Summary
- Altura has a real operating footprint: the Saratov company is a current RIPE NCC member, is associated with AS44158 and AS44552, originates visible IPv4 space and sits between several upstream and smaller regional networks. That supports a case for local interconnection, hosting and communications services, but it does not prove customer scale or pricing power.
- The financial constraint is severe. Public 2025 accounts report RUB10.417 million of revenue, RUB10.347 million of cost of sales and RUB16,000 of net profit. The implied gross spread is RUB70,000, or about 0.67% of revenue, leaving little internally generated cash for hardware replacement, compliance work or a bad debt.
- Altura's defensible value is different from its revenue. Public IPv4 blocks, routing relationships, local numbers, a functioning virtual-PBX product and accumulated operating knowledge may be useful assets. Their value is conditional on clean control, continued membership, customer retention and transferability; none should be confused with recurring earnings.
- The most credible path is a focused regional infrastructure role: business connectivity, local numbering, managed voice, address stewardship and interconnection for customers who value reachable support. Commodity household broadband and generic hosting are structurally harder because national operators and cloud providers can spread fixed costs across far larger bases.
The incentive is relevance, not scale
A small regional communications company survives by selling exceptions to the scale economy. The national carrier can usually buy backbone capacity, routers, software and advertising more cheaply per subscriber. A large cloud can automate provisioning and spread data-centre costs across thousands of servers. A mobile operator can bundle voice, broadband and applications into one bill. Altura cannot win a general contest against those cost curves.
It can still win particular transactions. A business may pay for a Saratov number that can be configured quickly, for an engineer who knows the local access path, for a static address, for a route to another regional network that avoids an unnecessary trip through Moscow, or for one supplier to combine connectivity, hosting and telephony. A smaller provider may value upstream reach and a local handoff. A customer moving offices may value keeping a number and call logic intact. These are narrow forms of convenience and risk transfer, not raw bandwidth.
That distinction identifies who pays and why. Local companies pay monthly access, hosting or voice charges and usage fees. Smaller networks may pay for transit, a port, address service or an associated technical service, although public routing observations do not reveal the commercial terms. Altura benefits from recurring bills and from using the same network and staff across several products. Upstream carriers, equipment vendors, software suppliers, utilities, landlords and the number-resource registry collect their own fees before the owners receive a return.
The downside moves in the opposite direction. When a circuit fails, a route leaks, a PBX is attacked or a regulator demands a network change, Altura bears the first response cost. Customers bear interruption and migration risk. If the company has promised service credits, it may also bear contractual penalties. Shareholders bear the residual downside because the profit buffer is exceptionally thin. A customer can often buy a replacement connection; Altura cannot replace lost recurring revenue as quickly if a large account leaves.
The correct question is therefore not whether Altura can become a miniature cloud. It is whether enough buyers attach a premium to local control, responsiveness and network continuity to cover all the costs that scale normally hides. The 2025 accounts suggest that the answer was, at best, barely.
Identity and the control boundary
The company behind the current record is a Saratov limited-liability company with tax identifier 6452005911 and registration number 1036405208950. RBC Companies' public profile says the business dates to July 1995, has eight employees, and is active. Its principal registered activity is telephone communications, with other wired-communications activity among the additional classifications. Evgeny Alekseev is the general director and owns 95%; Mikhail Tsymbal owns 5%. This is concentrated control in both legal and practical terms.
That ownership matters because the network records also point back to Alekseev. Public routing records for AS44158 identify the RIPE organisation handle ORG-DL87-RIPE and Altura's maintainer, while the same organisation handle appears on AS44552. A public RIPE NCC member list includes "Altura" ltd. as a registry based in Russia. The contact address in current network records matches the company's Kiseleva Street address. The evidence therefore supports a common administrative boundary around the company, its registry membership and both autonomous systems.
It does not settle every economic boundary. AS44552 carries the name SCTS-NET, commonly rendered as Saratov Digital Telephone Network, and public address records identify parts of that space with SCTS. The same ASN is nevertheless registered to Altura's RIPE organisation, maintained by Altura's network role and linked to AS44158. That could represent an Altura-operated access network, a customer or affiliated service, an inherited arrangement, or a commercial operation run under another name. Public data cannot determine which entity books the related revenue, owns every physical asset or bears every customer obligation.
The product boundary is clearer for voice. The A-Connect service rules explicitly name ООО «Альтура» as the operator, use the same tax identifier and address, and describe a platform for VoIP calls, customer accounts and Saratov, Engels and Moscow geographic numbers. A-Connect is therefore not merely a directory listing that happens to resemble the company; it is a live, company-operated service surface.
Older public references sometimes call the firm "Design-studio Altura," reflecting its history as a web developer before communications became the main business. The common tax identifier links that older trading description to the present company. The shift helps explain the current mix of hosting, data services and telephony, but it should not be used to assume that all historical products still contribute revenue.
The practical control test is simple. Altura appears able to maintain registry records, originate routes, administer the voice service, allocate numbers within its service and contract under its own legal name. It has not publicly demonstrated ownership of a large fibre estate, a modern hyperscale-style facility, or every address its networks route. The investable boundary is thus the legal company plus the rights and contracts it can prove, not every device, customer network or historical marketing claim visible around the ASNs.
What the business appears to sell
Altura's public descriptions span five related lines: internet access and data transmission, hosting, data-centre service, fixed and IP telephony, and interconnection. The lines share infrastructure but have different economics.
Connectivity produces monthly recurring revenue, installation fees and possibly usage or capacity charges. Its direct costs include upstream bandwidth, leased fibre or building access, customer equipment, field work and support. It is attractive when many customers share the same backhaul and access plant. It is unattractive when each new site needs a bespoke build or when a customer can switch to an already-wired national operator.
Hosting can reuse power, racks, addresses and network operations, but generic shared hosting is highly commoditised. IPinfo's current AS44158 profile estimates roughly 1,095 domains hosted across 23 addresses. Domain count is an imprecise measure: one server can host hundreds of low-value sites, parked domains can inflate the total, and a valuable application can use one domain. Still, the concentration on a small number of addresses is consistent with shared hosting rather than a broad cloud-compute estate.
Voice is more visible. A-Connect's current price page offers a geographic number for a RUB1,600 one-time payment and RUB350 per month. A call-centre module is RUB450 monthly, call queues RUB450, call recording RUB350, extra internal numbers RUB20 each and extra recording storage RUB450 per gigabyte. The service homepage promises rapid self-provisioning, number choices in 40 large Russian cities and support. Its call-rate table lists per-second billing and rates for Russian destinations. The product is a genuine small-business communications service, not merely a conceptual capability.
Interconnection may be the most strategically important but least financially transparent line. Historical company descriptions say Altura operated Saratov-IX, and public routing policy distinguishes upstreams, customers and exchange relationships. A local exchange or hub can lower latency and reduce paid transit for participants. It can also create a reason for regional networks to maintain a relationship with Altura even if they buy national transit elsewhere. But there is no current public port list, traffic curve, service-level schedule or fee table for Saratov-IX. The historic claim that 70% of the city's providers were connected should be treated as an old company assertion, not a current market-share statistic.
The portfolio logic is cross-subsidy. A business connection can lead to a static address, voice lines, hosted equipment and support. A network customer can buy transit and use local interconnection. A hosting customer may consume addresses and call support but little transit. The risk is that management cannot see which line creates cash. Revenue can grow while value falls if a low-margin resale or one-off equipment sale expands faster than the recurring services. Conversely, a modest revenue line can create value if it anchors a sticky customer or lowers the cost of the wider network.
Network evidence: real, local and limited
The strongest evidence of operating substance is AS44158. BGP.Tools shows the ASN as active, registered in November 2007, and originating 95.141.192.0/20, a block of 4,096 IPv4 addresses. Its current view shows three upstreams: RETN, Mezhdugorodnyaya Mezhdunarodnaya Telefonnaya Stanciya and Kvant-Telecom. It also observes four downstreams and eight peers. A public RPKI certificate record links 95.141.192.0/20 to AS44158, supporting route-origin authorisation for the visible aggregate.
AS44552 adds another layer. IPinfo's AS44552 profile reports 6,144 unique IPv4 addresses across three /21 blocks, no observed IPv6 addresses, two upstreams or peers in its current graph, and no downstream networks. Its traffic classification shows a pronounced day-night rhythm, which is more consistent with end-user or access activity than with an always-on hosting-only network. Router geolocation points mainly to Saratov, with two addresses associated with Pugachyov. Those signals support a regional access role, but geolocation and traffic classification are probabilistic.
Together, the two ASNs expose as many as 10,240 routed IPv4 addresses under the same RIPE organisation, before considering overlaps, assignments and operational restrictions. AS44158 appears to be the hub: it connects to larger upstreams and several smaller networks, while AS44552 takes reachability from AS44158 and ER-Telecom in the current IPinfo view. This is more than a paper allocation. Recent public probes reached addresses in both networks from Moscow in roughly the low-to-mid teens of milliseconds, and the route tables remain visible.
The footprint is nevertheless small. AS44158 originates one IPv4 aggregate. The public views show no originated IPv6 prefix for either ASN. No IPv6 in these datasets does not prove that Altura uses none inside its network, but it does indicate no visible dual-stack reach for the two ASNs. That matters because IPv4 scarcity can support present value while IPv6 absence can increase future technical debt. Customers increasingly expect dual-stack service, and an operator that relies entirely on IPv4 must keep managing address pressure, translation and reputation.
Routing records also require caution. Registry policy can be old, observed BGP paths can change by vantage point, and a downstream relationship is not automatically a paid transit contract. BGP.Tools currently reports four downstreams; IPinfo reports five. The difference is useful in itself: it shows why customer counts cannot be read straight from an AS graph. What can be said is that several independent networks choose paths through or alongside AS44158, which creates an operating surface where Altura can sell reachability, interconnection or support.
Network quality is not established by having an ASN. The evidence needed for that judgment would include packet loss, latency by destination, availability over time, route-security practice, incident history, capacity headroom and repair intervals. Public records establish control and reach, not service quality. They are necessary evidence for an operator claim, but not sufficient evidence for an economic moat.
Revenue is not the same as value
The latest accounts create the central tension. RBC Companies reports 2025 revenue of RUB10.417 million, down from RUB11.689 million in 2024. That is a decline of about 10.9%. Cost of sales was RUB10.347 million, leaving an implied gross spread of RUB70,000. Net profit was RUB16,000. The gross margin was about 0.67%; the net margin was about 0.15%.
Those are not normal cushions for a network operator. A modest router, server, storage repair, licence renewal, legal dispute or uncollected invoice can exceed the entire annual profit. A price concession of one percentage point across the revenue base would be about RUB104,000, more than the reported gross spread. A single RUB100,000 bad debt would have the same effect. The accounts may use broad cost classifications, and a small private company can report economics differently from a listed carrier, but classification cannot erase the basic fact that reported revenue and reported cost were almost equal.
The eight-person staff figure implies revenue of roughly RUB1.30 million per employee per year, or about RUB108,500 per employee-month. That is revenue, not pay. It must cover wages, payroll costs, upstream service, facilities, power, equipment, software, regulatory work and all other costs. The ratio signals a labour-intensive, low-revenue operation, especially if staff provide network monitoring and customer support outside office hours.
Revenue measures what customers were billed under accounting rules. Value measures the future cash that assets, contracts and capabilities can produce. Altura may possess value not expressed by one year's profit: established geographic number access, customer relationships, two long-lived ASNs, registry standing, an RPKI-covered aggregate, local routing knowledge and scarce IPv4 space. It may also own equipment whose replacement value exceeds its book value. None of these becomes distributable value unless the company can preserve control and earn cash from it.
IPv4 illustrates the difference. A 2026 market survey from IPv4Center gives an indicative average of about US$25 per address, while IPv4 Global's market reports show that prices vary by block size and moved materially through 2025 and 2026. Applied mechanically, US$25 would put the visible AS44158 /20 near US$102,400 and 10,240 addresses near US$256,000. Those are not valuations of Altura. Some addresses may be assigned to customers, operationally indispensable, non-transferable in the assumed form, reputationally impaired or subject to legal and registry review. A sale could destroy service revenue, and transaction prices can differ sharply from an average.
The thought experiment is still informative. At the Bank of Russia's 8 July 2026 official US dollar rate of RUB76.12, Altura's 2025 revenue was roughly US$137,000. The notional address figure can therefore be comparable with, or larger than, a year of revenue even while earnings are negligible. This explains why address stewardship matters. It also explains why a buyer should separate an operating-company valuation from a resource valuation and verify exactly which resources are transferable. RIPE NCC transfer rules allow qualifying IPv4 transfers but impose process and holding-period conditions.
The preferable outcome is not liquidation of scarce resources. It is to use them to raise service value: sell static-address business access, host applications that require public reachability, support smaller networks and charge for accountable allocation. If Altura cannot earn more from operating the space than its risk-adjusted alternative value, the business has an allocation problem. If it can, the resource supports a durable niche even without cloud scale.
Unit economics: a very small spread must do too much
Public data does not disclose customer count, traffic volume, average revenue per account, churn, transit commits or product-level margins. It is therefore impossible to calculate a verified contribution margin per circuit or PBX account. It is possible to set boundary conditions.
First, the voice prices are low. A geographic number at RUB350 per month produces RUB4,200 a year before call charges. One hundred such base subscriptions produce RUB420,000, only about 4% of 2025 company revenue. If the entire RUB10.417 million came from that fee, it would require roughly 2,480 account-equivalents active for a year. That is not a customer estimate; Altura also sells calls, modules, connectivity and other services. It shows that the base fee alone cannot carry the cost structure without either substantial volume or profitable add-ons.
A call-centre account taking the number, call-centre module, queues and recording would pay roughly RUB1,600 per month before calls, storage and extra extensions, using listed component prices. One hundred such bundles would produce about RUB1.92 million annually, still less than one-fifth of company revenue. The economic variables are therefore active account count, call minutes, termination cost, number rental, fraud loss, support time and module attachment. A low advertised monthly fee can acquire customers, but support-heavy small accounts can be negative contributors.
Second, business connectivity can produce much higher revenue per account but creates concentration. If an average business circuit generated RUB10,000 per month, about 87 full-year account-equivalents would equal total company revenue. At RUB50,000, only about 17 would. These are scenarios, not claims about Altura's prices. They show how quickly a small revenue base can become dependent on a few institutions, buildings or downstream networks.
Third, the reported gross spread leaves no visible tolerance. A one-percentage-point improvement in cost of sales would add roughly RUB104,000 before other effects, more than six times reported net profit. Conversely, a one-point deterioration would erase it. That means pricing discipline, contract indexation and cost allocation matter more than headline subscriber growth. Altura's A-Connect site said prices would rise from January 2026 after the VAT rate increased to 22%. Passing through tax protects nominal revenue but does not itself improve real margin.
Fourth, network products have step costs. Transit may be bought on committed capacity; once traffic exceeds a commit, the marginal cost changes. A new customer may fit on existing ports and produce strong contribution, or may trigger a router, access build or larger upstream commitment. Hosting has similar thresholds for power, storage and replacement servers. Voice has termination and number costs that track usage more directly. Product-level accounting is essential because the same RUB1 of revenue can have very different capital and support requirements.
The unit that matters is not a megabit or a telephone number in isolation. It is a customer-location-product bundle after upstream, access, support, failure, billing and compliance costs. Altura creates value when that bundle produces a repeatable cash contribution without requiring disproportionate bespoke work. The 2025 statements do not demonstrate that this is happening at an adequate rate.
Cost and capital needs
Altura's visible cost stack has at least seven layers. It must buy upstream reach or transport; maintain routers, switches, servers and power protection; pay for premises, power and access; employ people who configure and repair services; maintain billing, voice and security software; meet communications obligations; and pay registry and numbering costs. A company can own some components and rent others, but it cannot avoid the economic load.
Companium's public financial summary reports RUB5.3 million of fixed assets for 2025. The composition and age are not public. If these are primarily network and facility assets, replacement need is material relative to revenue. A simple five-year replacement allowance on even half that amount would be RUB530,000 a year, more than seven times the reported gross spread. That is not a depreciation forecast: the assets may include long-lived items, may already be depreciated, and replacement can be staggered. It is a stress test showing that current reported margin does not visibly fund routine renewal.
Registry cost is small in carrier terms but large beside Altura's profit. The RIPE NCC 2026 billing procedure sets an annual contribution of EUR1,800 per existing LIR account, before any applicable resource charges. At the Bank of Russia's 8 July euro rate of RUB86.89, that service fee alone was about RUB156,000. It was more than twice Altura's 2025 implied gross spread and almost ten times reported net profit. The fee is presumably already embedded somewhere in operating costs; the comparison demonstrates fixed-cost scale, not an additional deduction from profit.
Equipment inflation raises the renewal hurdle. CNews reported that several Russian telecom-equipment makers increased prices by 10% to 30% from January 2025, citing components, exchange rates, logistics and wages. A small operator has less purchasing leverage and a smaller pool of spares than a national carrier. Delaying replacement protects cash temporarily but increases outage and security risk. Buying used equipment lowers entry cost but can increase failure, support and power costs.
Power is another asymmetry. A large facility can negotiate supply, optimise cooling and spread backup systems across many racks. A small room still needs batteries, perhaps a generator, cooling and monitoring. One lightly used server can consume almost as much operational attention as a full one. Altura's best capital strategy is therefore density: fill existing access routes, numbers, racks and support capacity before adding another fixed layer.
Working capital can be equally important. Upstream and registry bills are due on schedule; business customers and public institutions may pay later. Voice fraud can create wholesale charges before the retail account is collected. Imported or scarce spares may require prepayment. With only RUB16,000 of reported profit, a cash reserve or committed credit line is more relevant to continuity than the income statement alone. Public sources do not disclose that liquidity.
Upstream and supplier dependence
Altura is not an independent internet. It buys reach from larger networks and depends on them for paths outside its regional sphere. Current BGP observations list RETN, Mezhdugorodnyaya Mezhdunarodnaya Telefonnaya Stanciya and Kvant-Telecom above AS44158. AS44552 is observed through Altura and ER-Telecom. Multiple upstreams reduce single-provider failure risk, but the protection depends on physical diversity, sufficient spare capacity and routing policy. Three contracts that share one duct or one building entrance are not three independent paths.
The bargaining balance is uneven. Upstreams can spread infrastructure and compliance costs across far more traffic. Altura's purchase is small to them, while the upstream bill can be material to Altura. A minimum commit, port charge or transport circuit can keep cost fixed even if customer traffic and revenue fall. Conversely, competition among upstreams can help Altura negotiate if it has credible alternative paths and a useful local handoff.
Interconnection is the counterweight. Traffic exchanged locally does not need to consume the same amount of paid upstream transit, and low-latency local routes can improve customer experience. AS44158's observed connections to several regional networks are therefore economically meaningful. Yet the savings depend on traffic balance and useful destinations. Peering with a small access network saves little if most user demand goes to large content platforms reached elsewhere. A local exchange becomes more valuable when it attracts caches, cloud on-ramps, government networks and major access providers.
Voice has its own supplier chain: number resources, termination carriers, SIP software, payment processing and customer devices. A-Connect's rates distinguish calls associated with Altura's numbering resources from parked or other numbers, suggesting that resource control affects cost. Fraud controls and wholesale settlement are crucial because an attacker can generate expensive traffic quickly. The customer terms place credential-security responsibility on the subscriber, but reputational and collection risk can still return to the operator.
Hardware and software supply is exposed to geopolitical friction. Replacement optics, line cards, storage, batteries and security subscriptions may have foreign content even when bought from a Russian vendor. Longer logistics chains increase inventory needs. An eight-person operator cannot maintain deep expertise across every discontinued platform. Standardising equipment and holding critical spares may therefore create more value than expanding a long tail of products.
Registry dependence is also cross-border. RIPE NCC invoices are in euros, and its sanctions transparency report explains that the Dutch association must comply with EU restrictions and that banking checks can affect invoicing and payment. That report is not evidence that Altura is sanctioned. It is evidence that Russian resource holders face a payment and documentation channel outside domestic control. Membership continuity, accurate records and early payment are therefore operating tasks, not administrative trivia.
Customer and market concentration
No reliable public source gives Altura's current customer count or revenue concentration. The absence is consequential. With RUB10.417 million of revenue, one RUB1 million annual account would represent 9.6% of sales; a RUB500,000 account would represent 4.8%. A handful of business circuits, hosting customers or wholesale relationships could therefore determine the year.
The network graph reveals a plausible concentration surface. BGP.Tools observes four current downstream networks beneath AS44158, while IPinfo's view includes five. If any are paying customers, each may be technically and commercially important. But a downstream can be affiliated, settlement-free, historical or visible only from some collectors. The graph cannot be converted into a revenue table.
Public procurement data is similarly limited. A company-data service reports four historical government procurements totalling about RUB1.022 million. The amount is meaningful beside one year of current revenue, but the record does not show that these contracts are active, recurring or concentrated in one buyer. It supports the possibility of institutional customers rather than a conclusion about present dependency.
The product mix may diversify billing without diversifying risk. Ten services sold to one university, building owner or regional network remain one counterparty. Hosting, voice and access can also fail together if they share a site, power feed or upstream. True diversification means different customers, industries, locations, access paths and payment cycles.
Saratov itself is a market dependency. Local knowledge and interconnection are Altura's advantage, but economic weakness or consolidation in the region can shrink the addressable base. A national provider can tolerate a slow year in one city; Altura cannot diversify geographically without incurring new support and transport costs. Its Moscow-number offer expands product reach, but the operating centre and network evidence remain strongly Saratov-focused.
Concentration is not always bad for a small operator. A few sophisticated customers can reduce acquisition and billing costs, and long contracts can justify dedicated infrastructure. The danger appears when prices do not compensate for bargaining power, bespoke support and renewal risk. Altura should prefer contracts with clear indexation, installation recovery, minimum terms and defined service boundaries. Revenue bought through underpriced custom work would worsen the thin-margin problem.
Competition and substitutes
Altura faces different substitutes in each product, and no single competitor defines the market.
For household broadband, the substitutes are straightforward. A current 101internet marketplace page displays no available Altura household tariffs while showing Saratov offers from TTK, MTS and Rostelecom around RUB500 to RUB600 per month for 100 to 200 Mbit/s, with promotional variation. The page is a commercial aggregator and prices are address-dependent, so it is a market signal rather than a tariff audit. It nevertheless shows the problem: a mass-market buyer can compare familiar brands, bundles and installation offers quickly. Altura has little reason to chase that segment unless it controls an underserved building or access route.
For business access, MTS, Dom.ru, Rostelecom, regional operators, mobile broadband and fixed-wireless links are substitutes. MTS's Saratov business page combines fixed and mobile access with a much broader digital-service catalogue. Dom.ru's business offer targets offices and provides individually quoted service. A national operator can bundle mobile lines, cloud applications, security and support. Altura's answer must be faster local resolution, useful routing, flexible installation or a site the larger provider serves poorly.
For voice, substitutes include a mobile corporate plan, a national SIP provider, a hosted collaboration platform, call-centre software and an on-premises PBX. A-Connect's low monthly prices and local numbers are attractive, but the functions listed on its site—menus, queues, recording, extensions and statistics—are widely available. The service wins when local numbering, simple setup and reachable support outweigh brand, integrations and geographic scale. It loses when customers want deep CRM integration, unified messaging, global numbers or a large compliance organisation.
For hosting, Russian data-centre and cloud providers such as Selectel, VK Cloud, Cloud.ru and Yandex Cloud offer automation, elastic resources and larger facilities. Selectel's public pricing makes the comparison immediate and has also announced 2026 price changes across services. Altura cannot profitably duplicate every cloud feature. It can offer local hands, a known engineer, simple colocated equipment, legacy workload support or a short path to Saratov users. Those advantages apply to a narrow set of buyers.
For interconnection, the substitute is paid transit or another exchange. If national backhaul is cheap and latency is acceptable, a smaller network can avoid a local exchange. Content-delivery caches can also move popular traffic closer without broad local peering. Altura's exchange role is defensible only if the participants and traffic are useful, the facility is reliable and the commercial terms beat the alternative.
For scarce addresses, customers can use carrier-grade NAT, lease IPv4 space, buy addresses, adopt IPv6 or purchase a cloud service that includes public reachability. Altura's addresses are valuable because those substitutes have costs and limitations. That value will decay if IPv6 adoption removes demand faster than scarcity raises price, or if abuse reputation makes the space harder to use.
The best strategy is not to deny substitution. It is to identify the cases where switching cost reflects real operational benefit: a local number customers know, a static address tied to allowlists, a route integrated with several sites, equipment requiring local hands, or a smaller provider that values one regional relationship. Those are monetisable frictions. Brand loyalty alone is too weak.
Regulation, geopolitics and operating risk
Communications regulation allocates obligations to the licensed operator, not merely to the network device. Russia's communications law makes licence records public and states that licence rights cannot simply be transferred to another person. That increases the value of a clean operating entity but also binds compliance to Altura's legal continuity. A buyer cannot assume that buying equipment or addresses alone transfers every right to serve customers.
Data retention is a direct cost category. Russian rules introduced in 2018 require communications operators to retain specified user communications for periods of up to six months. Storage, secure access, integration and audit work scale differently from revenue. A large carrier spreads the fixed layer across millions of accounts; a small operator cannot. Even where equipment is supplied or prescribed, staff time, rack space, power and failure handling remain real.
Filtering obligations are not abstract for Altura. A 2013 Saratov court report described litigation in which Altura said its equipment could not block particular web content in the requested manner; the court left the order in place. The technology and legal framework have changed since then, but the episode illustrates a persistent economic point: regulatory requirements can force an operator to alter equipment and processes without creating new billable demand.
Tax changes also flow through pricing. A-Connect's price and call-rate pages announced increases from January 2026 in response to the VAT rise to 22%. Passing the tax through protects the operator from absorbing the full increase, but customers see a higher bill and can compare substitutes. With a reported gross margin below 1%, Altura has little ability to absorb such changes as a retention gesture.
Sanctions and geopolitical restrictions work through several channels even without a company-specific designation. They can complicate euro payments, vendor support, software licences, hardware imports, bank settlement and counterparty reviews. RIPE NCC says it freezes registration, rather than use, when applicable sanctions affect a resource holder, and notes that banking institutions consider US restrictions. For Altura, the risk is operational friction and reduced optionality: an address transfer, equipment purchase or cross-border service can take longer or fail.
Domestic network policy can also alter customer experience independently of Altura's engineering choices. Blocking, traffic-management mandates or regional mobile restrictions can make users blame the retail provider for outcomes it does not fully control. The operator still bears support calls and reputation loss. Clear incident communication and path-level monitoring are economic controls because they shorten diagnosis and protect trust.
Cybersecurity risk is amplified by the product mix. Hosting can attract abuse complaints. Public addresses can acquire poor reputation. Voice accounts can be hijacked for toll fraud. An exchange or transit service can propagate bad routes. A small team must run access controls, backups, patching, abuse response, route filters and customer authentication while maintaining service. One serious incident can consume more cash and attention than a year of profit.
Finally, key-person risk is unusually visible. The 95% owner is also the general director, and his network contact history extends back many years. That continuity is an advantage when relationships and tacit knowledge matter. It is a weakness if routing, supplier negotiation, licence knowledge or customer trust cannot be transferred to other staff. Succession and documented authority are therefore part of enterprise value.
Unofficial signals: useful questions, not verdicts
Public customer feedback is sparse and old. 2IP's Altura profile displays eight reviews, a rating of 2.91 out of five, 9,153 measurements and an average ping of 56 milliseconds. A 2016 reviewer praised low gaming latency and the provider's interconnection; a 2020 reviewer complained about unstable latency, dropouts and slow support. Earlier reviews are also mixed. Eight self-selected reviews over many years cannot establish current service quality, but the disagreement identifies the right test: low local latency is valuable only if availability and support are consistent.
A current 2GIS listing shows one rating and a score of one. A single unexplained rating has almost no statistical weight. It is worth monitoring, not repeating as fact about performance. The same listing confirms that the company remains locally discoverable at the Kiseleva Street address and operates by prior appointment.
The household marketplace's absence of current Altura tariffs is another weak but useful signal. It may mean Altura is not actively acquiring mass-market subscribers through that channel, has limited coverage, sells directly, or simply has stale marketplace data. It aligns with the older review claim that the company focused on organisations and smaller providers, but it does not prove a deliberate business strategy.
Historical marketing material claimed the region's largest data centre, more than 500 hosted sites and links to 70% of city providers. The first two claims may have been plausible at the time, and IPinfo now estimates more than 1,000 hosted domains on AS44158. Domain count is not site count, facility size or revenue. The 70% claim lacks a current participant list. These claims should be converted into diligence requests: rack count, occupied power, customer count, traffic, current participants and contracted fees.
Routing measurements are stronger than reviews but still incomplete. Current probes reach both ASNs, and AS44158 has multiple visible upstreams. That supports ongoing activity. It does not reveal whether failover works during an outage, whether paths are physically diverse, or how often customers experience loss. An acquirer or large customer should request independent monitoring over weeks, not a single successful traceroute.
The proper use of these signals is triangulation. Old praise for local routes matches the observed network graph. Complaints about instability match the generic risk of a thinly staffed regional operator but do not prove a current problem. The lack of mass-market tariffs matches a business-oriented product set. None should override contracts, invoices, traffic records or measured service levels.
What would change the judgment
The present judgment is cautious: Altura controls useful regional infrastructure but has not shown that it converts that usefulness into an adequate financial return. Several facts could improve or weaken that view quickly.
First, product-level gross margin would resolve the largest uncertainty. If connectivity, voice and hosting each generate healthy contribution but one non-recurring item distorted 2025 cost of sales, the reported 0.67% spread understates earning power. If every line is thin, the business needs repricing or consolidation.
Second, customer concentration and contract duration would reveal resilience. A broad base of auto-paying voice accounts plus multi-year indexed business circuits would be positive. Dependence on one institution, reseller or downstream with an annual renewal would be negative.
Third, proof of physical path diversity would make the multi-upstream graph more valuable. Maps, contracts and failover tests should show different entrances, carriers and power domains. Logical diversity over one conduit is a false comfort.
Fourth, the exact control and transfer status of each IPv4 block matters. Clean registration, documented assignments, low abuse incidence and an active RPKI posture support value. Unclear customer rights, stale records or reputational blocks reduce it. Publishing ROAs and visible IPv6 routes for the whole active footprint would strengthen the operational case.
Fifth, a capital plan would test continuity. The age and support status of routers, servers, batteries, cooling and storage should be matched against replacement cash. RUB5.3 million of fixed assets is only useful if the equipment can keep delivering and if replacement does not consume more cash than the business generates.
Sixth, current exchange evidence would distinguish legacy reputation from an active platform. A participant list, traffic statistics, port capacity, outages and fees would show whether local interconnection is a present business or a historical label.
Seventh, A-Connect operating data would show whether the visible product can scale. Active numbers, average modules per customer, call minutes, fraud losses, churn, support tickets and acquisition cost would reveal whether RUB350 entry pricing leads to profitable relationships.
Eighth, upstream contracts would show bargaining and concentration. Price per committed megabit, burst terms, physical routes, currency clauses and renewal dates determine whether traffic growth creates margin or merely a bigger bill.
Ninth, liquidity and debt data would clarify survival capacity. A cash reserve, low debt and unused credit could offset weak reported profit. Overdue payables or reliance on owner funding would make the network more fragile than its routing visibility suggests.
Tenth, staff coverage and succession would determine whether eight people are a lean team or a single point of failure. Named responsibility for network operations, security, voice, billing, sales and after-hours response would increase confidence.
Eleventh, a verified service record would settle the local-support thesis. Availability, mean time to repair, customer retention and independently measured latency would show whether Altura's smaller scale produces responsiveness or under-resourcing.
Judgment
Altura is economically interesting because the network is more substantial than the income statement and the income statement is more sobering than the network map. The company has survived for decades, remains a RIPE NCC member, operates a live voice product, controls visible routing resources and occupies a useful place in Saratov's network fabric. Those are real advantages. A shell could not reproduce them instantly.
Yet survival and value creation are different. RUB10.417 million of revenue is small for an operator responsible for connectivity, hosting and telephony. RUB70,000 between revenue and cost of sales is not enough evidence of renewal capacity. RUB16,000 of net profit offers virtually no downside protection. The 10.9% revenue decline makes the thin spread more concerning because fixed network costs do not fall neatly with sales.
Altura's opportunity is to charge for control, not capacity alone. Local numbers, static addresses, direct support, regional routes and hands-on infrastructure can justify a premium for the right customer. The company should avoid products where the buyer compares only megabits, virtual CPUs or bundled entertainment. It should measure contribution by customer and recover bespoke access and support costs explicitly.
Who benefits if that strategy works? Saratov businesses gain a local alternative and a supplier capable of combining communications services. Smaller networks gain an interconnection and upstream option. Altura's staff and owners gain recurring income from infrastructure already in place. The regional market gains diversity. Who bears the downside? Altura absorbs fixed costs, compliance work and first-line outages; customers bear continuity risk; owners bear capital calls; and downstream networks bear route dependency.
The verdict is therefore neither that Altura is too small to matter nor that scarce addresses make it valuable regardless of earnings. It matters precisely because it controls a modest local operating surface. Its value will persist only if that surface produces cash after transit, labour, equipment, registry, regulation and failure costs. The public evidence proves relevance. It does not yet prove an adequate return on that relevance.