Summary

  • INTER-SVYAZ-GROUP's economic test is not whether a regional fibre network can charge high headline prices. It can. Its 2022 public tariff table lists cottage broadband at 1,530 to 6,612 rubles per month, apartment broadband at 660 to 1,020 rubles per month, and a separate Moscow Silk business-quarter table that reaches 90,000 rubles per month for 1 Gbit/s service. The harder test is whether enough connections sit on already-built routes for those prices to survive installation, support, upstream transit, peering, equipment replacement and customer churn.
  • The public evidence shows a real but small operator: a corporate identity tied to Moscow legal registration and Barvikha customer service, a predecessor dating to 2003, FTTH and GPON positioning, two visible IPv4 origin prefixes totaling 1,536 addresses, RIPE LIR registration, one clearly declared PeeringDB MSK-IX connection at 1 Gbit/s, and 2025 financial disclosures in business registries around 47.1 million rubles of revenue and about 17.0 million rubles of net profit. It does not show subscriber count, route mileage, renewal backlog, customer concentration, active licence detail beyond contradictory aggregator summaries, or the share of revenue coming from cottages, apartments, business connectivity, telephony, security systems and integration work.

Start with one incremental connection on an existing route. That is the cleanest unit for this company because almost every public fact about INTER-SVYAZ-GROUP points to the same economic boundary: fibre in a few affluent and semi-rural settlements west of Moscow, not a national access platform. If the next customer is near a distribution point, the operator can collect monthly revenue while spending relatively little new trenching capital.

If the next customer requires a new drop, optical terminal, field visit, splice, documentation, support relationship and future repair obligation, the same tariff has to repay a much heavier cost stack before it becomes retained margin.

The company itself describes the route boundary in unusually concrete terms for a small provider. It says the predecessor ZAO INTER-SVYAZ-GROUP had operated in telecoms since 2003 and was converted into LLC form in September 2016. It says the business was built around FTTH, that it built its own fibre-optic networks and communication nodes, and that these networks served residential and business customers in Razdory, Barvikha, Zhukovka, Usovo, Gorki-2, Znamenskoye, Sareevo, Soloslovo and other settlements in the Odintsovo district of the Moscow region. That is not a broad urban footprint.

It is a local-network statement around a particular geography where households can be valuable but route density can be uneven.

That geography matters because the revenue side is published in two different shapes. The first is the residential and small-business internet table that entered force on January 1, 2022. For cottages, the table lists a 4,800-ruble connection fee and monthly prices of 1,530 rubles for 10 Mbit/s, 3,552 rubles for 30 Mbit/s, 4,584 rubles for 50 Mbit/s and 6,612 rubles for 100 Mbit/s. For apartments, the table is much lower: up to 50 Mbit/s at 660 rubles per month and up to 100 Mbit/s at 1,020 rubles per month.

The same table says GPON subscriber equipment and optical cable entry works for cottages are not included in the connection price, and it warns that xDSL-connected cottages may not exceed 30 Mbit/s.

The second shape is the Moscow Silk business-quarter table, also effective from January 1, 2022. There the monthly prices are in a different league: 7,270 rubles for 10 Mbit/s, 10,800 rubles for 20 Mbit/s, 14,000 rubles for 30 Mbit/s, 19,300 rubles for 50 Mbit/s, 25,000 rubles for 100 Mbit/s, 32,000 rubles for 200 Mbit/s, 38,000 rubles for 300 Mbit/s, 49,000 rubles for 500 Mbit/s and 90,000 rubles for 1 Gbit/s. The connection price is 2,400 rubles, and the page separately lists monthly fees for small routed IPv4 blocks.

This table is critical because it shows why a small operator might still produce meaningful profit: one business customer on an already-served building can contribute as much monthly revenue as several apartment customers or several lower-tier cottage customers.

The contribution calculation therefore begins with route reuse. If a cottage 100 Mbit/s customer pays 6,612 rubles per month and does not require material new civil works beyond the customer-paid or separately billed items, the gross revenue is 79,344 rubles per year before VAT treatment, support, upstream traffic, power, office labour, billing, routing, optical maintenance and capital renewal. If that same customer requires several visits and bespoke fibre work, the first-year economics can be much weaker even before churn risk.

The tariff notice explicitly excludes GPON optical terminal equipment and optical cable entry works from the included connection price, which is an important disclosure: management does not appear to be absorbing every premise-specific cost into a flat installation fee.

The apartment customer is a different business. A 100 Mbit/s apartment plan at 1,020 rubles per month produces 12,240 rubles per year. That is much closer to national-operator price anchors and much less able to carry long truck rolls, custom field labour or underutilised access electronics. It only makes sense if the building density is good, the drop cost is small, and support load is contained. The tariff table's distinction between cottage and apartment pricing is therefore not cosmetic.

It tells readers that INTER-SVYAZ-GROUP is trying to price the physical difficulty of the last mile, not simply selling a generic "100 Mbit/s" commodity.

The business-quarter economics are different again. A 100 Mbit/s Moscow Silk line at 25,000 rubles per month produces 300,000 rubles per year. A 1 Gbit/s line at 90,000 rubles per month produces 1.08 million rubles per year. Those numbers can fund redundancy and higher support expectations if utilisation is healthy and if the operator already has presence in the facility. They can also exaggerate apparent pricing power if the table covers a narrow building context rather than a repeatable addressable market. The public record does not show how many customers actually buy these plans or whether the table is still transacting at those rates.

It is usable evidence of price architecture, not proof of sales mix.

The network evidence supports the idea that this is a real autonomous network, not only a reseller brand. RIPE records AS44772 as INTSG-AS and link it to LLC INTER-SVYAZ-GROUP. The RIPE organisation object gives the Russian registration number matching the company's OGRN, lists the organisation type as LIR, and places the registered address at Novoslobodskaya Street 67/69, suite VIII, room 5, Moscow. The aut-num object shows AS44772 was created in March 2008 and was last modified in October 2024.

It records upstream import policies from INETCOM CARRIER and DK Svyaz and route-server or peer policies involving MSK-IX, PITER-IX, Hurricane Electric and MACOMNET. That does not tell us traffic volume, but it does define an operator with its own routing policy.

RIPEstat's routing-status endpoint is more current and more quantitative. For AS44772, it reports two visible IPv4 prefixes and 1,536 IPv4 addresses, with 327 of 327 RIS IPv4 peers seeing the route at the query time. It reports zero visible IPv6 prefixes and zero visible IPv6 /48 equivalents. RIPEstat's announced-prefixes endpoint also shows the two visible prefixes, 185.129.56.0/22 and 195.28.14.0/23, over the recent measurement window. That gives the clearest public address-footprint calculation: 1,536 IPv4 addresses, not millions of endpoints, and no presently visible IPv6 origin in the RIPEstat view.

Several public ASN databases agree on the two IPv4 blocks and the lack of visible IPv6 origination, though not every third-party page is internally consistent. BGP.tools lists AS44772 as active under RIPE, with two IPv4 prefixes and no IPv6 prefixes, and names INETCOM CARRIER, DK Svyaz and BiMajLink as upstreams. IPinfo similarly identifies two IPv4 ranges totaling 1,536 addresses and classifies the ASN as an ISP, while showing no IPv6 addresses in its summary. IPLocate and IPIP also report 1,536 IPv4 addresses and zero IPv6.

IP2Location, however, shows the two IPv4 blocks and also lists an IPv6 block that conflicts with the RIPEstat and other public summaries. The conservative conclusion is not that IPv6 is impossible, but that the visible operational evidence used for this analysis should treat AS44772 as an IPv4-heavy small network unless a current route announcement proves otherwise.

Peering evidence gives the operator a partial cost-control mechanism. PeeringDB describes INTER SVYAZ GROUP as regional, with an open peering policy, self-disclosed traffic level of 5-10 Gbit/s, and one listed exchange connection. The API detail for that connection shows MSK-IX Moscow, speed 1,000 Mbit/s, IPv4 address 195.208.210.21, operational status, route-server peer status and no IPv6 address in the PeeringDB record. The company's own services page says internet connectivity is provided both at MSK-IX sites and directly at INTER-SVYAZ-GROUP nodes, which it says helps reserve transmission channels.

Together, these facts show a wholesale-access strategy: keep some local or exchanged traffic away from pure transit while maintaining upstream paths for the rest.

The same services page names a 10 Gbit/s data network capacity between access nodes and says the access technologies are GPON, Gigabit Ethernet and Fast Ethernet. It lists Cisco Systems, Ericsson and Eltex for routing and switching, and Calix, Ericsson and Alcatel-Lucent for GPON. Those vendor names are useful because they show a mixed equipment estate, not a single modern commodity access box. They also frame renewal risk.

A small operator that has used xPON equipment since 2005 may have strong operating experience, but it must eventually replace optical terminals, OLT cards, batteries, routers, switches, customer-premise devices and test equipment. The public statements do not disclose the age, model mix or replacement reserve.

That is why the revenue-to-address calculation is only a rough sanity check. If 2025 revenue was about 47.1 million rubles, as several business registries report from official filings, then annual revenue per visible IPv4 address is roughly 30,664 rubles, or about 2,555 rubles per month. This is not ARPU. Consumer ISPs can put many customers behind NAT, business customers can buy static subnets, and some addresses may be network infrastructure. But the calculation is still useful because it says the company is not monetising a huge visible address estate.

Revenue is likely coming from a bounded local access and integration business layered on a small number-resource base.

The company-level financials are better than the operating disclosure. RBC's company page reports 2024 revenue of 42.054 million rubles, profit of 17.033 million rubles and average headcount of 13. T-Bank, Companium and Xfirm show 2025 revenue around 47.1 million rubles and profit around 17.0 million rubles, with 2025 headcount shown as 9 by some registries. If those figures are accurate, the net margin is high for a small fixed-network operator: about 40.5 percent in 2024 and about 36.1 percent in 2025. That does not prove broadband access alone is that profitable.

It may include integration work, low depreciation in simplified reporting, timing effects, owner-controlled costs, a mature network with limited current capex, or business service revenue. It does show that the public filing story is not one of a distressed micro-operator losing cash.

The same numbers also show why renewal is the decisive question. A mature fibre operator can report healthy profit in years when the network is mostly built and incremental work is light. That profit must be judged against future replacement obligations, not only current labour and transit. If old GPON equipment, business routers, customer terminals, batteries or ducts require a renewal cycle, reported profit can disappear quickly unless prices hold or business customers carry the load. The public filings visible through aggregators do not give cash-flow detail, depreciation schedules, capital expenditure, route kilometres or debt.

A reader should resist the temptation to treat profit margin as free cash flow.

Control boundary is also visible but thin. The legal entity is LLC INTER-SVYAZ-GROUP, registered in 2016 after conversion from the predecessor ZAO. The company website lists INN 7707372454, KPP 770701001 and OGRN 1167746837906. The legal address is in Moscow; the subscriber service office is in Barvikha, in the Odintsovo district. Business registries name Alexander Tsyperson as general director and identify Ilya Yushvaev and Galina Klinovskaya as current shareholders with 64 percent and 36 percent stakes respectively.

Those public registry facts draw a boundary around the company, but they do not show operational decision rights, customer contracts, related-party leases, network ownership by route, or whether any passive infrastructure is controlled by affiliated real-estate entities.

The predecessor history is economically meaningful. The company says the original ZAO had worked in telecoms since 2003, that xPON equipment had been used since 2005, and that the business was among early Russian adopters of GPON and blown-fibre construction. It also says experience from telecom networks was used to expand into video surveillance, structured cabling, corporate and private data networks, telephony, seamless WiFi and DECT systems for offices and private households. This points to a hybrid model: recurring access revenue plus project and support work around the same customer base.

For a small regional ISP, that hybrid can be attractive because each customer relationship can produce installation, integration and maintenance income beyond monthly internet service. It can also create lumpy revenue and labour demand.

The telephony evidence reinforces the hybrid model. The services page says telephone service uses numbering resource allocated by a Federal Communications Agency decision dated February 7, 2013. A public telephone-operator database lists INTER-SVYAZ-GROUP with a total numbering capacity of 1,100 subscribers. Again, this is not proof of active lines. It is proof that the company has regulatory and numbering apparatus beyond a simple broadband-only resale proposition.

In practice, local telephony can help retain business customers and property-management clients, but the economics are usually defensive: fixed voice may support bundles or legacy needs more than it drives growth.

Customer concentration is the hardest missing variable. The public website names settlement coverage and service categories but not subscriber count, customer names, top-ten revenue concentration, renewal rates, churn, route utilisation or SLA obligations. The Moscow Silk table is attractive, but without a customer count it can be a single building opportunity rather than a repeatable business segment. The cottage table is also attractive, but without route maps and connection mix we cannot tell how many customers are incremental on built fibre and how many require costly field work.

The apartment plans look competitive but low-margin unless clustered. A serious valuation of the business would ask for subscribers by tariff, connected premises passed, homes passed but not connected, average installation cost, churn by settlement, trouble tickets per 100 lines and renewal capex per access node.

Unofficial network signals are consistent with a consumer-access component, but they should be given low weight. IPinfo describes a pronounced day-night rhythm as a consumer or eyeball-network signature and tags at least one IP in the ASN with BitTorrent and VPN signals. AbuseIPDB shows a single historical report on one IP address with zero percent confidence of abuse. These are not strategic facts. They do not prove customer quality or network misuse. They simply make the network look like an active access network with ordinary residential or small-business internet behaviour, rather than a silent corporate-only ASN.

The competitive constraint comes from the price anchors around Moscow and Moscow region. Large national operators advertise or are listed in current public sources with much lower home broadband prices in dense buildings. MTS support describes home-internet, TV and mobile bundles and its Moscow-facing pages show 100 Mbit/s or higher packages far below INTER-SVYAZ-GROUP's cottage 100 Mbit/s price. Beeline's public Moscow home-internet pages show 100 Mbit/s and promotional prices far below the 6,612-ruble cottage plan.

Rostelecom's public pages emphasise address-based connection, xPON, equipment supply and broad digital bundles, while tariff aggregators list Moscow and Moscow-region Rostelecom plans at a fraction of INTER-SVYAZ-GROUP's highest residential price. Megafon-facing home-internet listings also reinforce the same consumer price anchor. These are not like-for-like comparisons for a Barvikha private house that may need a bespoke fibre drop. They still matter because customers compare a monthly bill with the alternatives visible to them.

The national-operator threat is not only price. Rostelecom, MTS, Beeline and Megafon can bundle mobile, TV, streaming, cybersecurity, WiFi hardware, loyalty benefits and family SIM plans. CNews reported that Rostelecom had 15.6 million broadband subscribers in 2025 and that MTS had 5.35 million broadband subscribers. Those subscriber bases let national operators amortise platforms, advertising, call centres, procurement and software over far larger scale than INTER-SVYAZ-GROUP. They can also tolerate promotional pricing in selected locations. A small local operator cannot win a subsidy contest on generic apartment internet.

Its advantage must be physical locality, customer familiarity, faster field response, bespoke private-house construction and business-site knowledge.

Moscow's own competition roadmap makes that pressure explicit in a different way. The city document says Moscow had more than 300 providers offering broadband access as of 2025 and a very high share of private organisations in broadband provision. It also describes the city as having high telecom accessibility, mobile penetration around two SIM cards per resident and more than 30,000 public WiFi access points. INTER-SVYAZ-GROUP's actual service territory is not central Moscow apartment broadband, but the wider environment affects price expectations, labour markets and customer tolerance for outages.

A local operator in an affluent district may serve customers who want a known engineer and a stable line, yet those same customers can use mobile data or national fixed operators as fallback threats in negotiation.

Wholesale access cost is partly mitigated by peering, but only partly. An MSK-IX presence can reduce the cost or latency of reaching local Russian networks, caches and exchange participants if traffic volumes justify the port and operational overhead. PeeringDB's 1 Gbit/s listed port is not large relative to modern city networks, and the company's own 10 Gbit/s between access nodes does not tell us internet transit capacity. BGP.tools reports three upstreams; IPinfo also lists three upstreams. That is positive for resilience compared with a single-homed small ISP.

But redundancy costs money: port charges, cross-connects, routers, optics, monitoring, route filtering, on-call skill and occasional emergency work. A small network must have enough paying customers on each route for redundancy to be more than a marketing phrase.

Suppliers create a second cost channel. The operator names Cisco, Ericsson, Eltex, Calix and Alcatel-Lucent in its routing, switching and GPON stack. Some of those brands are international, some have Russian market presence through local channels, and all require parts, software knowledge and compatible replacements. The public record does not show whether the installed base is current, refurbished, locally supported, or partially legacy. That matters because the customer-facing promise is not just speed.

GPON and Ethernet access networks require optical splitters, OLT capacity, ONTs, spares, power, cabinets, fibre documentation and staff who understand the old as well as the new plant. The tariff can fund renewal only if management has priced the hidden replacement cycle, not just current bandwidth.

Regulatory evidence is mixed and should be treated carefully. The company website has a licences page but the public crawl does not expose detailed licence entries. Business registries report active communications licences, but they do not agree on the count: T-Bank displays four current licence records and Companium states nine active licences, split between EGRUL and Roskomnadzor sources. T-Bank's event log also says several licence activities were suspended on January 20, 2025, while the same page still presents communications licensing as active.

Xfirm and Companium report a 2020 Roskomnadzor inspection with violations and arbitral administrative cases in 2016 and 2020. The right conclusion is not that the company lacks authority to operate. It is that licence status, activity scope and any suspensions need direct regulator confirmation before publication or investment reliance.

The court and inspection signals are modest but not irrelevant. Xfirm lists three arbitral matters with the company as defendant, including a 2025 lease-performance dispute involving Moscow City Telephone Network with a claim amount of 372,026 rubles, a 2020 Roskomnadzor administrative matter, and a 2016 administrative matter concerning business activity without registration or special permission. Companium lists two arbitral matters and the 2020 inspection. These are small signals in the public record, not proof of systemic operational weakness.

They do matter because a small operator's margin can be sensitive to exactly this kind of friction: rent for network sites, licence paperwork, inspections, small penalties and field-access disputes.

The capital structure is also under-disclosed. Public registries show a 10,000-ruble charter capital, which is standard for many Russian limited-liability companies and not a useful measure of network value. Aggregators report no large public debt story and show microbusiness status, but the company could still have lease obligations, supplier credit, related-party infrastructure arrangements or customer advances that do not appear in summary pages. If the network has been built over two decades, the historical replacement cost may be much larger than the visible balance-sheet clues.

Conversely, if much of the civil plant is already depreciated and routes are stable, the economics can be very good for years.

There is a strong local-density argument in the company's favour. The named settlements are not random low-income rural villages. Barvikha, Zhukovka, Gorki-2, Usovo and nearby areas include high-value private housing, business premises and institutional customers whose willingness to pay for reliable bespoke connectivity can exceed mass-market apartment prices. A local operator with known routes, known property managers and field crews near the service office can respond faster than a national call-centre process.

The high cottage and Moscow Silk tariffs make sense only if some customers value that locality and if alternatives are imperfect at the specific address.

There is also a strong compression argument against complacency. Russia's broader telecom market has been growing in nominal terms, with TMT Consulting estimating 2025 telecom market growth of 7.7 percent to more than 2.2 trillion rubles, while Nexign and TelecomDaily figures reported by Kommersant put market growth at 6.5 percent to 2.3 trillion rubles. Both sources point to price increases and fixed broadband growth, not a collapsing demand environment.

But nominal market growth can coexist with local price compression when national operators push bundles, when mobile substitutes improve, or when customers demand higher speeds at the old price. A local 100 Mbit/s cottage plan priced above 6,000 rubles per month needs service differentiation, not only bandwidth.

The speed ladder itself raises a renewal issue. In the apartment market, 100 Mbit/s is no longer premium. In business markets, 1 Gbit/s is increasingly normal for many offices, cameras, cloud backups and WiFi-dense premises. INTER-SVYAZ-GROUP's Moscow Silk table prices 1 Gbit/s as a high-end 90,000-ruble service, while mass-market competitors advertise hundreds of Mbit/s for households at far lower prices. The contexts differ, but customer expectations move across contexts.

If a business customer can get a national operator build, a wireless backup, or a different fibre provider at a lower price, the local operator has to justify its premium through uptime, route diversity, service response, static addressing, private wiring knowledge or integration support.

The company's broader service menu may be the answer. Video surveillance systems, structured cabling, low-voltage networks, telephony systems, WiFi, DECT and home automation are labour-intensive, locally specific services. For a private household, gated settlement, office or public building, the provider that already knows the fibre route may also know the cameras, access points, PBX and automation wiring. That can reduce churn because switching broadband provider becomes more complex than changing a home router. It can also improve contribution because installation and maintenance work can carry project margin.

The public record, however, does not break revenue by service line, so we cannot tell whether the company is an ISP with side projects or an integration contractor with recurring broadband attached.

Pricing of static IPv4 addresses is another small clue. The residential table charges 180 rubles per month for a /30 and 600 rubles for a /29. The Moscow Silk table charges 216 rubles for a /30, 708 rubles for a /29, 984 rubles for a /28 and 1,440 rubles for a /27. In a world of scarce IPv4, those add-ons can matter for business customers, cameras, VPNs and small servers. But a total visible origin footprint of 1,536 addresses limits how much static-address revenue can scale. It also makes IPv6 absence more strategically relevant over time, even if many Russian access customers still operate comfortably on IPv4 and NAT.

Labour is probably a central cost, not an incidental one. The 2025 headcount reported by registries is small, around nine employees, while 2024 average headcount is reported at 13 by RBC. A small headcount can produce strong revenue per employee if the network is mature and the customer base is stable. It can also mean operational fragility: a few engineers may hold route knowledge, customer relationships and legacy-equipment expertise. Local support labour is part of the product. The question is whether the price premium funds enough staff depth for holidays, illness, emergencies, regulatory paperwork and night work.

Installation economics are equally sensitive. The cottage connection fee of 4,800 rubles is low relative to the likely all-in cost of bespoke fibre construction if no route is nearby, which is why the exclusions for GPON subscriber equipment and optical cable entry matter. If the customer pays those separately, the operator can protect capital. If competitive pressure forces the operator to waive or discount them, the payback period lengthens. The Moscow Silk connection price of 2,400 rubles suggests the building economics are different: the operator may already have presence or a structured pathway, so the marginal activation cost is lower.

Support distance is the hidden denominator behind all of these prices. A repair in an apartment building can be a short lift ride from a basement cabinet, or it can be a scheduling problem with a building manager. A repair in a private-house settlement can involve gates, weather, access permissions, long drops, damaged ducts, customer-owned indoor wiring and a different expectation of personal service. None of those items appears directly in the tariff table, but the gap between apartment and cottage pricing strongly suggests that the operator recognises them.

If a national operator can serve the same address with a standardised process, INTER-SVYAZ-GROUP's premium has to be justified by faster local response and better knowledge of the physical route. If national coverage is absent or slow to repair, the local operator's premium can be rational.

Backhaul concentration is the matching network risk. A small autonomous network can look resilient because it has multiple upstreams and an exchange port, yet still depend on a small number of physical paths from its settlements to Moscow interconnection points. The public record proves routing relationships more clearly than route diversity. It does not show whether two upstreams enter through physically separate ducts, whether MSK-IX access depends on a single metro path, whether access nodes have independent power protection, or whether private-house fibre laterals are mapped well enough for fast repair.

For this reason, "redundancy" should be treated as a claim with layers. Routing redundancy is visible. Physical route redundancy, power redundancy and field-spares depth are not.

The same point applies to utilisation. PeeringDB's self-reported 5-10 Gbit/s traffic band could be healthy for a small regional operator, but it is too broad to price the network. If the access network has many customers using little peak bandwidth, the operator can hold down transit and upgrade costs. If a small number of business or high-end residential customers drive evening peaks, the operator may need expensive upgrades before subscriber revenue obviously justifies them.

The company-published 10 Gbit/s between access nodes sounds comfortable, but without peak utilisation, oversubscription policy and traffic split between peering and transit, it cannot answer the margin question. The best interpretation is that the company has enough infrastructure to be credible; the unresolved question is whether spare capacity is abundant or merely advertised.

The most favourable reading is that INTER-SVYAZ-GROUP has already built the difficult local plant, operates in high-value settlements, earns high cottage and business prices where alternatives are weaker, controls some routing and peering, and supplements access subscriptions with integration services. Under that reading, the 2024 and 2025 profit figures are plausible signs of a mature regional operator harvesting a dense local network.

The correct management priority would be disciplined retention: protect service quality, keep routes documented, avoid underpriced bespoke builds, use MSK-IX and multiple upstreams to manage bandwidth cost, and reserve cash for equipment renewal.

The less favourable reading is that the public margin is temporarily flattered by underinvestment, ageing equipment, low reported depreciation, thin staffing or a handful of business customers. Under that reading, national operators can compress the apartment and lower-end broadband price, mobile data can cap customer willingness to pay for backup connectivity, and one lost building or one route dispute can have an outsized effect. The licence and court/inspection ambiguities do not prove this scenario, but they remind readers that small-operator friction can consume management time and cash.

One judgment-changing watchpoint is public-address slack, because it connects the network-resource evidence to the customer mix more directly than headline speed does. INTER-SVYAZ-GROUP's visible routing footprint is only 1,536 IPv4 addresses, while the tariff pages sell static subnet options and the services page advertises business-facing systems such as video surveillance, corporate data networks, telephony, WiFi and low-voltage integration. Those are exactly the use cases where customers may ask for predictable addressing, remote access, VPN endpoints, camera access, monitoring systems or small routed blocks.

If most access customers are ordinary households behind NAT, the limited IPv4 pool is not a near-term margin problem. If the growth opportunity is business-quarter service, private-house security systems, managed camera networks or offices that want fixed addressing, the same pool becomes a scarce monetisable input and a possible constraint on expansion.

The economics of that constraint are ambiguous, which is why it deserves monitoring rather than a forced conclusion. Static-address charges in the public tables create incremental monthly revenue with little extra field work once the customer is already connected. That improves contribution and can make a business line more valuable than its bandwidth alone suggests. But the address pool is finite, and the current public routing evidence does not show a visible IPv6 origin that would let the company reduce long-run dependence on IPv4 for new services.

A small operator can manage this by rationing static blocks, using carrier-grade NAT for residential customers, charging more for business addressing, recovering unused assignments, or introducing IPv6 where customer equipment and support processes are ready. Each option has a different cost: rationing can limit sales, NAT can complicate support, higher static pricing can irritate customers, reclamation consumes staff time, and IPv6 adds operational work before it adds obvious revenue.

This watchpoint would change the judgment in both directions. A clean IPv6 deployment, stable RPKI-valid prefix visibility, disciplined static-address pricing and evidence that business customers accept NAT or pay for public blocks would strengthen the case that the operator can keep monetising a small address base while preserving cash for renewal. By contrast, signs of address exhaustion, customer complaints about remote access, static blocks being given away to retain accounts, or business customers moving to larger providers for cleaner addressing would weaken the retained-margin story even if subscriber counts were stable.

In a local network, scarcity can be a premium feature when it is priced and managed. It becomes a margin leak when it is used as an unpriced retention concession.

The most important facts that would change the judgment are specific and obtainable. First, subscribers by tariff and settlement would show whether revenue is diversified across many small lines or concentrated in a few business customers. Second, route kilometres, homes passed and connected-premise density would show whether each incremental customer is genuinely low-cost. Third, capex and maintenance history by year would reveal whether profit is being reinvested or harvested. Fourth, churn by customer class would show whether premium cottage pricing is sticky.

Fifth, upstream invoices, peering traffic share and port utilisation would show whether network cost scales efficiently. Sixth, a direct current licence extract would resolve the contradictory public licence summaries.

Until those facts are visible, the analytical conclusion should stay narrow. INTER-SVYAZ-GROUP appears to have a real local network, a defensible identity and a tariff structure that can produce retained margin when incremental connections sit on existing routes. The company is not priced like a generic Moscow apartment broadband reseller. It is priced like a local fibre and systems operator serving private houses, selected apartments and business premises where route knowledge has value.

But the public evidence does not prove that density is deep enough to fund the next renewal cycle before national alternatives push customers toward lower monthly prices. The investment question is therefore not "does it have a network?" The evidence says yes. The question is "how many profitable connections are close enough to that network, and how much of today's profit is already spoken for by tomorrow's redundancy and replacement?"

Sources