Summary

  • The economic incentive is not growth for its own sake. DATASFERA has to earn enough recurring gross margin from local internet, hosting, colocation and network-related services to keep Ukrainian workloads reachable when the power grid, imported equipment supply and repair labor market are all under stress. The public numbers show a real but small operating base: Opendatabot reports 2025 revenue of UAH 10.529 million, net profit of UAH 334,900, assets of UAH 4.529 million, liabilities of UAH 1.001 million and four employees. At the National Bank of Ukraine's official exchange rate for August 10, 2026, that revenue is about USD 235,000 and the reported profit is about USD 7,500. That is enough to prove continuing commercial activity; it is not enough, by itself, to fund major data-center renewal or sustained emergency capital expenditure without customer prepayment, owner funding, supplier credit, group support or higher-value protected workloads.
  • The Elias Ward judgment is cautious: DATASFERA is investable as a local continuity utility only if its protected customers value Dnipro proximity, Ukrainian jurisdiction, direct support and short repair loops more than they value hyperscale abstraction or cheaper unmanaged access. Its evidence base supports a regional network and facility role, not a large cloud platform. The strongest public facts are the legal registration, VAT status, provider-register presence, visible revenue, the Civilization payment and tariff surface, RIPE membership, PeeringDB facility listing, AS35319 routing evidence, public IPv4 announcements, DTEL-IX and RETN visibility, and third-party facility claims of 64 cabinets, half a megawatt of utility input and a half-megawatt diesel generator. The weakest area is ownership and control boundary: some routing records point to ALTAIR MEGA LLC while some payment, registry and facility records point to DATASFERA. A buyer should therefore underwrite the company on confirmed contracts and cash collection, not on the whole apparent network group.
  • Unit economics are the center of the judgment. If DATASFERA's 2025 revenue were spread evenly over 64 cabinets, it would equal about UAH 13,700 per cabinet per month before power, cooling, rent, transit, labor, taxes, repairs and overhead. That calculation is only a stress test because the revenue mix is not disclosed, but it shows why the company cannot compete by selling cheap space alone. If the Civilization retail tariffs are part of the same commercial surface, then UAH 140-300 monthly broadband plans and a UAH 60 static IP option imply a high-volume, low-ticket cash model, not a deep enterprise cloud model. The company needs either high utilization, protected local customers, or a bundled revenue stack in which access, static addressing, hosting, colocation, cross-connects and repair support reinforce each other.
  • The facts that would change the judgment are concrete: signed long-term colocation or protected-hosting contracts, evidence of cabinet utilization, power-usage billing, fuel reserve policy, generator test logs, battery autonomy, transit contracts, spare-router and optical inventory, insurance terms, customer concentration, churn by segment, cloud-migration losses, unpaid receivables, and the exact legal relationship between DATASFERA, ALTAIR MEGA, INTER-NETWORKS and the Civilization brand. Without those facts, the right conclusion is not that DATASFERA is weak. It is that the public record supports a resilient local operator thesis but does not support a claim that public profits alone can finance wartime replacement cost.

The Incentive Starts With One Protected Workload

Start with one local workload that cannot simply disappear into a foreign cloud region: a municipal office connection, a local business server, a small hosting account, an edge router for a school, a neighborhood fiber customer who needs payment and support in Dnipro, or a rack-mounted system whose owner wants Ukrainian jurisdiction and a technician nearby. The buyer is not purchasing a generic megabit or a generic virtual machine. The buyer is purchasing continuity under constraints.

The practical question is whether DATASFERA can charge enough for that continuity to cover the parts of the service that are invisible when everything works: power redundancy, cooling, diesel, batteries, routing diversity, cross-connect access, cabinet density discipline, customer support, field repair and the option to replace damaged imported electronics.

That incentive is sharper in Dnipro than it would be in a stable wholesale data-center market. Ukraine's macro environment makes local infrastructure both more necessary and harder to finance. The World Bank's 2026 reconstruction update estimates very large national recovery needs and identifies energy as one of the sectors with heavy damage. The United Nations reported that attacks on energy infrastructure across the 2025-2026 winter damaged generation, transmission and distribution assets and triggered emergency outages. For an access provider or colocation operator, those facts do not create an abstract risk.

They raise the cash cost of being trusted: diesel has to be available, battery autonomy matters, technician response has to continue through alarms, and customers must be told honestly what is protected and what is merely best-effort.

The recurring revenue available to absorb that cost is visible but small. Opendatabot reports that DATASFERA generated UAH 10.529 million of revenue in 2025, up from UAH 6.116 million in 2024 and UAH 4.546 million in 2023. That is meaningful growth, especially under wartime conditions. It also reports UAH 334,900 of net profit in 2025, a 3.18 percent margin, with four employees. At the August 10, 2026 official exchange rate of UAH 44.7579 per dollar, the 2025 net profit converts to roughly USD 7,500.

The profit figure is not a free-cash-flow statement, and it does not tell us depreciation, owner loans, related-party costs or capital expenditure. But it is enough to frame the incentive: DATASFERA cannot treat continuity as a free feature. It has to price it explicitly or cross-subsidize it from other cash flows.

That is why the public headline should not be "small Ukrainian provider survives." Survival is a state, not a strategy. The strategic question is whether DATASFERA can sell continuity at a price that pays for the replacement cost of the equipment and supply chain behind it. A customer may choose local hosting because the workload needs low-latency access, data locality, human support, local payment, or a cabinet that can be visited. A customer may reject local hosting because hyperscale cloud absorbs hardware renewal, power engineering and geographic redundancy.

DATASFERA's economic task is to defend the gap between those choices. It wins when the local customer's avoided downtime, avoided migration cost and operational familiarity are worth more than the premium required to keep local infrastructure alive.

Identity And Control Boundary

The legal entity is clear. Opendatabot identifies TOV "DATASFERA" by USREOU code 37733537, founded on May 24, 2011, registered in Dnipro at Shevchenka Lane 9, premises 101, with UAH 1.416 million of statutory capital. It lists the main activity as wired telecommunications, names Olena Orel as director, and shows ownership split among Alina Vovk, Liudmyla Uhnych and Pavlo Dukh. YouControl and Clarity Project independently support the same core identity: registered status, Dnipro location, statutory capital, wired-telecommunications activity and VAT registration. The legal base is therefore not the weak part of the file.

The weak part is the operating perimeter. The public web shows a shared Civilization service surface. The Civilization contact page lists payment details for TOV "DATASFERA", TOV "ALTAIR MEGA", TOV "INTER-NETWORKS" and other entities on the same customer-facing site. The tariff page advertises residential and settlement-specific internet plans.

The "about" page claims that the company's staff operate the largest switching node in Dnipro around the clock and that commercial providers in the oblast are customers, but that claim appears on the Civilization brand site rather than in DATASFERA's statutory filing. The public offer PDF on the same site names INTER-NETWORKS as the provider for that offer. Those facts do not discredit DATASFERA. They warn the analyst not to collapse every Civilization, ALTAIR MEGA and INTER-NETWORKS claim into DATASFERA's own balance sheet.

Routing evidence has the same ambiguity. RIPE's member page for LLC "DATASFERA" gives a Gogolya 15 address in Dnipro, a DATASFERA email contact and Ukraine as the serviced area. PeeringDB lists a Datasfera facility at Gogolya 15 in Dnipropetrovsk with DTEL-IX exchange presence, RETN carrier presence and ten local network presences. Yet RIPE database records and RIPEstat identify AS35319 as held by ALTAIR MEGA LLC, and bgp.tools gives ALTAIR MEGA LLC as the registered network while Hurricane Electric labels AS35319 as LLC DATASFERA.

The better interpretation is an affiliated or overlapping operating group in which DATASFERA, ALTAIR MEGA and the Civilization surface are related in practice but not identical in public evidence.

That matters economically because credit, liability, contracts and replacement obligations attach to legal persons. If a customer contracts with DATASFERA, the customer's enforceable counterparty is DATASFERA, not the whole network ecosystem implied by BGP pages. If a route is announced by AS35319 under ALTAIR MEGA's LIR object, the route may help DATASFERA's service but does not, by itself, prove DATASFERA owns the router, transit contract or address block. For a small local operator, this boundary can be normal. Ukrainian regional providers often share brands, billing platforms, technical staff and infrastructure.

But the investor or enterprise buyer must demand the corporate map before treating public network scale as DATASFERA's standalone capacity.

Revenue, Pricing And Unit Economics

The disclosed financials are the cleanest starting point. Opendatabot's 2025 row gives UAH 10.529 million of revenue, UAH 334,900 of net profit, UAH 4.529 million of assets, UAH 1.001 million of liabilities and four employees. Those figures imply revenue per employee of about UAH 2.63 million and net profit per employee of about UAH 83,700. They also imply liabilities equal to about 22 percent of assets and revenue equal to about 2.3 times assets.

For a small telecom operator, that combination is plausible: fixed network assets can carry recurring monthly revenue, while a small staff can run the platform if installation, major fiber work, accounting or specialist repairs are partly outsourced or shared across related entities.

The margin is the critical signal. A 3.18 percent net margin in 2025 is not a platform margin; it is an operating-company margin. It says the company is collecting cash and producing an accounting profit, but it leaves little visible room for heavy renewal. The same source shows revenue growth from UAH 3.759 million in 2020 to UAH 10.529 million in 2025, which is strong in local currency. But inflation, wage pressure, imported equipment costs, higher power resilience requirements and currency depreciation all reduce the real comfort that top-line growth might otherwise imply.

The National Bank of Ukraine's 2026 inflation commentary expected inflation moderation over time but still described an economy operating under attack-related destruction and external financing needs. That is the world in which DATASFERA has to buy routers, optics, batteries and generator inputs.

The Civilization tariff page helps anchor the access side of the revenue stack, subject to the identity caveat. In Dnipro it lists UAH 140 per month for a 60 Mbps plan, UAH 179 for a 100 Mbps plan and UAH 300 for a 500 Mbps plan. In surrounding settlements it lists combinations such as UAH 180 for 20 Mbps, UAH 250 for 50 Mbps and UAH 300 for 100 Mbps, plus a separate Karnaukhivka set from UAH 130 to UAH 200 and a static-IP option at UAH 60 per month. Those numbers are not enterprise colocation prices. They are mass-market, low-ticket prices.

If DATASFERA participates materially in that retail base, the operator's economic strength depends on scale, low churn, disciplined support costs and the ability to install fiber without turning every new subscriber into an unrecovered capital outlay.

The connect page reinforces that point. The advertised installation package includes up to 150 meters of optical cable for the subscriber line, mounting from the pole to the premise, subscriber optical equipment, splicing, signal testing, a copper patch cord and initial setup of one device or Wi-Fi router. Hidden cable work, internal LAN construction and a router can be excluded or charged separately. That is a classic access-provider way to prevent a low monthly plan from swallowing the whole installation cost. The operator tries to define what is included, what is not included and when the customer bears extra work.

In a stable market, that protects margin. In a wartime market, it also protects repair capacity because field labor and optical materials are not free.

The facility-side unit economics are more demanding. Data Center Map describes Datasfera as a colocation provider with 64 cabinets, half a megawatt of utility input, a half-megawatt diesel generator, 2 x 10 Gbps internet, last-mile options including fiber, DSL and WiMax, and connection to operators present in the data center. 2IP's listing adds a similar identity as a Dnipropetrovsk data center founded in 2010, with two hermetic zones, 400 kW first-category power supply, a 0.5 MW redundant diesel generator, meet-me-room service and N+1 precision air conditioning.

These are third-party directory claims, not audited engineering reports, but they describe a capital-heavy business.

If the 2025 revenue were spread evenly across 64 cabinets, the result would be about UAH 164,500 per cabinet per year, or about UAH 13,700 per cabinet per month. At the August 2026 exchange rate that is roughly USD 306 per cabinet per month before power, cooling, transit, rent, staffing, tax, repair and capex. The calculation is deliberately conservative because DATASFERA's revenue mix may include access, IP services, hosting, cross-connects, field work and other items, not just cabinet rent. Still, the stress test shows the economic direction. A 64-cabinet facility cannot be healthy if it prices continuity like cheap residential access.

It needs either high cabinet utilization, meaningful power pass-through, premium workloads, recurring network services, or a related-company model where shared infrastructure economics sit outside DATASFERA's standalone accounts.

Cost And Capital Burden

The capital burden splits into three categories. The first is sunk infrastructure: room, cabinets, cooling, power distribution, meet-me-room access, routers, optical plant and billing systems that already exist. The second is recurring operating cost: power, transit, exchange participation, rent, wages, taxes, bank fees, support and routine maintenance. The third is resilience and replacement: diesel, batteries, generator service, spare optics, spare routing gear, cable repair, security, insurance and emergency labor. Public financial statements do not reveal the split.

The analyst has to infer pressure from the business model and source evidence.

The sunk infrastructure appears real enough. PeeringDB's facility page places Datasfera at Gogolya 15 and records one exchange, one carrier and ten local networks. Data Center Map and 2IP describe cabinets, power and data-center features. RIPE membership and IP-resource records show operational contact and network-resource participation. These sources together make it unlikely that DATASFERA is merely a reseller with no infrastructure surface. They do not, however, prove the age, condition, ownership or debt status of the assets.

The recurring cost base is partly visible through network relationships. AS35319 is announced, has ten IPv4 prefixes in RIPEstat's public view and is seen by bgp.tools and Hurricane Electric with upstream or peer relationships involving RETN, Eurotranstelecom and Enterra. RIPEstat's neighbor view also shows AS35320, AS48964 and AS9002 on one side of the path and AS48085 on the other. The exact commercial relationship cannot be read from BGP alone, but the network is not isolated. Connectivity depends on upstreams, exchange paths and route policy that carry cost or reciprocal value.

No public evidence shows IPv6 origination, and multiple sources list zero IPv6 prefixes for AS35319. That absence is not fatal in a small local market, but it is a modernization gap for enterprise-grade hosting.

Power is the harder cost. Ukraine's energy regulator stated in 2026 that electricity prices for non-household consumers are formed by competitive market segments, while the regulator sets transmission and distribution tariffs. That means DATASFERA's power bill is exposed to market and system conditions rather than a simple protected household tariff. The United Nations and World Bank context shows why that matters: damaged generation and grid infrastructure create outage risk and system stress. A data center that advertises diesel backup transfers some outage risk from customer to operator.

The customer pays for continuity, but the operator buys fuel, maintenance and battery autonomy. If tariffs do not recover that cost, every outage can turn a paid service into a margin drain.

The reported balance sheet cannot carry unlimited resilience. UAH 4.529 million of assets and UAH 334,900 of net profit can support ongoing small-company operations, but a serious refresh of generators, UPS systems, chillers, routers, optical access gear and security would likely exceed one year's net profit. Imported electronics add currency exposure. Even if suppliers invoice in hryvnia, replacement cost is often connected to dollar or euro hardware chains. The NBU exchange rate of UAH 44.7579 per dollar on August 10, 2026 is a reminder that a UAH revenue base must still defend foreign-currency equipment economics.

That creates the central capital test. If DATASFERA's customers are mostly low-ARPU households and a few small public contracts, then continuity investment must be shared across a large base and protected by strict service terms. If the company has protected hosting or colocation customers that pay for power density, redundancy and support, then a small revenue line can still be attractive if those customers are sticky and prepaid. The public record does not reveal the mix.

The correct underwriting answer is therefore conditional: DATASFERA's business can work if the recurring protected workload base is strong enough; it becomes fragile if the facility claims are real but utilization or premium pricing is low.

Suppliers, Routes And The Price Of Redundancy

The network evidence shows redundancy but not immunity. BGP.tools lists AS35319 with three upstreams: RETN, Eurotranstelecom and Enterra. Hurricane Electric observes four IPv4 peers and ten originated IPv4 prefixes. RIPEstat's announced-prefixes data showed ten prefixes active across the late July to early August 2026 measurement window. IPinfo and IPLocate both report 6,400 IPv4 addresses and no IPv6. PeeringDB's Datasfera facility lists DTEL-IX and RETN, while the Colomap DTEL-IX page lists Datasfera among facilities reaching the exchange.

Those facts support a local interconnection and transit story: DATASFERA or its adjacent operating group has more than one path into the wider internet.

Redundancy, however, is not the same as contractual resilience. BGP can show observed adjacent networks; it cannot show minimum commits, service-level terms, payment status, physical diversity, last-mile separacy, repair priority or whether two paths share the same duct, power dependency or building entry. A buyer should ask whether RETN, Eurotranstelecom and Enterra paths enter the facility separately, whether routes are accepted through route servers or direct sessions, whether critical customers receive separate handoff equipment, and whether route filters and RPKI state are monitored.

Public evidence can say the network is visible and multi-homed. It cannot prove that an outage in one building, substation or fiber corridor will leave every protected customer online.

The supplier problem extends beyond transit. A local data-center operator relies on power suppliers, fuel logistics, generator service, battery vendors, cooling parts, optical cable, SFP modules, routers, switches, splicing teams, security, bank payment rails and often imported customer-premise gear. The Civilization connect page's included installation materials are small but illustrative: even a household connection consumes optical cable, subscriber equipment, splicing and labor. A rack customer consumes a different bundle: power circuits, cooling headroom, cross-connects, support and spare parts.

Both models turn supplier availability into customer continuity.

DATASFERA's visible position is best understood as a risk-transfer business. The customer pays the provider to absorb technical burdens the customer does not want to manage alone. In ordinary broadband, that burden is installation, support and network operation. In hosting and colocation, it is power, cooling, space, connectivity and physical security. In wartime Ukraine, it also includes continuity under power disruption and repair scarcity. The transfer only works if the provider prices the burden. If DATASFERA sells continuity as a cheap add-on, it absorbs risk without capital recovery.

If it sells continuity as a measured product with power limits, service tiers and explicit repair terms, it can defend margin even while remaining smaller than national cloud competitors.

Customer Concentration And Demand Signals

Customer concentration is the largest unanswered commercial question. Public procurement data visible through Opendatabot and YouControl shows recurring small contracts for unlimited internet access or provider services with a Dnipropetrovsk special school. Opendatabot shows tenders of UAH 12,000 for 2023, 2024, 2025 and 2026 and summarizes UAH 60,000 from the top public buyer. YouControl similarly lists UAH 12,000 contracts in 2024, 2025 and 2026. These are useful because they prove a recurring public-sector service relationship. They are not large enough to anchor the company.

A single UAH 12,000 annual contract is about 0.11 percent of DATASFERA's 2025 revenue, and UAH 60,000 is about 0.57 percent. The public-sector record, at least the visible part, is a signal of presence rather than concentration.

The private customer base is not disclosed. The Civilization about page claims that commercial providers in the region are customers of the company's Dnipro switching node. If true for DATASFERA or the adjacent group, that is strategically stronger than a retail-only base because provider customers buy interconnection, locality and network convenience. But the claim is not a customer list, and it is not in DATASFERA's financial note. It should be treated as a market signal. A customer concentration review would need the top ten customers, revenue by product, receivables aging, churn, prepaid balances and contract duration.

Unofficial signals point in mixed directions. Portmone lists DATASFERA as an internet-payment recipient with the same USREOU code, an IBAN and a 4.6 rating from 14 users, and says three Portmone customers paid the company last month. This is a tiny signal, not a market share estimate. It supports retail billing activity and consumer-facing recognition. IPinfo reports a small number of hosted domains and pingable IPs, while also tagging at least one IP in the ASN as VPN and at least one as BitTorrent in a recent window.

Those tags are not allegations against DATASFERA; they are normal internet-reputation observations that matter because hosting and access networks inherit the behavior of their users. A provider with low margins cannot ignore abuse handling, because bad reputation can increase support cost and degrade customer trust.

The pricing page gives a separate demand clue. Plans at UAH 140, UAH 179 and UAH 300 per month require either many subscribers or low service cost to move the revenue line. A rough sensitivity check shows the scale: if all 2025 revenue came from UAH 179 monthly accounts, it would correspond to about 4,900 account-years; if all came from UAH 300 monthly accounts, about 2,925 account-years; if all came from UAH 140 monthly accounts, about 6,267 account-years. Those are not subscriber estimates because DATASFERA's revenue mix is not known and the tariff page belongs to a shared brand. They are boundary checks.

They show that low-ticket access alone needs a sizeable base, while colocation or provider customers can move revenue with far fewer accounts.

Competition, Substitutes And Why Locality Still Matters

DATASFERA's competition is not just another Dnipro ISP. It competes with several substitutes at once. A household can choose another fiber provider, mobile broadband or satellite backup. A small business can move mail, files and applications to a hyperscale cloud region outside Ukraine. A software company can use managed cloud infrastructure instead of a local cabinet. A public body can tender connectivity annually and switch if price or reliability disappoints. A carrier can peer or buy transit elsewhere.

A rack customer can move to Kyiv, western Ukraine, Poland or a pan-European hosting provider if the migration cost and data-location requirements allow it.

The alternative set is visible in public facility directories. Scrutica's Ukraine profile lists multiple operational colocation and hosting facilities across the country, including Kyiv, Lviv, Odesa and Dnipro-related locations. Data Center Map and DC Byte frame Datasfera as one facility among a broader data-center market. PeeringDB's global context shows that interconnection buyers compare facilities by exchange access, carriers, networks, geography and operational data. DATASFERA therefore cannot defend itself by saying it has cabinets. Many providers have cabinets.

It has to defend the specific value of Dnipro locality, customer familiarity, access to local operators, payment convenience, and a continuity posture customers believe.

Hyperscale cloud is the cleanest substitute for workloads that can be rewritten, replicated or moved. It absorbs hardware procurement, much physical security, power engineering, global connectivity and platform management. But it does not automatically solve every Ukrainian workload. It can create foreign-currency bills, data-governance concerns, latency to local users, dependence on remote support, and migration complexity. For small customers, cloud competence itself is a scarce resource.

DATASFERA's opportunity is in the workloads whose owners know they need continuity but do not have the staff, budget or appetite to rebuild everything around a foreign platform.

There is also a middle substitute that can be more dangerous than hyperscale cloud: a customer keeps the local access circuit but removes the protected workload from the local facility. In that case DATASFERA still carries the field-support burden, payment friction and last-mile repair work, but loses the higher-margin hosting or colocation layer that could have paid for power resilience. The company should therefore watch not only subscriber counts but wallet depth.

A household broadband customer at UAH 179 per month and a business customer buying local access, static addressing, backup handoff, cabinet space and remote-hands support have very different economic value even if both appear as active customers. The strategic defense is bundling: make the local network, address space, cross-connect, support desk and facility service useful together. If those pieces are sold separately at commodity prices, the highest-value workloads can migrate away while the provider is left with the lowest-margin obligations.

That bundling has to be credible. A protected workload buyer will ask what happens when power fails, when an upstream flaps, when a router dies, when a fiber cut lands outside office hours or when a customer needs a hand near the rack. A regional operator does not need to pretend it is global. It does need to be precise about what it can do faster than a remote cloud or distant data center. The answer may be local hands, nearby access plant, direct knowledge of Dnipro customers, Ukrainian paperwork and faster informal escalation. Those advantages are real only when backed by staff, spares and contract language.

That opportunity comes with a price ceiling. The provider cannot charge like a premium global cloud for a service that lacks public evidence of multi-region redundancy, audited uptime, deep staffing and large spare-part inventory. The source record supports a regional continuity utility: local access, facility presence, network routes, retail pricing, public tenders and data-center directory listings. It does not support a claim of hyperscale equivalence. The right product position is practical: keep local workloads reachable, make support tangible, disclose limitations, and charge enough for the resilience customers actually receive.

Regulation, War And Public-Service Risk

DATASFERA appears in the provider-register context. Opendatabot states that the company is in the register of providers of electronic communications networks and services, and the national regulator's public pages describe the register and the general authorization framework for electronic-communications activity. This matters because the company is not just a private hosting shop. It operates in a regulated communications environment where provider status, customer terms, service notices and lawful obligations matter.

The wartime overlay changes the business risk. RIPE NCC's Ukraine/Russia page explains that it has continued to treat members under standard procedures while recognizing payment and administrative difficulty for Ukrainian members due to conflict and sanctions. RIPE Labs' Ukraine resilience analysis describes how Ukraine's decentralized internet and local interconnection helped resilience, but also how power outages affected measurement probes and local stability.

Chatham House's account of Ukrainian internet resilience similarly describes physical damage, targeted infrastructure disruption and reconstitution of routes in occupied or contested areas. DATASFERA sits in that environment. Its local interconnection value increases when decentralization matters, but its operating burden increases when physical infrastructure and power are threatened.

The legal and commercial terms visible on the Civilization offer PDF also illustrate how risk is passed back to users, even though the offer names INTER-NETWORKS rather than DATASFERA. It defines service access, billing, customer acceptance, maintenance interruption, zero-balance disconnection and restoration after force majeure. This is standard ISP discipline. It means the customer is not buying infinite uptime. The provider defines when service is available, what the subscriber must pay, how equipment damage is handled, and what happens under extraordinary events.

For DATASFERA, the economic lesson is the same: continuity promises must be legally bounded or priced as a premium service.

The regulatory risk is not that public records show a current sanction or insolvency problem. The accessible records do not support that conclusion. The risk is evidence depth. YouControl says the company has licenses and court documents, but some sections are behind paid access. Clarity warns that tax registry data has limitations under martial-law restrictions, while it also shows no tax debt and a valid VAT certificate in the accessible snapshot. Opendatabot reports a VAT number and financials but does not replace audited accounts. A buyer should not overread positive registry status as operational assurance.

It is a necessary condition, not a resilience certificate.

Uncertainty And Contradictions

The main contradiction is naming. DATASFERA is the legal entity in the assignment, the payment recipient on Portmone and Civilization, the RIPE member name, and the facility name in PeeringDB and Data Center Map. ALTAIR MEGA is the RIPE organisation behind AS35319 in official database records and the registered network name in RIPEstat, bgp.tools, IPinfo and IPLocate. Hurricane Electric labels AS35319 as LLC DATASFERA, and some prefixes are described as LLC "DATASFERA" while others are described as ALTAIR MEGA, Enterra or Pavel Dukh. This is not rare in regional networks, but it is analytically important.

It prevents a clean standalone map from public evidence alone.

The second uncertainty is revenue mix. Public financials show total revenue and profit, not product segmentation. The article can calculate stress tests, but it cannot assert how much revenue comes from households, public-sector internet, hosting, colocation, cross-connects, wholesale customers, static IPs or related-party services. The difference matters. A UAH 10.5 million revenue base built from stable protected hosting customers is much more resilient than the same revenue base built from volatile low-ARPU residential accounts.

A UAH 10.5 million revenue base with high cabinet utilization is different from one in which the facility is underused and access customers carry the overhead.

The third uncertainty is cost visibility. No public source gives DATASFERA's electricity bill, diesel stock, generator age, PUE, cooling capacity, battery runtime, spare inventory, transit cost, rent or staff compensation. Data Center Map and 2IP describe facility features, but neither is an audited engineering inspection. The public accounts show small profits, not maintenance backlog. Because continuity is the thesis, the missing cost details are not minor. They are the difference between a company that can charge fairly for resilience and one that survives until a large repair bill appears.

The fourth uncertainty is customer concentration. Public tenders are too small to define the business, and private customer names are unavailable. The Civilization site's claim that commercial providers are customers is strategically interesting, but it lacks names, volumes and contract terms. The source set therefore supports "possible wholesale/local interconnection value" but not "proven diversified enterprise customer base." That distinction should remain visible in any investment, credit or procurement judgment.

Facts That Would Change The Judgment

The first changing fact would be a contract book. If DATASFERA can show multi-year protected-hosting or colocation agreements with monthly recurring charges, power pass-through, prepayment, termination penalties and a diversified customer list, the margin problem becomes less worrying. If the contract book instead shows month-to-month low-price access with high arrears, the visible growth becomes less attractive.

The second changing fact would be a facility evidence pack. Generator make, maintenance records, load tests, fuel autonomy, UPS topology, battery replacement dates, cooling redundancy, maximum and average power draw, rack utilization, fire suppression, access control, environmental monitoring and incident logs would determine whether the 64-cabinet and half-megawatt claims translate into sellable continuity. A half-megawatt generator is only valuable if it is maintained, fueled, tested and integrated into a working electrical design.

The third changing fact would be a network contract map. Transit agreements, physical path diagrams, cross-connect inventory, route policy, RPKI operations, route-server dependence, DDoS protection, abuse handling and IPv6 plans would show whether AS35319 and the Datasfera facility provide robust routing or merely basic reachability. Public BGP says the network is visible. It does not price the routes or prove physical diversity.

The fourth changing fact would be corporate linkage. If DATASFERA, ALTAIR MEGA and INTER-NETWORKS share ownership, assets, staff, contracts or cost centers, the economic analysis should be consolidated. If they are looser brand or billing neighbors, DATASFERA must be analyzed on its own revenues and obligations. The public record hints at linkage but does not prove consolidation.

The fifth changing fact would be churn and cloud displacement. If customers are leaving for foreign cloud, national hosting providers or other local fiber operators, DATASFERA must raise quality or narrow its product. If customers are adding local redundancy because war has made nearby support more valuable, then DATASFERA can justify premium continuity pricing. Both stories are plausible. The public record does not decide between them.

Final Judgment

DATASFERA's economic case is credible but narrow. The company has a legal identity, visible revenues, continuing profits, provider-register evidence, a customer-facing payment surface, tariffs, a facility listing, RIPE presence, public routing signals and a Dnipro locality that can matter under wartime conditions. It is not an empty shell in the public record. But the same record shows small absolute profit, thin public disclosure, an ambiguous operating perimeter and a capital-heavy promise if customers believe they are buying real continuity.

The right judgment is therefore conditional rather than diplomatic. DATASFERA can be valuable if it prices protected local continuity above cheap access, keeps utilization high, makes customers pay for power and support reality, and uses related-network infrastructure without hiding legal boundaries. It is vulnerable if it tries to fund data-center resilience from low household tariffs, if ALTAIR MEGA or other adjacent entities own the most important network assets without clear customer guarantees, or if public revenue growth masks replacement costs that have not yet hit the accounts.

For customers, DATASFERA should be treated as a local continuity provider to diligence, not as a cloud substitute to assume. For investors or creditors, the focus should be cash conversion, capex backlog and contract enforceability. For suppliers, the question is whether payment terms reflect currency risk. For the company itself, the incentive is blunt: every protected workload must either pay for the resilience it consumes or be served at a clearly bounded best-effort level. In wartime infrastructure, the operator that underprices continuity is not generous. It is transferring future repair cost onto its own balance sheet.

Sources