat 2025 as a comfortable peak year rather than as a stable base.

Why this matters: Partner is no longer trading against a survival question. It is trading against the quality of allocation between network, content, and dividends. That is a material difference, because it affects both the quality of future earnings and the credibility of the 2025 improvement.

The hurdle for the next 2 to 4 quarters is fairly clear. Partner needs to show that growth in fiber, ARPU, and TV can translate back into fixed-line profitability even without one-offs, that cash generation remains positive against NIS 450 million to NIS 500 million of CAPEX, and that the cash cushion is not being depleted faster than the business can rebuild it. If those three conditions hold, 2025 will look like a solid new base year. If they do not, this report will be remembered mainly as the moment when the balance sheet improved before discipline was truly tested.