Summary

  • AMOS has an operating footprint, not merely a corporate registration. Public records connect the Alushta company to five active Russian communications licences, RIPE NCC Local Internet Registry status, AS208701, one routed IPv4 /22 and an allocated IPv6 /29. The IPv4 network is visibly active and takes reachability from Miranda-Media and CrimeaCom South.
  • The 2025 accounts are profitable but economically mixed. Revenue increased 13.9% to RUB41.495 million, cost of sales reached RUB32.824 million and net profit declined 2.5% to RUB5.644 million. The implied gross spread was RUB8.671 million, or 20.9% of revenue. Growth without profit growth is evidence that the marginal rouble of sales may be less attractive than the existing base.
  • Resource-holder status gives AMOS control and optionality, not an automatic moat. Its 1,024-address IPv4 block is scarce, while the much larger IPv6 allocation could support future migration. But there is no visible IPv6 origin announcement, the IPv4 route currently has no validating Route Origin Authorisation, and public routing views show no downstream networks. Those gaps limit the claim that registry assets already produce differentiated wholesale demand.
  • The strongest strategy is density rather than breadth: retain households, hotels, small businesses and institutions along existing plant; bundle Internet and television where the bundle reduces churn; charge properly for difficult private-home connections; and use local support as a measurable service advantage. A general contest in speed, content or multi-service discounts against larger regional operators would put AMOS in the price-taker role.
  • The judgment would improve if AMOS disclosed stable or rising contribution margin by product, low customer churn, disciplined installation payback, resilient physically diverse upstream paths, live IPv6 service and valid route-origin protection. It would worsen if 2025 growth came from discounted connections, if one or two contracts explain most of the increase, or if replacement capital and compliance spending exceed recurring cash generation.

The management incentive is to stay locally necessary

A small fixed-network operator does not need cloud economics to survive. It does need a reason for customers to keep paying when a larger rival can advertise more speed, more television, mobile service and a lower introductory price. AMOS's reason can be local reach: a cable already entering a building, an engineer who can visit Partenit or Maly Mayak, a relationship with a hotel or apartment manager, or a bundled bill that is easier than replacing both Internet and television. Those advantages are commercially real when they lower acquisition cost, shorten repair time or increase customer reluctance to switch.

They are not the same as scale. A large operator spreads core routers, billing software, regulatory staff, television rights and advertising across a much larger subscriber base. AMOS must support similar categories of work with a company that RBC Companies reports has 12 employees. Its 2025 revenue was RUB41.495 million. That is enough to sustain a local enterprise; it is not enough to absorb careless expansion, a major equipment cycle or a sustained price war without consequence.

The economic chain begins with the payer. Households pay monthly fees for fixed Internet and television. Hotels, sanatoriums, shops and offices may pay for connectivity, television distribution, public addresses, installation, surveillance or support. Building owners can influence access to multi-dwelling properties. AMOS benefits when one drop, one field visit and one support relationship generate several recurring services. Customers benefit from local availability and continuity. The suppliers of transit, content, equipment, electricity, software, poles, ducts and registry services are paid before the owner receives a return.

The downside is concentrated in AMOS. A household can switch to mobile broadband or another fixed provider where coverage overlaps. A hotel can seek a second carrier. An upstream can insist on a minimum commit. A television supplier can raise fees. A regulator can require a network change without creating revenue. AMOS remains responsible for the call, the truck roll and the restoration even when the fault begins outside its network.

This makes the central question narrower than whether AMOS is a genuine provider. The evidence says that it is. The question is whether local demand is differentiated enough to earn more than the cost of maintaining that status. The 2025 accounts say the existing business earned money. They do not yet show that growth improved the return.

Identity, ownership and the operating boundary

The legal company is the limited-liability company whose full Russian name translates as Broadcasting Company AMOS. Public company records give tax identifier 9101000959 and registration number 1149102039350. They place its legal address in Kiparisnoye, within the Alushta municipality. The Russian record dates the underlying company to May 1998, while the current registration number was issued in 2014. T-Bank's public counterparty profile identifies Yuri Kotov as director since July 2018 and Darya Gulivataya as the 100% owner. Concentrated ownership can make investment decisions fast, but it also makes succession and governance important to the value of a small technical business.

The operating classification is more useful than the historic broadcasting name. The principal registered activity is wired communications. Additional activities include data processing and hosting, software and technology consulting, advertising and computer repair. Registered activity codes are not proof of current revenue, but wired communications matches the visible customer and network evidence.

Licensing adds substance. The same counterparty profile lists five active communications licences, including two issued in April 2025. A licence does not prove service quality or subscriber count, and rights have to be read with their geographic and service conditions. It does show that the company maintains a formal operating position beyond an address allocation.

Public service listings consistently describe AMOS as an Internet and cable-television provider in greater Alushta. A local Alushta business catalogue lists Internet, cable television, IPTV and satellite television. Other listings identify service points in central Alushta, Partenit and Maly Mayak. A Yandex Services profile carries the company's own description of operating since 2005 and offering 74 analogue and 148 digital television channels. Those channel counts are marketing claims on a platform profile, not an independently audited current lineup.

The boundary of the business should therefore be drawn carefully. AMOS appears to control customer access plant in several Alushta-area settlements, retail service relationships, communications licences and Internet number resources. It may install antennas, private branch exchanges or surveillance systems. Public evidence does not establish how many kilometres of fibre or coaxial cable it owns, how many buildings it serves, whether it owns every headend and tower it uses, or which services are subcontracted. It also does not establish that every route listed in an Internet registry is a paying commercial relationship.

That distinction matters in valuation. The company is not the sum of every address, route and service label around its name. It is the cash-generating rights, contracts, plant, staff capability and customer relationships it can retain and transfer. The strongest diligence would reconcile the legal entity, licences, network inventory, subscriber ledger, supplier contracts and resource registrations to one control schedule.

A bundled local model, not a miniature cloud

AMOS's public service surface suggests three economic layers. The first is recurring household connectivity. The second is television, which can be sold alone or bundled with Internet. The third is field and technical work: private-home connections, antenna and satellite installation, surveillance systems, small-office communications and related support. Each uses some of the same people and local reputation, but each has a different margin profile.

Fixed Internet is attractive when the access plant is already built. The next subscriber in a wired apartment building may need only a drop, customer-premises equipment and activation. Much of the core network and support cost already exists. The same connection is unattractive when a new rural or private-home customer requires a long cable run, difficult rights of way or repeated travel. A price that looks profitable at the monthly-service level can destroy value if installation cost is not recovered before the customer leaves.

Television can improve retention because replacing two services is more inconvenient than replacing one. It can also reduce margin. Channel and platform costs do not disappear when a local operator is small, and consumers increasingly substitute online video or larger operators' interactive packages. AMOS should value television by its effect on household contribution and churn, not by channel count. A large lineup that customers will not pay for is a liability disguised as a feature.

Technical services can create useful one-off cash and open the door to recurring connectivity. A surveillance installation at a hotel, for example, may lead to managed connectivity or maintenance. An antenna job can produce a television customer. The danger is labour absorption. Twelve employees cannot simultaneously run a network, answer faults, perform complex installations and pursue many bespoke projects without trade-offs. Revenue from hardware resale or one-off work can expand the top line while contributing little after equipment and labour.

The portfolio therefore needs a contribution hierarchy. Recurring access on existing plant is likely to be the most valuable if churn is low. Television is valuable when it raises retention more than it adds content and support cost. New construction is valuable when the installation fee, contract term and expected monthly margin recover the build within a defined period. Technical projects are valuable when priced for labour and risk or when they lead to high-quality recurring accounts.

Nothing in the public evidence supports treating AMOS as a cloud provider. Hosting appears among its registered activities, and the IPinfo profile for AS208701 observes only about 20 hosted domains on a handful of addresses. That is compatible with some local hosting or service infrastructure, but not with a broad compute platform. The correct alternative is not to imitate a cloud. It is to use local plant, public addressing and field support where a remote platform cannot solve the last-mile problem.

The resource footprint is real, with unfinished security and IPv6 work

The clearest technical evidence is AS208701. The current RIPE Database record assigns the autonomous system to AMOS's Local Internet Registry organisation and records it as created in June 2019. The organisation record uses the same legal registration number and Kiparisnoye address as the Russian company. That alignment strongly supports common control of the corporate and network identities.

The network originates 45.88.52.0/22, an IPv4 block containing 1,024 addresses. BGP.Tools reports one active IPv4 prefix, no active IPv6 prefix, two upstreams and no downstream networks. IPinfo's range page has observed many addresses responding to probes and recent routes into the block. The day-night traffic pattern on the ASN page is consistent with an access network used by people, although such third-party classification is probabilistic.

One /22 is economically meaningful for a provider of AMOS's size. Public addresses can support subscribers that need direct reachability, business services, network equipment, servers and operational separation. Scarcity gives the block an alternative value. A mid-2026 market commentary from IPv4 Global cited indicative purchase prices of US$28 to US$40 per address for /22 to /24 blocks. Multiplying those figures by 1,024 produces a rough range of US$28,672 to US$40,960 before brokerage, diligence, reputation, tax and transfer constraints.

That calculation is not a valuation of AMOS or even of its block. The space may be essential to customer service. Its transfer may be restricted by contracts, registry requirements or sanctions. Address reputation and clean registration matter. Selling the block could force carrier-grade address sharing, reduce service quality or destroy revenue. The useful point is that management allocates a scarce resource: it should know the contribution earned per public address and compare operating use with the cost of substitutes.

The IPv6 position is more revealing. The RIPE Database has allocated 2a07:3d40::/29 to AMOS and contains a route6 entry pointing to AS208701. Yet current public routing views show no originated IPv6 prefix. Registration has therefore moved further than visible deployment. A /29 is more than enough space for a regional access provider; address scarcity is not the barrier. The missing evidence is routing, customer delegation, resolver reachability and measured dual-stack use.

Route-origin protection is another unfinished task. A current RIPEstat validation query returns an unknown state and no validating Route Origin Authorisation for the IPv4 /22. The route has an Internet Routing Registry record, but that is not the same as cryptographic origin authorisation. Creating a narrowly configured authorisation would not prevent every BGP incident, but it would let networks that perform route-origin validation distinguish the authorised origin from an unauthorised one.

These gaps do not make the network unreal. They show the difference between holding resources and operating them to a high standard. AMOS has numbering independence, a visible route and multiple paths. It has not publicly demonstrated complete route-origin protection or customer-facing IPv6. Resource-holder status is thus a foundation for differentiation, not proof that differentiation has been achieved.

Revenue grew; economic value did not grow with it

The 2025 accounts are the most important evidence in the case. RBC Companies reports revenue of RUB41.495 million, up from RUB36.420 million in 2024. It reports 2025 cost of sales of RUB32.824 million and net profit of RUB5.644 million. Xfirm independently presents the same 2025 revenue and profit and shows revenue up 13.9% while profit fell 2.5%.

The arithmetic deserves attention. Revenue increased by RUB5.075 million, but profit decreased by about RUB145,000 from roughly RUB5.789 million in 2024. The implied 2025 gross spread was RUB8.671 million, or 20.9% of sales. Net margin was 13.6%. Cost of sales absorbed 79.1% of revenue. The gross spread averaged about RUB723,000 a month; net profit averaged about RUB470,000 a month.

This is not a distressed income statement. A 13.6% net margin is respectable for a small communications business. The concern is the direction of the marginal economics. If revenue rises 13.9% and earnings fall, the added sales may have arrived with high direct cost, higher operating expense, tax effects or unusual items. Public summaries do not provide a full income statement or cash-flow statement, so no single explanation can be verified. The pattern still puts the burden of proof on growth.

The employee figures sharpen the scale. Using the reported 12-person average headcount, 2025 revenue was about RUB3.46 million per employee and net profit about RUB470,000 per employee. Revenue per employee is not productivity in isolation; transit, content, equipment and power flow through the top line. It does show how little room exists for a strategy that adds labour faster than contribution.

The cost structure also reveals an asymmetry. A one percentage-point reduction in 2025 net margin equals about RUB415,000, nearly a month of average reported profit. A RUB1 million equipment replacement equals more than two months of average profit. A major bad debt or underpriced build could be material even though the annual margin looks healthy. The relevant reserve is cash and available credit, neither of which is disclosed in the public summaries.

Value creation should be measured against a conservative replacement cycle. Routers, optical terminals, switches, batteries, cooling, television headend equipment and customer devices age. Software and compliance work recur. If depreciation is lower than the cash needed to replace the actual network, accounting profit overstates distributable cash. Conversely, if plant is young and most 2025 cost of sales was a one-off expansion that creates future recurring revenue, the accounts may understate future earning power. The missing capital-expenditure schedule is therefore central.

Management should separate four movements: price increases on the installed base, subscriber additions on existing plant, revenue from new construction, and one-off equipment or project sales. The first two usually have the best economics. The third can be attractive if payback is controlled. The fourth can flatter revenue. Without that bridge, a larger top line can be mistaken for a stronger franchise.

Unit economics must be reconstructed from contracts, not slogans

AMOS does not publish enough current tariff, subscriber or churn data to calculate verified household unit economics. Market prices can still establish useful boundary conditions. A current Alushta comparison page lists Miranda-Media Internet and television packages beginning around RUB650 a month, with 200 Mbit/s and 500 Mbit/s fibre offers in the market. Volna's official home-Internet page markets a combined fixed, television and mobile proposition across Crimea, while an Alushta partner page advertises plans from about RUB500 a month.

Those are competitor and promotional prices, not AMOS prices. Used as a sensitivity, RUB500 to RUB750 a month would require about 4,600 to 6,900 full-year account-equivalents to produce all RUB41.495 million of 2025 revenue. AMOS certainly has other revenue and some customers may buy more than one service, so this is not a subscriber estimate. It demonstrates the scale relationship: low household tariffs require thousands of recurring bills, while business, installation and television revenue can reduce the required account count but may carry different costs.

For each household, the useful metric is lifetime contribution, not monthly revenue. Monthly contribution equals the bill less content, upstream, payment, customer equipment support and other costs that move with the account. Lifetime contribution then depends on churn and the cost to acquire and connect the customer. A RUB650 bill with RUB250 of monthly contribution produces RUB4,800 a year before shared overhead. A RUB12,000 private-home build would need two and a half years merely to recover installation cost at that contribution level. These figures are examples, not claims about AMOS's costs; they show why management needs address-level payback rules.

Density changes the result. Ten new apartments in one building can share fibre, switching, travel and sales effort. One distant house cannot. A hotel may justify a dedicated route if it signs a term contract and buys support, television or surveillance. A seasonal rental property may suspend or cancel service outside the tourist period. The same advertised speed can therefore carry very different value.

Price is only one lever. Installation fees can recover construction. Minimum terms can protect payback. Annual prepayment can improve cash. Equipment rental can turn replacement into recurring revenue, although it also adds inventory and failure obligations. Business packages can charge for service levels, public addresses, managed Wi-Fi or faster restoration. AMOS should avoid hiding bespoke support inside a household tariff.

The 2025 profit pattern suggests one of three things. The company may have acquired lower-margin customers, incurred the cost of expansion before receiving the full recurring benefit, or faced inflation and compliance costs that offset price and volume. Product-level contribution by cohort would distinguish them. If accounts connected in 2025 generate attractive contribution in 2026 without another build cycle, the growth was sensible. If they require discounts and repeated service work, the company has bought revenue.

The cost base is small in absolute terms and stubborn in structure

A regional fixed network combines variable costs with a large set of step-fixed costs. Transit may be bought on a committed port or capacity tier. Television content and platform charges can depend on subscribers or minimum guarantees. Electricity, site rent, monitoring and software continue even when traffic is quiet. A small support team cannot be reduced every time one customer leaves. These characteristics make utilisation more important than raw speed.

Direct cost of sales averaged about RUB2.735 million a month in 2025. Public accounts do not divide that figure among transit, content, materials, equipment, subcontractors and wages classified into direct cost. That absence prevents a verified gross margin by service. It also means the 20.9% company-level gross spread should not be compared casually with another operator whose accounting classification differs.

Capital comes in uneven increments. An optical line terminal, core router, storage expansion, battery bank or headend change can require cash before any new customer pays. Imported components and advanced telecommunications equipment face constrained supply channels. The European Commission's current sanctions summary describes broad restrictions on advanced technology exports to Russia. Separate EU measures for Crimea restrict certain telecommunications goods, technology and related assistance. Even when AMOS buys through a domestic distributor, limited vendor support and longer replacement chains can raise inventory needs and shorten useful planning horizons.

Registry cost is visible but not the main burden. The RIPE NCC's 2026 billing procedure sets an annual contribution of EUR1,800 per Local Internet Registry account, with additional charges for certain assignments. For AMOS, the fee buys registry service and resource-management capability. The larger cost is the staff discipline needed to keep records, routing policy, abuse handling, reverse DNS and security current.

Compliance creates another fixed layer. Russian rules require communications operators to retain specified communications data for periods of up to six months. A government-law summary describes the storage obligation. Technical rules for threat-countermeasure equipment require operators to provide power, access and a management channel; the installation conditions include four hours of backup power and a management channel of at least 100 Mbit/s. Whatever equipment is centrally supplied, site space, power, network integration and staff time remain local operating demands.

The correct capital policy is therefore selective replacement plus density-led expansion. AMOS should first protect the core, backup power, security, monitoring and the access segments with the highest recurring contribution. It should add plant where signed demand meets a payback threshold. It should resist broad construction justified by homes passed rather than customers contracted.

Upstreams are suppliers, substitutes and strategic constraints

Public routing policy names several adjacent networks, but current observation consistently identifies two principal upstreams: Miranda-Media and CrimeaCom South. Multiple providers are better than one because they allow path choice and failover. They do not guarantee resilience. Both contracts could enter the same building, share regional transport or depend on common infrastructure. Logical diversity is valuable only when capacity, physical routes, power and operations are independently survivable.

The bargaining imbalance is considerable. BGP.Tools reports Miranda-Media with scores of originated prefixes, four upstreams, more than 70 inferred downstreams and connections at major exchanges. CrimeaCom South also operates a broader regional network than AMOS, and a separate network profile counts 8,192 originated IPv4 addresses and visible IPv6. AMOS originates 1,024 IPv4 addresses and has no observed downstream network. Its traffic purchase is likely material to itself and small to its suppliers.

Those upstreams can also compete for the end customer. Miranda-Media sells retail fibre and television in Alushta. Crimea-wide mobile and fixed bundles add further pressure. A supplier that controls backbone scale and also markets at retail can capture both wholesale margin and customer relationship. AMOS needs either a cost-efficient contract, a last-mile location the supplier does not reach economically, or service quality that keeps the retail account local.

The RIPE record also includes import and export policy with the Crimea Internet Exchange route set and with other regional networks. That indicates planned or historic interconnection beyond two default routes. It does not prove settlement-free peering, current traffic, port speed or savings. A network can have an exchange policy while sending most useful traffic through paid upstreams. The diligence questions are concrete: how much traffic takes each path, what are the commits, which paths share ducts, what capacity remains at peak, how often failover is tested and what content is reached locally?

Local exchange can be an economic counterweight if it removes paid transit and improves latency. The value depends on traffic mix. Exchanging routes with another small access provider saves little if users spend most of their time on large video, search and messaging platforms reached elsewhere. Caches and direct content links can matter more than a long peer list. AMOS should measure roubles saved per port and customer experience improved, not count routing neighbours as trophies.

Supplier concentration extends beyond transit. Television relies on content and platform providers. Customer routers and optical terminals rely on vendors and distributors. Billing, payment and messaging rely on software and financial channels. A 2025 banking tariff document lists TRK AMOS Internet as a supported household payment recipient, evidence of an established billing channel. Any interruption to a convenient payment method can still increase collection work for a small team.

The local market rewards density but limits the ceiling

Alushta offers a recognisable local market rather than an unlimited addressable one. A Republic of Crimea small-business portal gives the municipality a population of about 53,800 at the start of 2025. Rosstat's 2024 average was 54,446 residents, split between roughly 30,100 urban and 24,300 rural residents. A dispersed municipality creates both opportunity and cost: larger operators may prioritise dense urban blocks, while a local operator can serve villages and private homes, but long access lines and truck travel weaken unit economics.

Tourism changes demand through the year. Hotels, sanatoriums, restaurants, rental properties and seasonal households need connectivity and television, and many value rapid local repair. Visitor demand can increase bandwidth and support calls in summer while some properties reduce usage in winter. It can also create lucrative business accounts if contracts include managed Wi-Fi, surveillance or television. AMOS should plan capacity to the seasonal peak while pricing contracts over the full year.

Concentration risk can arise even without a dominant named customer. The company may depend on a few building owners, hotels or settlement-level access routes. Losing access to one large apartment complex can remove many subscribers at once. A cable cut on one branch can affect several villages. A hotel account may combine Internet, television and technical service, making product diversity look greater than counterparty diversity.

No reliable public source provides AMOS's current subscriber count, churn, average revenue per account or top-customer share. The current network graph shows no downstream autonomous systems, so the case is primarily retail and small-business rather than visible wholesale transit. That can be stable: thousands of modest monthly bills are often better than one wholesale contract. It can also be costly to support if the footprint is geographically scattered.

Contract durability should be measured in cohorts. Apartment customers on existing plant may be retained by convenience. Private-home customers may stay longer because switching requires another build. Hotels may be valuable but negotiate hard and demand rapid restoration. Public institutions can be sticky but may pay through formal procurement and slower cycles. A single blended churn rate hides those differences.

The most attractive customers are those who use sunk plant, pay reliably, take more than one service and generate few avoidable calls. The least attractive are those who require bespoke construction, buy on promotion, consume intensive support and can switch after the discount. The 2025 revenue increase should be decomposed along precisely those lines.

Competition turns speed into a commodity

AMOS does not compete with one provider. It competes with a set of substitutes that changes by address and customer type.

In wired apartment buildings, Miranda-Media can offer fibre and television at visible retail prices. The current Alushta comparison page shows entry packages around RUB650 and higher-speed offers with television. Volna promotes an integrated proposition that combines fixed Internet, television and mobile service in one payment. A customer comparing headline speed and bundle price can therefore choose a larger platform with broader marketing and product range.

Mobile broadband is another substitute, particularly for a seasonal resident, light user or home that is expensive to wire. It may not match fixed service for consistent capacity, gaming, multiple televisions or business use, but it can cap the price a fixed provider charges. A second mobile connection is also an easy backup for a hotel or office, weakening the premium for basic fixed availability.

Satellite service can reach difficult properties, and AMOS itself is listed as an installer. Satellite can complement the business, but it can also substitute for the company's fixed build where terrain makes cable uneconomic. The rational strategy is not to insist on owning every access path. It is to capture installation and support revenue while reserving network capital for locations with sufficient density.

Online video competes with the television bundle. Customers may keep Internet and drop linear channels. Larger providers can spread content and platform cost across more subscribers. AMOS needs to know whether television reduces churn enough to pay for itself. An analogue channel count is not a durable competitive advantage when consumers compare on-demand libraries, device support and price.

Business customers can substitute national or regional carriers, dual mobile links, managed service companies or cloud-hosted applications. AMOS's local advantage is practical: a known engineer, a fast site visit, a public address, a flexible cable route or one invoice for several local services. Those advantages must be specified in contracts. Vague claims of personal service will not support a premium after an outage.

The larger providers' scale does not guarantee better service at every address. Their call centres can be remote and their build thresholds rigid. AMOS can win where it has installed plant and local knowledge. But it cannot assume loyalty. The right competitive response is selective: match market speed where plant allows, avoid uneconomic promotions, publish clear restoration expectations, and charge separately for service features that businesses value.

Cloud scale affects AMOS indirectly. Large content networks reduce the value of generic hosting and make application delivery less dependent on a local server. They also raise customer expectations for always-on access. AMOS should not spend to reproduce remote compute services. It should make the path to those services reliable and use local support where the cloud has no field presence.

Regulation and geopolitics narrow strategic freedom

AMOS operates in Crimea, which Russia controls and which Ukraine, the EU and the United States regard as Ukrainian territory. That status is not a footnote to the cost base. It affects licences, suppliers, payments, registry diligence, asset transfer and the set of possible buyers.

Ukraine imposed sanctions on the company in December 2023. The official presidential decree 851/2023 enacted a sanctions decision covering listed legal entities, while a specialist Ukrainian sanctions record matches AMOS by both Russian registration number and tax identifier and gives an expiry date in December 2028. This is a specific company designation under Ukrainian law. It should not be presented as an EU, US or United Nations designation without separate evidence.

Regional restrictions apply independently. In June 2026, the EU extended its Crimea and Sevastopol measures to 23 June 2027. The regime restricts investment and certain exports, including telecommunications goods and technology for use in Crimea. The practical effect for a small operator is reduced access to foreign capital, equipment, technical assistance and counterparties. A transaction that is ordinary inside the local Russian market may be unacceptable to an EU-connected bank, vendor, insurer or acquirer.

US rules also distinguish the transmission of communications from the supply of network capacity and equipment. The Ukraine-/Russia-related regulations authorise certain telecommunications and Internet communications involving Crimea but do not generally authorise the sale or lease of telecommunications equipment, technology or transmission capacity under the cited provision. The details require transaction-specific legal analysis. Economically, the rules narrow vendor and financing choice even where end-user communication remains permitted.

RIPE NCC membership adds another jurisdiction. RIPE NCC is based in the Netherlands and must follow applicable EU sanctions. Its Ukraine and Russia guidance says that where applicable sanctions cover a holder, registration of number resources is frozen rather than their use: the holder cannot acquire or transfer resources, but resources are not automatically deregistered. There is no public basis here to say that AMOS's resources are frozen by RIPE NCC. The point is that registry optionality depends on cross-border legal status and due diligence, not only possession of a /22.

Domestic requirements add cost and operational exposure. The company has five active communications licences and must maintain the systems, subscriber records and network controls required of an operator. Filtering, data retention, lawful-access and personal-data obligations have a larger cost per customer for a small provider. Failure can threaten licences or create remediation expense. Compliance is part of network capital, even though it does not increase advertised speed.

Geopolitical risk also enters physical operations. Regional transport, power, banking and logistics can be disrupted. A small operator has less inventory and fewer alternate routes than a national carrier. It may nevertheless be locally faster to improvise and restore. The valuable capability is tested continuity: spare optics and switches, backup power, physical path records, supplier alternatives, documented authority and customer communications.

Unofficial signals identify the diligence questions

Public reviews of AMOS are too sparse to establish service quality. They are useful as prompts for measurement.

The 2IP provider page shows thousands of user-initiated speed measurements but only one written review, a positive 2021 comment about Internet, cable television and prices. Recent displayed tests vary widely, from roughly 13 Mbit/s to more than 80 Mbit/s downstream, with different upstream speeds and latency. Such tests depend on tariff, Wi-Fi, device, time and destination. They prove that users continue to measure the network, not that every subscriber receives a particular service level.

A local review directory presents two more recent complaints. One October 2025 reviewer said a 100 Mbit/s service delivered around 15 Mbit/s in the best periods and criticised the response from technicians. Another March 2026 comment reported difficulty reaching support. The directory itself displays only two reviews. Two self-selected accounts cannot support a general conclusion about thousands of possible service months. They identify the correct operating tests: peak-hour throughput, repeat fault rate, time to answer, time to restore and performance after a technician visit.

The contrast between the speed-test page and complaint page is instructive. Some recent tests are close to 90 Mbit/s in both directions; others are far lower. That variation could reflect different plans, access technologies, Wi-Fi conditions or congestion. AMOS should be judged by controlled wired measurements and fault records, not isolated screenshots. Management should know the 95th-percentile peak utilisation of each access segment and upstream, not only average traffic.

Public listings show multiple service addresses and long operating history, which supports local embeddedness. They also show inconsistent opening hours and no robust current public tariff archive. That may simply reflect stale directories. It may also indicate that sales rely on phone and local relationships. For a small company, poor public information raises acquisition cost and gives larger rivals an advantage in comparison shopping.

The official network evidence is stronger than review chatter. The ASN is active, the /22 is routed and several paths are visible. Yet it also raises questions: no observed IPv6, no validating route-origin authorisation and no visible downstream network. Those are verifiable engineering items that management can change. The reviews are not verdicts; they are a reason to connect engineering metrics with retention and refunds.

Strategy should allocate scarce cash, not imitate scale

AMOS has four realistic strategic paths.

The first is disciplined harvesting of the installed base. Management would focus on reliability, modest price indexation, television retention, public-address services and low-cost upgrades in buildings already reached. Capital intensity would remain controlled. This path protects cash but can allow the network to age if replacement is deferred too far.

The second is density-led expansion. AMOS would extend fibre or upgrade coax only where signed households, a hotel, a business cluster or a building agreement meets a payback threshold. It would recover exceptional installation cost through fees or minimum terms. This is the most credible growth path because it uses local knowledge without assuming a large market.

The third is a business-service niche. AMOS could sell monitored connectivity, dual access, managed Wi-Fi, surveillance support, public addressing and defined restoration to hotels and local enterprises. The revenue per account can be higher, but so can concentration and service liability. Contracts need explicit scope, escalation, indexation and penalties that AMOS can actually support.

The fourth is consolidation or asset partnership. A larger operator might value last-mile plant, building access, licences, subscribers or the /22. AMOS might buy transport or platform services while keeping the local relationship. Any transaction would face ownership, licence, registry and sanctions diligence. The address block should not be treated as freely separable from the operating network.

The wrong strategy is broad imitation. Building generic hosting, matching every television feature, discounting to win low-density homes and promising premium support at mass-market prices would consume the same small team and cash pool. Strategy without a capital schedule is marketing.

The 2025 result provides a simple management rule: every growth initiative must explain why profit fell while revenue rose and show how the next rouble will behave differently. A useful monthly report would bridge subscribers, average bill, churn, new-build cash, direct content cost, upstream cost, support visits and contribution by service. It would also reserve cash for replacement before declaring earnings available to the owner.

Operationally, three low-drama investments deserve priority. First, create and maintain route-origin authorisation for the /22. Second, turn the registered IPv6 /29 into measured dual-stack service, beginning with the core and business customers. Third, test upstream failover and document physical path diversity. None creates a glamorous new product. All make the resource-holder footprint more defensible.

Commercially, AMOS should sell outcomes. A household buys stable evening access and a television service that works. A hotel buys guest connectivity and rapid restoration in season. A business buys continuity, a public address and an accountable local contact. Pricing those outcomes is more durable than advertising the same headline speed as a larger rival.

What would change the judgment

The present judgment is cautious but not negative. AMOS is a profitable local operator with real network resources and a long service history. Its latest growth did not increase profit, and the technical record shows unused or underdeveloped elements. Several specific facts would change that view.

First, a revenue bridge and product-level contribution statement would settle the marginal-economics question. If the 2025 increase came from connections built early in the year that now deliver strong monthly contribution, the profit decline may be temporary. If the increase came from hardware resale, discounting or expensive custom work, growth has made the business busier rather than more valuable.

Second, subscriber cohorts would show durability. Low annual churn, high on-time payment and rising bundle penetration on existing plant would support pricing power. A large share of promotional or seasonal customers, repeated suspensions and expensive reconnections would weaken it. Top-ten customer and building-manager concentration should be disclosed separately from subscriber count.

Third, a physical network map tied to ownership and contracts would establish the moat. It should distinguish owned fibre and coax, leased routes, building access, core sites, backup power and shared failure points. A route that appears diverse in BGP but shares one duct should not receive a resilience premium.

Fourth, supplier contracts would reveal bargaining risk. The needed facts are transit commits and price escalators, television minimums, equipment lead times, vendor support, payment terms and alternate suppliers. Two upstream names are encouraging only if capacity and physical routes are genuinely independent.

Fifth, a five-year capital plan would test the earnings. It should list the age and replacement cost of core routers, optical equipment, switches, storage, headend systems, batteries, cooling and customer devices. If normalised replacement capital fits comfortably inside recurring cash after tax, the reported margin is valuable. If the network requires a large near-term refresh, 2025 profit is not freely distributable.

Sixth, routing security and IPv6 are objective watchpoints. A valid, appropriately scoped Route Origin Authorisation for 45.88.52.0/22 would improve route hygiene. A visible 2a07:3d40::/29 announcement, customer delegations and measured traffic would show that the company can use what it holds. Continued absence would turn an allocated resource into administrative potential rather than operating advantage.

Seventh, service metrics would answer the unofficial complaints. Peak-hour throughput by segment, packet loss, uptime, mean time to answer, mean time to restore, repeat faults and service credits should be tracked over at least a year. A small operator can charge for local accountability only if it measures and delivers it.

Finally, a clear legal and sanctions memorandum would define strategic freedom. It should cover the Ukrainian designation, current EU and US regional restrictions, RIPE NCC treatment, supplier exposure, transferability and possible buyer jurisdictions. The conclusion may be that the existing local business can continue while many cross-border transactions cannot. That difference matters to both operating value and exit value.

A profitable niche still has to earn its replacement cost

AMOS has more substance than its scale suggests. It holds communications licences, serves a visible local market, maintains its own autonomous system, originates a /22 and has more than two decades of corporate history. The 2025 net margin shows that the company can earn money. Its challenge is not legitimacy. It is allocation.

The resource footprint gives management options: public addresses, independent routing, multiple upstreams and a path to IPv6. Those options become a moat only when paired with reliable service, secure routing, dense customer plant and contracts that recover capital. A registry entry alone does not make customers pay more. A /22 does not repair a cable. Two upstreams do not guarantee two physical paths.

The accounts provide the final discipline. RUB5.075 million of extra revenue coincided with RUB145,000 less profit. That fact does not prove a broken model, but it blocks an easy growth story. AMOS must show that 2025 spending bought recurring contribution, or accept that it expanded as a price-taker.

Below cloud scale, relevance is earned through precision. Build where demand is signed. Bundle where it reduces churn. Charge for difficult service. Protect the route. Activate IPv6. Keep spares and cash for the network customers already trust. If AMOS can do those things while restoring profit growth, its local status is worth more than the addresses it holds. If it cannot, the upstreams, content suppliers and bundled rivals will capture the value while AMOS carries the downside.