Bezeq: Structural Separation, the yes Merger, and the 2026 Deadline

What Is Actually Stuck

The main article argued that Bezeq has already passed the operating test, but not yet the value-capture test. This follow-up isolates the layer that keeps that gap open: structural separation. This is not a side note in the legal section. It is the main bottleneck between stronger execution inside the subsidiaries and a telecom group that can actually behave like one group.

The annual filing states the problem plainly. Bezeq's license requires structural separation between the company and its subsidiaries. In practice, that means full separation between management teams, including the business, finance, and marketing systems, and a ban on transferring commercial information to a subsidiary, subject to limited exceptions. Bezeq also says explicitly that these restrictions put the group at a competitive disadvantage versus other telecom groups and create higher operating overhead.

That is the core of the story, because the issue is no longer whether yes can improve. It already has. The issue is whether the group can turn that improvement into full group value: broader bundle marketing, wider use of shared systems, cleaner operating synergies, and ultimately the ability to merge yes into Bezeq under a known tax path.

The product layer shows the asymmetry clearly. yes is allowed to sell communications bundles that include internet and TV without having to sell the TV service at a uniform standalone price, following the April 2021 amendment to the merger conditions. But the same annual filing says the Ministry of Communications has rejected various Bezeq requests in recent years to sell Bezeq internet together with yes TV services, including OTT. Some relief exists, but it is still far from a structure under which the group can market itself the way the competing telecom groups do.

Another important point is that the regulatory review is not binary. In the October 2025 public consultation, the Ministry of Communications put three policy models on the table: full cancellation of structural separation, functional separation, or separation of ownership over the passive infrastructure. Bezeq, for its part, submitted a response in November 2025 arguing for option A and rejecting the other two. So not every headline about "progress on structural separation" is equal. A partial model could still leave a large part of the friction in place.

The 2026 Hourglass

The December 2025 tax-ruling extension gives Bezeq another year, but it also imposes a concrete clock. The Tax Authority extended the ruling for the merger of yes into Bezeq until December 31, 2026. In the same letter, it repeated the wording used in prior years: it may decide not to extend the ruling beyond that date if there are no material developments during 2026 regarding the cancellation of structural separation.

This is where the nuance matters. That is a deadline for the current tax ruling, not necessarily for the merger itself. Bezeq states explicitly that even if the ruling is not extended, the company could still request a new tax ruling in the future. So December 31, 2026 is not a cliff for the merger concept. It is a clock on certainty, on the current tax path, and on the ability to close the move without reopening a fresh layer of uncertainty.

That also explains why the 2025 delay matters. In its April 8, 2025 work plan, the Ministry of Communications defined the structural-separation review as a key task and set publication of a decision by the end of 2025 as a success metric. In practice, the ministry was still sending data requests to telecom companies in June 2025, and by October 2025 it had only moved to a renewed public consultation. In other words, 2026 began after the regulatory timetable had already slipped.

The point of this chart is precisely that it is not dramatic. Bezeq guides to comparable EBITDA of NIS 3.7 billion to NIS 3.8 billion in 2026, almost identical to the NIS 3.74 billion reported for 2025, and a medium-term midpoint of about NIS 4.3 billion in 2029. Those are the base numbers. In the same filing, the company says the group targets do not include the effects of canceling structural separation, mergers and acquisitions, or accounting value changes. In other words, regulatory relief is not required for the business to keep running. It is required to unlock the layer of value above the base case.

Why This Is Not Only About the Ministry of Communications

It is tempting to read this as a story that runs only through a regulatory decision. That misses a second layer: labor and organizational execution. Bezeq itself says that past synergy programs between Pelephone and yes were implemented through collective agreements that included efficiency and synergy processes. It also says the companies still operate separately, even if they share a CEO and certain management layers, while continuing to use shared resources.

That matters because moving from structural separation on paper to a group that actually operates differently will not happen only through a signature from the regulator. It will also require organizational adjustments, shared functions, shared systems, and in some cases workforce arrangements. Bezeq does not quantify what a future yes merger would cost or how long it would take from a labor perspective. But it does show that the group continues to move meaningful efficiency layers through negotiated agreements rather than through spreadsheets alone.

Chairman Tomer Raved wrote in the opening letter that, during the last month, the group had signed collective agreements intended to protect employee rights and group stability in line with its multi-year plans. The investor presentation repeats the point in a more focused way: a new collective agreement at Bezeq Fixed-Line and principal understandings at Pelephone. A few weeks before the annual report, on February 25, 2026, Pelephone said its board had approved understandings with employee representatives that include the retirement of up to 150 permanent employees and a three-year renewal of the collective agreement.

That does not prove that a future yes merger will be blocked by labor. That would go too far. But it does prove that the group still runs efficiency and organizational change through negotiated labor channels. So even if the ministry advances, value unlock would still not be immediate after a regulatory step. There is a real execution layer here.

The 2026 clock, then, is not only regulatory. It is a combined clock. The Ministry of Communications needs to produce material progress, the Tax Authority needs to view that progress as sufficient to preserve the existing tax path, and the group itself will have to show that it can convert such a move into a workable operating structure rather than a theoretical win.

What Management Already Underwrites, and What Remains Optional

One of the strongest insights in the filing is that management is not building the base plan on top of structural-separation relief. That is true both for 2026 and for the 2029 targets. On one hand, the group guides to comparable EBITDA of NIS 3.7 billion to NIS 3.8 billion in 2026 and sets clear medium-term targets for 2029. On the other hand, it says explicitly that the targets exclude the effect of canceling structural separation, mergers and acquisitions, and accounting value changes.

That distinction matters because it separates two very different layers:

• The base layer: fiber keeps penetrating, yes keeps migrating to IP, Pelephone keeps improving, and the group keeps returning capital.
• The optionality layer: structural-separation relief, the yes merger, broader operating synergies, and a cleaner tax path.

In plain terms, Bezeq is telling investors two things at once. First, the business can keep improving even without a regulatory breakthrough. Second, if that breakthrough comes, it arrives on top of a base case that already stands on its own. That is exactly why structural separation is a value-capture bottleneck rather than a condition for the thesis to survive at all.

That point is especially clear in yes. In the 2025 investor presentation, the group shows yes ARPU of NIS 192 and a 2029 target of about NIS 215. It also shows yes TV + fiber subscribers at about 21% of total subscribers in 2025, with a target of about 50% by 2029. That means Bezeq is already underwriting meaningful improvement in the yes product, pricing, and fiber combination inside the base case, before structural-separation relief even enters the model.

So the story is not "either regulation gets solved or yes remains a problem." That would be too simplistic. yes is already on an operating-improvement path without the merger. What the merger and structural-separation relief are supposed to do is not invent the improvement, but convert it into fuller group value: broader bundle flexibility, more system sharing, more operating synergies, and probably a cleaner tax outcome.

That is also why the market has to separate operating improvement from value unlock. If yes keeps improving ARPU, bundle penetration, and its migration to IP, that can materially strengthen the base case. But as long as structural separation remains in place, part of that value stays trapped between the entities and does not automatically become fully accessible group value.

Conclusion

Structural separation is Bezeq's main bottleneck not because the business is weak, but because the business is now strong enough for the bottleneck to stand out. Once operations improve, the question moves to the next layer: can the group sell, manage, and merge like one telecom group, or will it keep operating through separation walls that create cost, delay, and competitive asymmetry.

The extension of the tax ruling through December 31, 2026 makes the story more urgent, but not in the simplistic sense. It is not the end of the road for a yes merger if that date passes. Bezeq itself says it could request a new tax ruling. But it is the end of the current tax path, which makes it a deadline on certainty and timing. The more 2026 passes without material progress, the more evidence the market gets that the gap between operating improvement and real value capture remains open.

Current thesis: 2026 is a proof year for value unlock, not for operating survival. Bezeq has already shown it can improve yes without regulatory relief. What is still missing is the move from subsidiary-level progress to full group-level value capture.

The real point: if the ministry moves toward full cancellation of structural separation, and if that progress is enough to preserve the existing tax path, the upside arrives on top of a business base that is already working. If not, Bezeq can still keep improving, but it will do so with higher overhead, weaker bundle flexibility, and another round of uncertainty around the yes merger.