Summary

  • The apparent SkyX retail offer creates a sharp economic promise. Residential plans start at 150 somoni a month for 30 Mbps, while business plans run from 12,000 somoni for 100 Mbps to 40,000 somoni for 500 Mbps. That business premium can fund reliability only if it buys measurable uptime, rapid restoration, protected capacity and route diversity; the public offer does not yet disclose those terms.
  • Network records show real but transitional capability. "Shabakai oson" became a RIPE NCC member in January 2026, received a large IPv6 allocation and on 25 June received a permanent transfer of 1,024 IPv4 addresses from Orien Invest. Yet the IPv4 route is still registered to originate through Orien Invest's AS48001, the company's own short-lived AS202726 is no longer present in the live RIPE database, and no current public evidence establishes independent peering or physically separate international paths.
  • Tajikistan offers both demand and danger. Fixed-broadband household penetration was only about 8% in March 2024, official figures say fixed prices fell sharply during 2025, and businesses increasingly depend on digital payments and remote services. The same reforms intensify price competition, established rivals already sell fibre access, Starlink arrived in February 2026, and power, wholesale, support and regulatory costs can absorb the margin before a young operator has enough customers to spread its fixed costs.

The first product is an outage someone cannot afford

The economic reason to own network capability is not that an autonomous network number looks impressive. It is that a provider which controls more of the path can make a stronger promise, diagnose failure faster and avoid paying another company for every operational decision. The customer buys the result of that control only when an outage is costly enough to justify a premium. A pharmacy that loses payment connectivity, a laboratory that cannot transmit results, a hotel whose guests cannot work, or a distributor cut off from its ordering system has a reason to pay for continuity. A household comparing streaming plans usually has much less appetite for an expensive second path.

That difference sits in plain view on the SkyX tariff page, the most plausible public-facing commercial offer associated with "Shabakai oson." SkyX displays unlimited household plans at 150 somoni a month for 30 Mbps, 250 somoni for 50 Mbps and 400 somoni for 70 Mbps. Its business plans are dramatically dearer: 12,000 somoni for 100 Mbps, 20,000 for 200 Mbps and 40,000 for 500 Mbps. Each business plan advertises a Wi-Fi 6E device and city Wi-Fi as benefits. The site separately lists network protection at 49 somoni, 100 GB of cloud storage at 99 somoni and support at 149 somoni a month.

The price structure is more revealing than the catalogue. Household access is offered at roughly five to 5.7 somoni per advertised Mbps each month. The corresponding business figure is 80 to 120 somoni. A business byte is not intrinsically worth 14 to 24 times a household byte. The premium has to pay for a different service commitment: dedicated rather than heavily shared capacity, a meaningful restoration target, better customer equipment, priority escalation, an accountable local technician, stronger route control or some combination of those things.

SkyX's public page does not say whether the business rate is dedicated internet access, whether speeds are committed or merely maximum, whether upload and download capacity are symmetric, or what service credits follow a missed target. It gives no latency, packet-loss, contention, installation, contract-term, fair-use or restoration data. The optional support plan promises round-the-clock online support, a home visit and equipment configuration, but the business plans do not publish a separate support standard. This does not show that the service is weak. It shows that the economic justification for the premium remains unproven in public.

The immediate management choice is therefore not simply whether to build more fibre. It is whether to allocate scarce capital to the specific controls that turn a large tariff into customer value. A second upstream bought over the same physical cable is not redundancy. A spare router without an engineer who can reach the site is not resilience. An impressive maximum speed without congestion management is not continuity. The provider creates value only when it can convert network ownership into fewer lost customer hours, then capture part of that avoided loss as recurring margin.

A company and a brand still coming into focus

The operating boundary needs care because the company is young in public network records and the SkyX site does not state the legal contracting name. The RIPE NCC member list identifies "Shabakai oson" LLC as a Tajikistan-based local internet registry. The live organisation entry says the membership was created on 13 January 2026 and is maintained under the name SKYX-MNT. A cached version of the company's former autonomous-network entry also pointed to skyx.tj, and the site presents home, business and city internet under the SkyX name.

Together those facts make SkyX the most plausible commercial expression of the company. They do not settle the legal question. The SkyX about page describes a new Tajik information-technology and telecommunications company whose mission is to provide citizens, organisations and entrepreneurs with reliable digital infrastructure. It does not identify "Shabakai oson" in its visible text, publish a corporate registration number, name an owner, provide financial accounts or set out customer terms. This article therefore uses SkyX tariffs as apparent commercial evidence and does not assume that every SkyX contract, asset or employee belongs to the named company.

The timeline is consistent with a launch or separation. RIPE created the company membership and a large IPv6 allocation on 13 and 14 January. A company-specific autonomous number, AS202726, appeared on 15 January in public mirrors with planned routing relationships to TT Mobile and Tojiktelecom. By 10 July, however, a live query for AS202726 returned no entry. Historical data is useful for reconstructing intent, but it is not a licence to describe two current upstreams.

The more consequential change came in June. The RIPE NCC transfer list records a permanent policy transfer of 185.208.96.0/22 from Investment and Production Company "Orien Invest" LLC to "Shabakai oson" on 25 June 2026. A /22 contains 1,024 IPv4 addresses. The current allocation entry names the company and shows the transfer date. This is more substantial evidence than a marketing claim because routable IPv4 space is scarce and operationally useful.

Even so, holdership is not the same as independent operation. A live RIPE search for the route still lists AS48001, the Orien Invest network, as the origin. The route entry was updated on the transfer date and is maintained by SKYX-MNT, suggesting an orderly handoff or continuing network arrangement. It does not reveal whether the companies share ownership, whether Shabakai pays Orien for transit, whether customers have migrated, or whether the route will move to another network. The transfer type is policy, not merger or acquisition. The consideration is not public.

This transitional boundary is economically central. If Shabakai owns customer contracts and address space but depends on Orien's routing, it has gained an asset without yet proving full path control. That may be a sensible low-risk launch sequence. It may also leave one of the largest operational dependencies outside its direct control. The judgment changes materially when the company publishes the legal relationship among Shabakai, SkyX and Orien, the assets included in the handoff and the agreements that protect service if the parties disagree.

What the business is actually trying to sell

SkyX's catalogue points to three revenue pools. The first is recurring household broadband delivered through Wi-Fi links, fibre to the building or GPON fibre to the home. The second is business connectivity sold at a much higher monthly rate. The third is a set of low-price add-ons: security, storage and support. City plans and free trial periods appear designed to place the brand in front of users before asking them to take a full household subscription.

The model can work because the same local access plant supports several products. A fibre feeder into a dense building can serve many flats, a shop and a small office. Once the splitter, building access, backhaul and support route are in place, each additional paying line can carry an attractive contribution. Security, storage and support can raise revenue per customer without another trench. City Wi-Fi can market the fixed product and reduce churn if access is genuinely useful away from home.

The model can also fail in several distinct ways. A cheap household line may never repay installation if the customer leaves after a few months. A free trial can fill the support queue with users who never convert. City Wi-Fi can become a maintenance burden with little direct revenue. Cloud storage and network protection can create data, licensing and incident-response obligations that exceed their 49 or 99 somoni price. A promised home visit can consume a large share of a 149 somoni support fee once transport and technician time are included.

Business access is supposed to repair those economics. One 12,000-somoni customer produces the same headline monthly revenue as eighty households on the entry plan. One 40,000-somoni contract equals one hundred top-tier household subscriptions. The temptation is to use a few corporate accounts to subsidise a broad residential footprint. That works only if the business customers are diversified and the expensive service does not require bespoke construction, reserved international capacity and constant manual support.

No public source discloses Shabakai's subscriber count, active lines, billed revenue, average contract life, gross margin, churn, collection period or bad debts. There is no evidence of a large corporate contract, public tender award or named customer. That means revenue cannot be inferred from tariff cards. The page shows what the company would like to charge, not what customers actually pay, how many can be connected or whether promotions and negotiation reduce the realised rate.

The cleanest business model would be geographically narrow at first. Dense Dushanbe buildings and business clusters offer more customers per metre of access fibre, shorter repair trips and greater scope to reuse backhaul. The public site does not provide a coverage map, so even that remains a strategy hypothesis rather than a description of deployed plant. Expansion should follow proven take-up and payback by micro-area. National ambition without local density would turn every new district into a fresh fixed-cost problem.

The residential price attack

SkyX enters a market where fixed prices have recently moved down. Its 150-somoni 30 Mbps household plan is aggressive beside established public offers. Babilon-T lists 30 Mbps at 240 somoni, 50 Mbps at 320, 70 Mbps at 550 and 100 Mbps at 770. Tcell's TezNet displays 30 Mbps at a regular 199 somoni and 50 Mbps at 299, with temporary promotional prices of 149 and 249. MegaFon Tajikistan advertised Dushanbe fixed plans in September 2025 from 150 somoni for 15 Mbps to 280 somoni for 30 Mbps, with different speeds for local and selected services.

These are advertised rates, not a controlled service comparison. Coverage, installation, hardware, traffic treatment, upload speed and promotional periods differ. SkyX distinguishes ordinary internet speed from faster local resources and social networks on household plans. It does not say whether the 30, 50 and 70 Mbps figures are guaranteed. Still, the direction is clear: the entry plan attempts to offer more headline capacity for the money than several visible alternatives.

Low price can buy share, but it does not create value by itself. At 150 somoni, a customer who needs a new optical terminal, router, drop cable and technician visit may be unprofitable for months. If the company absorbs all connection costs, payback depends on retention. If it charges installation separately, the effective first-year price is higher than the card suggests. Neither connection charge nor equipment ownership is stated. A customer acquired on price can also leave on price when an incumbent responds.

The regulator says that response is already happening. In a first-quarter 2025 industry update, Tajikistan's Communications Service said lower imported-internet costs allowed fixed tariffs to fall by an average of 55%, while plans below 5 Mbps were discontinued. Its full-year account later said plans below 30 Mbps had been removed and measured fixed speed reached 40.42 Mbps. The thresholds refer to different points in the policy year, but both show a market moving quickly towards more capacity at lower prices.

This is good for demand and hard on returns. The World Bank's Digital Foundations appraisal estimated fixed-broadband household penetration at only 8% in March 2024. The runway is large if falling tariffs bring new homes online. Yet every provider sees the same runway. An operator that builds on the assumption that 2026 prices will remain unchanged could find its revenue per connection falling before the fibre is fully depreciated.

Affordability still constrains the ceiling. A 2026 UNESCO review, drawing on ITU data, put the 2023 fixed-broadband basket at 5.33% of gross national income per person, well above the ITU's 2% affordability target. SkyX's 150-somoni entry point may broaden access, but many households will still compare it with mobile data or share one connection. The company needs high take-up within covered buildings, not merely a national statistic about unconnected homes.

Why the business tariff is the real thesis

The business plans are not a price attack. They sit in an established Tajik corporate range. A 2024 office-connectivity guide published by the Communications Service with MegaFon cited 600 somoni for 5 Mbps and 100 GB, 4,000 somoni for 30 Mbps and 1,000 GB, and a broad range up to 40,000 somoni for larger offices and stricter needs. SkyX's 12,000 to 40,000 somoni pricing is therefore plausible for the market. What remains unclear is what specification earns it.

For an SME, the willingness to pay depends less on nominal speed than on the cost of interruption. A 30-person services company may need only modest average throughput but cannot tolerate losing cloud applications or payment access for half a day. A video studio may need high sustained upload capacity. A bank branch may require routing, security, logging and a second circuit. A medical laboratory may value rapid repair more than peak download speed. A single tariff ladder based only on Mbps cannot express all of those needs.

That creates both an opportunity and an underwriting problem. Shabakai can price a managed outcome above commodity access if it measures the customer's exposure and designs the circuit accordingly. It can also accept a low-margin commitment by promising "business internet" without modelling peak traffic, route failover, equipment replacement and field response. The larger the contract, the more dangerous the second mistake becomes because a service credit or customer loss can erase months of margin.

The address-space transfer could support business value. Stable public IPv4 addresses are useful for virtual private networks, hosted services, remote access and customers that cannot complete an IPv6 transition. A /22 gives room for customer addressing, infrastructure and growth. But addresses are not resilience. The current route remains tied to AS48001, and the company publishes no policy for static addresses, route security, denial-of-service response or customer portability. Monetisation depends on operational packaging, not the size of the allocation.

The best proof would be a transparent service schedule. It should distinguish best-effort household access from committed business capacity; define availability, restoration and support response; explain service credits; state whether a backup circuit follows a separate physical path; and identify which customer equipment is included. This would narrow sales negotiation and prevent staff from making promises the network cannot fund. Without it, the high business prices may signal hidden quality, or they may simply be a tariff aspiration.

The cost stack beneath an unlimited plan

No Shabakai financial statements are public, so a precise margin model would be false confidence. The sector's cost anatomy is nonetheless visible. In a 2024 explanation of operator spending, the Communications Service used MegaFon Tajikistan as an example: 37% of 2023 cash receipts went to taxes, about 9% to maintaining existing infrastructure, 19% to technical development, 20% to internet traffic and interconnection, and 6% to employees. Those percentages belong to a mobile operator, not Shabakai. They illustrate why telecom revenue is not free cash.

A young fixed provider faces a different but equally unforgiving sequence. It pays for upstream capacity and local transport before knowing whether every household will renew. It buys optical line terminals, switches, routers, splitters, customer terminals, batteries and tools. It secures building access and lays feeder and drop fibre. It needs monitoring, billing and customer-support systems. It employs installers and network engineers whose salaries continue between activations. It carries spare equipment because waiting for an imported replacement can extend an outage.

The SkyX page itself identifies GPON, fibre to the building and point-to-point wireless as connection methods. Each has a different capital profile. GPON shares feeder fibre efficiently but places dependence on an optical line terminal and passive splitters. Fibre to the building can lower the cost per flat in dense blocks, though building power and distribution equipment become failure points. Wireless can reach a customer quickly but requires clean spectrum, line of sight, mounting rights and careful capacity planning. The cheapest installation method is not necessarily the cheapest life-cycle service.

"Unlimited" shifts the traffic risk to the provider. The household pays a fixed amount while video quality, device count and cloud use grow. If upstream transit is priced by committed capacity or peak use, a few heavy users can make a low tariff unattractive. Faster domestic caches and direct peering can reduce that cost, but only for traffic that is actually local or cached. The company must know busy-hour usage by access segment and the marginal cost of another Mbps before it promises more speed.

Power belongs in the same calculation. In a November 2024 winter-preparation notice, the regulator warned operators to keep base stations operating despite winter challenges and power shortages. Fixed access equipment, rooftop wireless links, building switches and customer routers all need electricity. Batteries, generators, fuel, maintenance and theft protection are part of reliability. A second data path cannot save a service if both paths and the access switch share one exhausted battery.

The 1,024-address transfer is also a capital decision, although its price is undisclosed. RIPE's transfer classification establishes a permanent change in holdership; it does not disclose cash consideration or any service agreement. The upside is a durable addressing asset and less dependence on borrowed customer addresses. The downside is tying capital to IPv4 while still paying another network to originate the route. The return depends on how many paying connections use the space and whether the handoff reduces operating costs or merely changes the registry name.

Two suppliers do not automatically make two paths

The historical AS202726 entry listed import policies from TT Mobile's AS43197 and Tojiktelecom's AS51346. On paper, two providers can improve resilience and bargaining power. In practice, both sessions could enter the same building, cross the same duct, depend on the same border route or converge on the same wholesale gateway. True redundancy requires physical, commercial and operational separation that a routing policy alone cannot prove.

The live evidence is narrower. The autonomous-number entry is gone, while the transferred IPv4 block still originates through Orien Invest. That leaves at least three possible interpretations. Shabakai may be staging a migration and using Orien temporarily. It may have decided that an independent network is unnecessary and will keep wholesale origination. Or the January plan may have been withdrawn while the commercial service proceeds through partner networks. Public data cannot choose among them.

Tajikistan's wholesale structure makes this more than a technical detail. The World Bank's project appraisal described state-owned Tajiktelecom as dominant in fixed retail and wholesale services and identified international connectivity as a bottleneck. In November 2023 the regulator said it had licensed two private companies for direct international high-speed channels. A 2025 reform then required Tajiktelecom to publish transparent, non-discriminatory terms for international wholesale broadband. The World Bank reported in June 2026 that the average dollar price of 1 GB of mobile data had fallen about 8% since the reform programme began.

Cheaper and more open wholesale access can lift Shabakai's gross margin. It also lowers rivals' costs and reduces the value of owning an exclusive path. The company must decide whether to spend on independent international reach, buy diverse wholesale service, or remain an access specialist. Owning every layer is not automatically rational. Independence creates value only where control improves price, quality or bargaining power enough to cover the extra capital and staff.

Domestic peering offers a cheaper intermediate step. Tajikistan's first national exchange, TJ-IX, began operating in Dushanbe in 2025. PeeringDB's current listing shows 16 connections and 610 Gbps of participant capacity, including Tojiktelecom, TT Mobile, Babilon-T, Tcell's network and several smaller providers. Neither Shabakai's former AS202726 nor Orien's AS48001 appears in that public participant list. PeeringDB is self-reported and absence is not proof that no local traffic arrangement exists. It is evidence that direct exchange participation has not yet been demonstrated.

Joining an exchange would not fix international dependence, but it could keep local traffic local, reduce latency and lower paid-transit demand. More important, it would make the network's operating boundary clearer. A published exchange port, route-security policy and active autonomous number would show that the company controls routing rather than only address holdership. Until then, "owning the network" should remain a question, not a conclusion.

The customer base has to pay for density

Tajikistan gives a new provider a real demand case. The World Bank's 2025 economic update reported rapid economic growth and argued that better digital connectivity could reduce trade costs, improve public services and support new forms of commerce. Its companion release said poor internet connectivity affected nearly half of customs checkpoints. The 2024 Digital Foundations project includes funding to connect at least 100 schools and nearby public entities, with the intention of attracting private capital into places that are currently unconnected.

But demand is not the same as an addressable customer base. More than 70% of Tajikistan's population lives in rural areas, according to the Asian Development Bank's 2025 Tcell financing announcement. A fixed provider gains density in Dushanbe and selected regional centres; it loses density across mountain settlements and long rural routes. Mobile and satellite technologies often spread the cost more efficiently outside urban clusters.

Shabakai's public offer appears city-oriented. It advertises city Wi-Fi, fibre-building access and home plans, but no coverage map names a district, town or building. This weakens both customer acquisition and capital planning. A household cannot know whether the attractive price is available. An analyst cannot tell whether a low tariff sits on an inherited dense network, a planned build or a few trial locations.

The ideal early customer mix would combine many low-support household lines with a modest number of high-value businesses in the same footprint. The households fill passive access capacity and create recurring base revenue. The businesses fund better backhaul, spares and engineering. Public institutions can add long contracts, but tender cycles and delayed payment can concentrate cash risk. A single 40,000-somoni account looks efficient until its cancellation strands a custom fibre run and reserved capacity.

Customer concentration is therefore one of the largest unknowns. If one corporate group or one public contract provides most revenue, Shabakai's apparent independence may conceal commercial dependence. If thousands of households provide most revenue, collection, churn and support efficiency become decisive. If Orien or related businesses are anchor customers, the relationship may make the launch economical but obscure the external market test. None of these cases is established by current evidence.

Management should publish operating measures that reveal the mix without exposing customer identities: active residential and business lines, premises passed, take-up, average revenue by segment, churn, contract duration, bad debt, installation payback and revenue concentration. Growth in "connections" is not enough. Free trials, city access accounts and inactive lines can inflate the number while consuming support and equipment.

Competition will not wait for the network handoff

The realistic substitutes are already visible. Tojiktelecom says it has provided internet, telephony and television across the country since 1996. Babilon-T has a 25-year operating history and public home-fibre tariffs. Tcell sells TezNet GPON alongside mobile service. TT Mobile sells MegaFon home access and can combine it with mobile distribution. NETS and Oshno appear in customer recommendations, while other fixed and mobile operators participate at TJ-IX. These competitors have brands, support channels, equipment and customer histories that a new entrant must replicate or beat.

Mobile broadband remains the most important substitute for households that cannot justify a fixed bill or installation. The regulator counted 4.9 million mobile and fixed internet users in the first quarter of 2025, of whom 4.7 million were mobile users. Mobile cannot always match fibre consistency, especially at busy hours, but it has national distribution and low switching friction. A household may tolerate variable performance rather than pay for two services.

Established operators are still investing. In October 2025 the Asian Development Bank agreed a $30 million local-currency loan to Tcell to expand and modernise mobile coverage, including rural areas and preparation for urban 5G. Shabakai cannot win an investment race across the whole country. It has to pick places where fixed density and local accountability matter more than national scale.

Starlink changes the upper end of the substitution set. Tajikistan's regulator signed a licensing agreement in October 2025, and the service officially launched on 5 February 2026. Satellite broadband will not be the cheapest solution for every Dushanbe flat, and public Tajik pricing, equipment availability and enterprise support terms still need comparison. It gives remote businesses and affluent continuity buyers a route that does not depend on the same terrestrial last mile. It can also serve as a backup rather than a replacement.

This puts a ceiling on an unproven business premium. A company asked to pay 12,000 to 40,000 somoni a month will compare a second terrestrial circuit, a mobile failover, Starlink, or a combination assembled by its own IT staff. Shabakai wins when one local contract makes those pieces work together and assumes responsibility for restoration. It loses when the buyer can obtain similar resilience with two commodity providers at lower cost.

Competition also disciplines the add-ons. Cloud storage competes with global platforms and local data services. Network protection competes with router features, endpoint software and managed-security providers. Support competes with the customer's own technician. These products raise account value only if they integrate with the access service and reduce total customer effort. Bundling labels onto third-party products would increase revenue but not necessarily value.

Regulation, power and politics sit inside the service promise

The provider does not control all causes of failure. Tajikistan's Electronic Communications Law gives the state broad roles in licensing, supervision, numbering, internet domains, channels, radio frequencies and suspension under legal authority. The Communications Service says it supervises network development, service quality, equipment sales and operator payments, while also enforcing obligations related to security and emergencies.

Compliance is not a one-off licence fee. A provider must maintain lawful customer records, respond to regulator instructions, meet quality requirements, manage abuse reports, protect data and keep technical contacts available. Cloud and security add-ons may increase those duties. The SkyX site does not publish a privacy notice, customer terms, licence details or complaint process in the visible pages reviewed for this article. Those omissions matter because reliable service includes reliable recourse.

Service quality is already politically salient. The regulator reported 2,758 complaints to its 4030 contact centre by 3 April 2025, with 876 still under review. In November 2025 it held a meeting with mobile operators after complaints about low speeds and temporary interruptions in several regions. These figures do not concern Shabakai specifically. They show that customers and officials are sensitive to the gap between advertised access and delivered performance.

Political intervention can also override technical redundancy. Internet Society Pulse rates Tajikistan's internet resilience below the Central Asian average. Access Now documented a roughly four-month shutdown in the Gorno-Badakhshan region beginning in November 2021. A private operator cannot sell its way around a lawful nationwide restriction, and a terrestrial second path may not preserve service when policy is the cause of disconnection.

The commercial response is transparency, not a promise of immunity. Contracts should distinguish provider-controlled outages from national restrictions, customer power failures and force majeure. Network reporting should still show what Shabakai controls: access availability, core incidents, upstream loss, mean restoration time and whether failover worked. Customers can then decide what risk remains theirs and whether a satellite or mobile backup is justified.

Geopolitics enters through supply and transit. Tajikistan is landlocked, and international routes ultimately depend on neighbouring countries and foreign carriers. Imported routers, optical equipment, batteries and software expose the provider to currency, customs and vendor-support risk. Public sources do not identify Shabakai's equipment or software suppliers, so a claim of Chinese, Russian, European or American dependence would be speculation. The decision test is supplier concentration: can the company replace a failed part and renew critical software if one vendor, border route or payment channel becomes unavailable?

Unofficial signals say customers will test the promise hard

There is not yet a substantial body of independent public commentary about SkyX or "Shabakai oson." Searches reviewed for this article did not reveal a durable set of customer reviews, installation reports or measured speed tests tied clearly to the company. That absence is consistent with a new commercial launch, but it cannot distinguish a very small live service from a pre-launch catalogue. It should not be converted into either a positive or negative satisfaction claim.

Broader Tajik customer discussion is much louder. In a January 2026 Reddit request for home-internet advice, commenters recommended established names including Tojiktelecom, NETS and Babilon while another said there was no good option. An August 2025 thread about recurring home outages included complaints about Babilon, interest in Starlink, a recommendation for Oshno and debate about whether the underlying problem was competition or central infrastructure.

These are anecdotes from small, self-selected samples. The claims are not audited, locations and service types differ, and dissatisfied users are more likely to post. They do not establish failure rates for any provider and say nothing directly about Shabakai. Their economic value is narrower: the things users discuss are weekly interruptions, busy-hour speed, support, regional availability and the hope of a substitute. Those are precisely the dimensions on which SkyX's low price and reliability language will be judged.

The absence of direct reviews also raises the importance of first-party proof. A new provider can publish measured availability, a real coverage checker, installation lead times, support channels and clear terms before it has a large review base. It can invite independent speed measurement without selecting only favourable results. That evidence would cost less than a national brand campaign and would help business buyers understand what they receive for the premium.

Who pays, who benefits and who carries the downside

The household pays 150 to 400 somoni each month and receives affordable fixed capacity if the line is available at its address. The immediate beneficiaries are family members using education, entertainment, communication and work services. Shabakai benefits from recurring revenue and the chance to sell support, storage and security. Equipment suppliers, upstream carriers, landlords and technicians receive cash before the shareholder knows whether the connection will repay its installation.

The business pays far more because interruption has a larger cost and because it may demand protected capacity. Its employees, customers and suppliers benefit when payments, cloud applications and communications continue. Shabakai captures value only if the price exceeds reserved capacity, equipment depreciation, field support, service credits and the capital cost of redundancy. If the provider underestimates any of those inputs, revenue can grow while value is destroyed.

The downside is asymmetric. Customers carry lost work and switching costs during an outage. Shabakai carries repair, refund, churn and reputation risk. An upstream carries only the obligations in its wholesale contract. The state carries part of the economic cost when digital services or public facilities fail. Employees carry the strain of restoring a young network with limited spares. Investors carry the residual capital loss if cheap tariffs never produce enough density.

The Orien transition adds another layer. Shabakai now holds the IPv4 allocation, while the route still names Orien's network. If that arrangement is protected by a long-term service agreement with clear escalation and migration rights, it may lower launch risk. If it depends on informal alignment, customers ultimately carry a hidden supplier concentration. The public record cannot tell which. A registry transfer alone does not allocate outage liability.

The same logic applies to apparent multi-provider plans. Two invoices can improve negotiating leverage, but only a tested failover design benefits the customer. Shabakai should map shared ducts, power, buildings, border crossings and wholesale gateways, then sell redundancy according to the weakest common dependency. Otherwise the company will charge for two paths while carrying the cost of only one real failure domain.

A capital allocation plan that could earn the premium

The first priority should be evidence of the existing service, not breadth. Shabakai needs an accurate map of fibre, wireless sites, splitters, power backup, upstream handoffs and paying premises. It should measure busy-hour utilisation and incidents by segment. This establishes which part of the apparent SkyX offer is live and where another customer adds contribution rather than congestion.

The second priority should be the handoff from Orien. Management should choose a stable end state: a protected wholesale-origination agreement, or an independent active network with at least two commercially and physically diverse upstreams. Either can be rational. Ambiguity cannot. Route security, abuse handling, address assignment and migration responsibility should match the chosen model.

The third priority should be dense access and spares. Every expansion area needs a payback threshold based on premises passed, expected take-up, connection cost and support distance. Optical terminals, power supplies and access switches should have local replacements. Field teams need realistic restoration routes. A large IPv6 allocation should be deployed to customer equipment rather than left as a symbolic asset, while IPv4 should be conserved for services that need it.

The fourth priority should be service design. Household access needs plain-language speed treatment and connection terms. Business access needs a service schedule, escalation path, restoration target and optional physically diverse backup. Network protection and cloud storage should be sold only when the company can identify the supplier, data location, security responsibility and margin after support.

The fifth priority should be measured distribution. The 10-day free plan and city-Wi-Fi benefits may be effective acquisition tools, but management should track conversion, support contacts, fraud and lifetime value. It should stop promotions that produce accounts rather than cash. Marketing spend should follow proven installation capacity; otherwise a successful campaign can create a queue of dissatisfied applicants outside coverage.

This sequence is deliberately unglamorous. It favours route contracts, batteries, inventory, measurement and support over a broader catalogue. That is exactly what the core question requires. Strategy without the resources to restore service is branding. A new provider earns trust when its spending pattern matches the outage risk it asks customers to transfer.

A conditional judgment

The evidence supports a real company building a network position, not a proven independent operator with established economics. The RIPE membership, IPv6 allocation and permanent IPv4 transfer show commitment. The SkyX offer shows a coherent fixed-broadband ambition and unusually aggressive residential pricing. Tajikistan's low fixed penetration and falling wholesale costs create room for a focused entrant.

The evidence also shows a business in transition. Its apparent public brand does not state the legal contracting company. Its former autonomous number is absent from the live registry. Its newly held IPv4 block remains originated through the transferor's network. The public site provides no coverage, customer count, licence, service level, ownership or financial data. There is no demonstrated participation at the national exchange and no current proof of two independent upstreams.

The base case is therefore a small or launching Dushanbe-focused access provider using partner routing while it builds a household and SME book. Residential prices win attention; a limited number of business contracts fund backhaul and support. Value creation is possible if the company keeps the footprint dense, converts trials, retains customers and formalises its network boundary. This is a scenario, not a disclosed operating result.

The upside case is stronger. Shabakai completes the resource handoff, deploys IPv6, activates independent routing or secures protected wholesale terms, joins domestic peering, and proves physical diversity. Business customers accept the premium because restoration is fast and reported. Dense buildings generate high take-up at low marginal cost. Falling wholesale prices then expand margin faster than retail competition reduces tariffs.

The downside case is a reseller carrying the costs of an operator. Cheap household plans attract support-heavy users and require subsidised installation. Business customers negotiate down list prices because no service level supports them. One upstream, one physical route, one power point or one anchor customer remains critical. New equipment and the IPv4 transfer consume cash before recurring revenue reaches scale. Incumbents and Starlink take the customers most willing to pay for continuity.

Can "Shabakai oson" make customers pay enough for reliability? The tariff page says it intends to try. The current evidence does not yet show that the premium reaches the parts of the network that create reliability. The answer turns on whether the June resource transfer is followed by route control, contractual protection, dense customer acquisition and published performance. Owning an address block is a start. Owning the outcome of an outage is the business.

Facts that would change the judgment

The first decisive disclosure would establish identity and ownership: a corporate extract, telecom licence, customer contract or legal notice that connects "Shabakai oson" to SkyX and explains the relationship with Orien Invest. The consideration and service agreements attached to the IPv4 transfer would show whether June represented a purchase, restructuring or operational separation.

The second would establish network control: a current autonomous number, active route announcements, signed route authorisations, named upstreams, physical path maps, exchange participation and tested failover results. Evidence that two providers use different ducts, power and border paths would support the reliability premium. Evidence that both converge on one wholesale or physical dependency would weaken it.

The third would establish customer economics: active paying lines by segment, premises passed, take-up, realised monthly revenue, gross margin, churn, installation cost, payback, support contacts, bad debt and revenue concentration. High household retention and low support intensity would make 150 somoni look efficient. Short tenure or heavy field work would make it look like an acquisition subsidy.

The fourth would establish delivered quality: monthly availability, busy-hour throughput, latency, packet loss, restoration time, incident causes and service credits. A published business schedule with consistent performance would justify the gap between residential and corporate tariffs. A list price without those measures would not.

The fifth would establish capital resilience: cash available for equipment refresh, spares, batteries and expansion; supplier concentration; equipment lead times; and a plan for winter power. The company does not need to own every layer. It needs enough cash and contractual control to keep the layers it sells working when a supplier, fibre, router or power source fails.