Summary
- Adamant is economically interesting because a Kyiv connectivity contract now buys more than bandwidth: it buys local repair labour, power continuity, data-centre hands, peering optionality and a provider relationship that may be cheaper to renew than to rebuild under pressure.
- The case is bounded. Adamant has credible local evidence in business PON, corporate fibre, colocation, secure internet, cloud and AS8788 peering, but it faces large-carrier substitution, mobile and satellite backups, cloud migration, electricity-cost pass-through and the risk that public evidence cannot prove every operational claim a buyer would want.
Renewal Is The Real Test
A useful way to price Adamant is to begin with a mundane procurement decision. A Kyiv accounting firm, clinic, logistics office, school vendor or public-sector contractor is coming to the end of an internet and hosting contract. It can renew a local Adamant line and perhaps keep a small server or backup service inside the provider's Kyiv data centre. It can move to a larger national carrier. It can buy a business mobile router as a second path. It can put the application stack into a foreign public cloud and reduce dependence on local hosting. It can run a self-managed office link with commodity equipment and an on-call technician. It can buy a satellite terminal for outages, even if the everyday traffic still goes over fibre. Those substitutes are not theoretical in wartime Ukraine; they sit inside ordinary renewal conversations because power cuts, route changes, repair queues and cyber incidents have made communications procurement less like an annual utility renewal and more like a resilience budget.
Adamant's own public positioning fits that problem. The company describes itself as one of Ukraine's early internet companies and dates its lineage to 1991, with the Adamant name appearing after a 1992 reorganisation (https://adamant.ua/en/company). Its home page and service pages present a Kyiv-based mix of business internet, home networks, domain registration, hosting, VPS, colocation, dedicated servers and data-centre services (https://adamant.ua/en/). The directory entry that anchors this article is ADAMANT, Ltd. (https://btw.media/en/directory/adamant-ltd-ua). The economic unit is therefore not a pure consumer broadband subscription. It is a Kyiv connection that may be tied to a router, a block of addresses, a hosted workload, a support queue, physical access to a server rack, and a relationship with engineers who know the local path.
The renewal decision turns on avoided cost. A larger Ukrainian carrier may offer broader national reach and a stronger balance sheet, but may not always deliver the same local support intimacy for a small Kyiv customer. A mobile backup may keep messaging alive but can be congested, power-dependent and unsuitable for all office systems. A cloud-only architecture can reduce dependence on a local server room but moves risk to foreign-platform billing, latency, identity controls and the quality of the last-mile connection that still reaches the user. A self-managed line may look cheaper until the buyer prices field labour, spares, firmware, monitoring and after-hours response. A satellite backup is useful when terrestrial links fail, but its economics are different from an always-on fibre connection and it has its own outage, capacity, policy and power risks. Adamant's renewal value is the difference between those substitute costs and the price of staying with a provider that can combine terrestrial access, hosting support and local operational knowledge.
The public evidence supports that bundle, but only to a point. Adamant advertises internet for business in Kyiv and Ukraine, with business tariff cards that include high-speed access, corporate networks and secure internet (https://adamant.ua/en/internet/for-businesses). Its broadband-access page describes PON connection for corporate clients, speeds up to 1 Gbit/s, connection priced from UAH 2,500 after survey, an IP-address block, equipment-configuration advice, round-the-clock technical support and monitoring, unlimited traffic and reserved access to world and UA-IX resources (https://adamant.ua/en/internet/broadband-access). Its corporate-networks page offers 100 Mbit/s to 10 Gbit/s and frames Adamant as designing, supplying equipment, installing, testing, documenting and maintaining fibre-optic company networks (https://adamant.ua/en/internet/fiber-optic-networks). Its secure-internet page offers a protected access node with DDoS, virus filtering, attack detection and other monitoring claims (https://adamant.ua/en/internet/secure-internet-site). Each of these is a sales statement, not an audited service record, but together they show the products Adamant wants buyers to price as more than raw bandwidth.
The procurement constraint is time. A Kyiv customer does not evaluate Adamant in a vacuum; it evaluates the cost of changing a working arrangement while managers are also buying batteries, generators, mobile failover, office security and cloud subscriptions. If the existing Adamant line is already in the building, the buyer may treat renewal as the lowest-disruption option even when a larger carrier has a more impressive national story. Conversely, if a procurement team has been forced to redesign continuity after outages or cyber incidents, it may require two terrestrial providers, a mobile route, a satellite terminal and a clear plan for moving workloads out of any single local facility. Adamant therefore wins when its local bundle reduces managerial work, and loses when the customer concludes that managerial work should be spent on reducing dependence on one provider.
That makes Adamant's contract value partly a labour-arbitrage story. The buyer is not simply deciding whether an Adamant engineer can run a cable cheaper than another engineer. It is deciding whether to rent a local operating routine: someone surveys the premises, knows the cabinet, monitors the access line, can talk to the data-centre desk, understands the customer's IP addressing and can coordinate a field response when the office loses power or the router fails. The rival routine could be stronger if it comes from a national carrier with deeper resources, or leaner if the customer has moved everything to cloud and only needs basic access. Adamant's defensible middle ground is the customer that is modern enough to need stable IP services but not so cloud-native that local hands no longer matter.
What The Company Actually Sells
Adamant's market surface is narrower than a national mobile operator and broader than a single-neighbourhood ISP. It sells home broadband, but the article's investment and resilience question is the business side: office PON, corporate fibre, secure internet, hosting, VPS, G-Cloud, colocation, dedicated servers, email, domain services and data-centre support. A buyer who only needs commodity home access can compare monthly prices. A buyer who needs a stable office connection, a public IP allocation, a hosted server, a support line and a predictable route to local exchange points is buying a more layered product.
The first layer is access. The official business page says the product set includes high-speed internet, corporate networks and secure internet, and describes internet connection for business in Kyiv and Ukraine (https://adamant.ua/en/internet/for-businesses). The broadband page is unusually useful because it exposes a few economic levers. A corporate PON connection is not a one-click activation; Adamant says the connection cost starts from UAH 2,500 and is calculated individually after surveying the customer's site (https://adamant.ua/en/internet/broadband-access). That survey language matters. It means the service cost depends on where the buyer sits relative to Adamant's plant, what building access is available, what cable path can be used, how much labour is needed, and whether the provider can reuse existing ducts, risers or termination points. In a calm market those are ordinary installation details. Under pressure they become a renewal moat: once the fibre is inside the building and the team knows the local path, switching to a different provider can mean another survey, another building-access negotiation, another waiting period and another failure mode.
The second layer is corporate-network construction. Adamant advertises corporate networks from 100 Mbit/s to 10 Gbit/s and says it can design, deliver equipment, install, start up, test, document, provide warranty and post-warranty maintenance, and support all elements of the system (https://adamant.ua/en/internet/fiber-optic-networks). This is the part of the value proposition that competes with self-management. A small institution could buy switches, contract a local installer and run its own intranet connections between offices, but then it owns the repair burden. Adamant's product language says the buyer can shift part of that burden to a local provider. The price is a recurring relationship rather than a one-off equipment purchase.
The third layer is security. Adamant's secure-internet page describes access separation, protection against unauthorised access, data destruction, modification and blocking, registration of threat attempts, DDoS protection, virus protection, traffic cleaning and application-layer attack monitoring (https://adamant.ua/en/internet/secure-internet-site). A buyer should not treat marketing text as proof of actual attack performance, but it tells us why a Kyiv enterprise might renew: it may want the same provider to carry the line, filter traffic, observe abnormal flows and support the site when something breaks. If a larger operator offers a cheaper line without the same local support context, the headline bandwidth price can understate the total operating cost.
The fourth layer is hosting and colocation. Adamant's data-centre page lists hosting, Cloud VPS, G-Cloud, dedicated servers, colocation, domains, mail and multimedia services (https://adamant.ua/en/data-center). The colocation page prices 1U server placement from UAH 945 per month with a 100 Mbit/s port and one IP address, with larger rack options, power assumptions and additional service charges shown separately (https://adamant.ua/en/data-center/colocation). The VPS page shows a low-entry cloud-server product and highlights unlimited traffic and round-the-clock support (https://adamant.ua/en/data-center/vps). The G-Cloud page says Adamant offers a protected cloud service to government entities and says its data centres are located exclusively in Ukraine, with the ability to place and duplicate cloud resources in several geographical areas (https://adamant.ua/en/data-center/g-cloud). These claims create a distinct procurement problem: Adamant can be the access provider, the hosting provider and the support provider. That concentration can be convenient and sticky, but it also creates dependence on one local operating team.
Kyiv Connection Economics Are Built On Distance
The cheapest way to misunderstand Adamant is to treat Mbps as the main unit of competition. In Kyiv, the real cost can be distance. The distance from Adamant's fibre to the customer's building determines whether a new line is easy or labour-intensive. The distance from a field crew to the fault determines repair time. The distance from the customer's equipment to a data-centre technician determines whether a reboot, cable move or disk swap takes minutes, hours or days. The distance from Adamant's network to local exchange points affects latency and transit economics. The distance from a buyer's office to a substitute carrier's closest feasible route determines whether switching is a procurement decision or a construction project.
Adamant's business broadband page exposes that distance cost by making connection conditional on a site survey and by stating that PON access is possible only in territory covered by Adamant's network (https://adamant.ua/en/internet/broadband-access). Its corporate-network page says Adamant has its own extensive fibre-optic network and that its specialists continuously expand it (https://adamant.ua/en/internet/fiber-optic-networks). Those statements should not be converted into a map of actual building coverage; the public page does not provide a verifiable GIS layer. But they are enough to identify the business model. Adamant monetises local network reach and local engineering capacity, not just wholesale transit resale.
This is where national-carrier substitution becomes more nuanced. A larger Ukrainian operator can often offer broader coverage, more national backbone options and a stronger capital base. The buyer may conclude that it wants that scale after a bad outage or a stalled repair. But a national carrier may still need to reach the exact building, negotiate the same risers, schedule the same field visit and support the same customer equipment. If Adamant already has the line in place, the incumbent's advantage is the avoided cost of change. The larger carrier is not simply cheaper or safer; it has to be safer enough to compensate for migration risk.
Mobile backup also looks different when measured against distance. A 4G or 5G router can bypass a cut last-mile fibre, but it depends on cell-site power, radio congestion, spectrum loading and indoor signal quality. It may be enough for messaging, payments and cloud dashboards; it may not be enough for a video-heavy office, a large file workflow, an on-premises server or a security camera system. DataReportal's 2025 Ukraine report shows a digitally mature market with 31.5 million internet users and 82.4% internet penetration at the start of 2025, as well as median fixed download speeds of 83.81 Mbit/s and median mobile download speeds of 31.23 Mbit/s (https://datareportal.com/reports/digital-2025-ukraine). Those national medians do not price Adamant directly, but they frame the substitution: fixed lines still matter for stable office throughput, even in a country with high mobile penetration.
Cloud-only architecture is another substitute with hidden distance. Moving a server out of Adamant's colocation rack and into a public cloud can reduce dependence on local power and local hardware repair. But the office still needs a line to reach the cloud. Identity, backup, storage egress and regulatory comfort still need to be managed. If the buyer has Ukrainian-data preferences or needs local staff who can touch equipment, a pure cloud move can create new frictions. Adamant's G-Cloud claim that its data centres are exclusively in Ukraine and that it targets government entities gives it a public-sector continuity angle, though buyers would still need to verify certifications, redundancy and contractual terms directly (https://adamant.ua/en/data-center/g-cloud).
The distance economics also shape repair priority. A provider that has concentrated customers in a known Kyiv footprint can plan spares, ladders, access permissions and engineering routes differently from a provider serving a thin scatter of sites. That can help Adamant if the customer is near its established plant and data-centre ecosystem. It can hurt Adamant if the customer is near a rival's stronger route or if building owners favour another carrier's access equipment. In renewal negotiations, the buyer should ask which ducts, risers, optical splitters, aggregation nodes and power points sit between the office and the provider core. Public pages cannot answer those site-specific questions, but they are the questions that turn a nominally cheap line into a resilient or fragile one.
Power Resilience Is No Longer A Back-Room Detail
The power question is central because Ukraine's communications reliability is tied to electricity. The International Energy Agency said in its September 2024 report that Russia's attacks on Ukraine's energy system had intensified, creating risks to power, heating and communications services (https://www.iea.org/reports/ukraines-energy-security-and-the-coming-winter). The IEA's executive summary said rolling blackouts and unscheduled interruptions had become normal in badly affected regions, with summer 2024 generation capacity more than 2 GW below peak demand of 12 GW and a possible winter deficit as high as 6 GW under stress assumptions (https://www.iea.org/reports/ukraines-energy-security-and-the-coming-winter/executive-summary). Its chapter on the energy system under attack said the August 26, 2024 strike involved more than 200 missiles and drones, that around 8 million households lost power without warning, and that Kyiv had its first unscheduled blackout since November 2022 (https://www.iea.org/reports/ukraines-energy-security-and-the-coming-winter/ukraines-energy-system-under-attack).
For Adamant, that context does not prove a specific outage. It changes how the buyer values the line. A business fibre connection that depends on mains power at the customer premises, in the building riser, at intermediate active equipment, and at the provider node is not resilient merely because it is optical. PON can help because passive optical distribution between active endpoints can keep more of the street path unpowered, but the customer's terminal, router, Wi-Fi and provider-side equipment still need electricity. The economic question becomes: who has backup power, how long does it last, what happens when the blackout schedule changes, and which party has a technician who can get to the fault?
Adamant's data-centre page gives some concrete power evidence. It says the data centre has two independent grid inputs and an independent input from a diesel generator, two guaranteed power inputs with load up to 120 kW, the ability to connect UPS power from two different inputs with allocated capacity up to 6 kW per cabinet, and monitoring of more than 250 power-supply parameters (https://adamant.ua/en/data-center). The same page says the data-centre modules use N+1 air conditioning and 24-hour monitoring of temperature and humidity. These details are material because hosting and colocation customers do not only buy rack space; they rent a power and cooling envelope that would be expensive to reproduce in an office basement.
The colocation page makes the cost pass-through more visible. Adamant lists rack products with port speeds and power allowances, and it states that electricity consumed by customer equipment can be calculated at the supplier's market price for electricity (https://adamant.ua/en/data-center/colocation). That matters for margins. When electricity becomes scarce or expensive, hosting economics are not fixed. A buyer may renew because Adamant's data-centre environment is more resilient than the customer's office, but the buyer may also face changing energy pass-through, generator terms, or power caps. Adamant's own news page includes a March 2026 item about updated conditions for using a diesel generator, which signals that generator economics are active customer terms rather than invisible infrastructure (https://adamant.ua/en/company/news).
The power context also explains Adamant's winter retail messaging. In November 2025 the company promoted an "Internet without light" offer for new subscribers, centred on PON connection, a low promotional monthly price, UAH 1 connection, 24/7 support and online continuity during cold and dark conditions (https://adamant.ua/en/company/news/promotion-internet-without-light-from-adamant). That is not proof of network-wide uptime. It is a bounded market signal: Adamant sees blackout resilience as a selling point, and it is using PON migration to defend or grow the subscriber base. A later promotion extended a UAH 1 switch to optics for existing Ethernet subscribers (https://adamant.ua/en/company/news/extension-of-the-promotion-switch-to-optics-for-uah-1-for-existing-subscribers). The commercial signal is clear. Power resilience has moved from engineering detail to advertised benefit.
The power-backup economics are also asymmetric. A household or small office can buy a modest battery and keep an optical terminal, router and laptop alive for part of a blackout, but it cannot easily reproduce the cooling, fuel, monitoring and access discipline of a data-centre rack. A larger enterprise can buy generators and UPS systems, but then it must manage maintenance, fuel contracts, noise, exhaust, insurance and staff coverage. Adamant's data-centre proposition competes with that self-supply burden. The customer pays a provider to aggregate power resilience across many racks. The risk is that the same aggregation creates a common point of stress if fuel, grid inputs, generator rules or cooling constraints tighten.
The exact commercial effect depends on who bears the next hryvnia of resilience spending. If Adamant absorbs battery replacement, generator maintenance and power engineering while retail tariffs remain promotional, margins compress. If it passes electricity and generator costs to colocation users, churn risk rises among price-sensitive customers. If business customers accept higher invoices because downtime is more expensive than the bill increase, Adamant's resilience spending can defend revenue. The public tariff pages show the terms in outline, especially the colocation treatment of electricity at the supplier's market price, but they do not disclose whether customers actually renew after power-related bill changes (https://adamant.ua/en/data-center/colocation).
This is why a simple blackout narrative is not enough. The relevant question is duration. A two-hour outage can be handled with office batteries and provider UPS. A longer scheduled cut tests generator availability, fuel logistics, staff cover and cooling headroom. Repeated cuts test maintenance intervals and component fatigue. Unscheduled cuts test monitoring and after-hours response. Adamant's public claims about two grid inputs, diesel-generator input, UPS options and monitored power parameters are meaningful because they point to these layers, but the thesis would be much stronger with observed runtime, fuel-replenishment history and customer outage statistics.
Routing Evidence Shows Optionality, Not Invulnerability
The public routing record supports Adamant's claim that it is not a single-upstream local reseller. RIPEstat's AS overview identifies AS8788 as "ADAMANT ADAMANT, Ltd." and marks it as announced (https://stat.ripe.net/AS8788). The RIPE WHOIS data for AS8788 lists internet exchange relationships with Giganet, DTEL-IX and UA-IX, uplinks through Cogent, UARNet and ETT, parity with Vodafone, Google, Datagroup, ColoCALL, United Networks of Ukraine, Ukrainian Telecommunication Group and others, and various customer or other peerings (https://stat.ripe.net/data/whois/data.json?resource=AS8788). PeeringDB lists Adamant as AS8788, type Cable/DSL/ISP, with an AS-ADAM-UA IRR set, traffic in the 5-10 Gbit/s range, open peering policy, European scope, three exchange points and four facilities (https://www.peeringdb.com/api/net?asn=8788).
That is useful evidence, but it has to be interpreted carefully. AS records do not show the lived performance of every customer connection. A route object does not prove spare field crews. PeeringDB is user-maintained and can lag reality. Still, routing data tells us whether the provider has multiple paths and local interconnection options. PeeringDB's exchange-point data shows Adamant at UA-IX, DTEL-IX and Giganet IXN with 10 Gbit/s ports in the listed records (https://www.peeringdb.com/api/netixlan?net_id=1755). Its facility data lists Ucomline Kiev, Datagroup Kiev, TSUA and Adamant, LLC, all in Kyiv (https://www.peeringdb.com/api/netfac?net_id=1755). RIPEstat's announced-prefixes data shows visible AS8788 prefixes including 212.80.32.0/19, 212.26.128.0/19, 91.205.16.0/22, 178.211.96.0/19, 194.32.108.0/23 and IPv6 space during the late June to July 2026 window (https://stat.ripe.net/data/announced-prefixes/data.json?resource=AS8788).
The company-side data-centre page aligns with the routing record. Adamant says its data-centre network access has 160 Gbit/s channel capacity, optical inputs from two opposite sides, connections to UA-IX, DTEL-IX and Giganet IX by two alternative channels, and parity channels with Google and Facebook platforms and other telecom operators (https://adamant.ua/en/data-center). That overlap between the official page and independent databases improves confidence in the general architecture. It does not prove every redundancy claim under attack, but it reduces the probability that Adamant is simply a last-mile brand with no meaningful interconnection posture.
The distinction between optionality and invulnerability is important. Multiple IXPs and upstreams can improve route diversity, reduce transit cost and allow traffic to move around certain failures. They do not solve building power, local fibre cuts, customer-premises battery depletion, cyber misconfiguration, or a citywide blackout that exhausts generator fuel. The 2022 measurement paper "Internet Performance in the 2022 Conflict in Ukraine" found that internet performance in Ukraine degraded after the invasion, with fixed and mobile speeds falling, latency rising and packet loss increasing, while routing paths changed for many RIPE Atlas probes (https://arxiv.org/abs/2205.08912). That study was about Ukraine broadly, not Adamant specifically. Its relevance is methodological: routing resilience is measurable at the network level, but customer experience still depends on the physical and power layers beneath it.
UA-IX also deserves careful treatment. Adamant says it was among the initiators of the Ukrainian Internet exchange point (https://adamant.ua/en/company). RIPEstat's AS15645 record is for the Ukrainian Internet Exchange itself, not Adamant, and lists AS8788 as one of many parties accepted in the exchange's import policy (https://stat.ripe.net/AS15645). That means UA-IX is evidence of Adamant's interconnection environment, not an Adamant asset to be counted as the company. Treating the exchange as evidence only avoids overstating Adamant's control surface.
Upstream dependence is therefore a priced risk rather than a binary flaw. RIPE's AS8788 record lists multiple named upstreams and peerings, including Cogent, UARNet, ETT, Google, Vodafone, Datagroup and others (https://stat.ripe.net/data/whois/data.json?resource=AS8788). That improves the case against single-carrier fragility, yet it does not remove dependence on external networks. Cogent reach, domestic academic or commercial routes, Ukrainian exchange fabrics, cache platforms and local peers all have their own maintenance windows, commercial terms and wartime exposure. If a customer needs specific international reach, a local IXP path may not be enough. If a customer mostly serves Ukrainian users, domestic exchange and cache relationships may be more valuable than a large foreign transit pipe.
The most useful private routing metrics would be precise and operational: peak and 95th-percentile utilisation by upstream and exchange port, packet-loss and latency by destination class, failover time when a primary upstream is withdrawn, cache-hit contribution by large platforms, number of customer prefixes with route-object hygiene problems, and the share of traffic that can remain domestic during international-path stress. Public databases show Adamant's route presence and some declared relationships. They do not show congestion, contractual headroom or the operational quality of failover. That is why the public record supports optionality, not invulnerability.
For an enterprise buyer, route diversity matters most when mapped to applications. A media company moving large files, a clinic reaching cloud software, a school platform serving local users and a public contractor using domestic government systems may value different routes. Adamant can look stronger where local exchange reach, Ukrainian hosting and support hands reduce friction. It can look weaker where the buyer needs guaranteed international performance, multi-region disaster recovery or carrier-neutral separation between access and hosting. The routing evidence is best read as a menu of resilience ingredients, not a final service-level answer.
Hosting Makes The Contract Stickier
Adamant's data-centre product changes the switching calculus because connectivity and compute can be bundled. A customer that only buys access can change provider if a better line appears. A customer that keeps a server, mail system, backup storage, domain account or managed VPS with the same provider has a more complex migration. It must schedule data transfer, DNS changes, firewall updates, certificate checks, IP address changes and perhaps physical hardware movement. Those frictions are not always bad. They can make the incumbent's support knowledge valuable. But they also increase concentration risk.
The official data-centre page is detailed enough to analyse the economics. Hosting starts from low monthly prices, VPS tiers show small entry packages, dedicated servers start at a higher monthly price, and colocation includes a port, an IP address and round-the-clock support (https://adamant.ua/en/data-center). The colocation page exposes add-ons such as ports to Adamant equipment, dedicated global-resource channels, additional power, extra wired connections, OS installation, software support and administration hours (https://adamant.ua/en/data-center/colocation). This is a classic regional-provider margin stack. The base product gets the customer into the facility; higher-margin services monetise hands, ports, administration, power and customer uncertainty.
In wartime pressure, hands matter. If an office server fails during a blackout, the cost is not only the replacement disk. It is the ability to reach the device, diagnose it, power-cycle it, swap parts and confirm application recovery. Adamant says its data-centre services include IP-KVM connection on request, visual control with photo and video fixation, primary diagnostics and equipment reboot by the duty system administrator (https://adamant.ua/en/data-center). A buyer with no in-house infrastructure team may price that as insurance. A buyer with a mature cloud architecture may price it as unnecessary legacy complexity.
The hosting question also affects public-sector and institution customers. Adamant's G-Cloud page says the service is aimed at government agencies, references protected information and a State Special Communications certificate received at the beginning of 2022, and says Ukrainian-only data-centre location can support secure internet access and cloud duplication (https://adamant.ua/en/data-center/g-cloud). Public-sector buyers, regulated contractors and local institutions often care where data is stored, who can access equipment and how quickly a local support team can respond. The evidence does not let us conclude which agencies use the service or how material the revenue is. It does let us conclude that Adamant is positioning around sovereign/local cloud trust rather than trying to outscale global hyperscalers.
The counterargument is that hosting stickiness can become a liability. If a buyer can rebuild in a major public cloud, use a larger national carrier for access and keep a mobile or satellite backup for continuity, Adamant's integrated bundle may look like local concentration rather than resilience. The right judgement depends on the application. A website with no locality requirement may not need a Kyiv VPS. A payment terminal, school platform, camera system, municipal workflow or office file server may still benefit from a local provider that can combine access, hosting and hands.
Enterprise hosting demand is especially sticky when it is attached to old software. Many organisations do not have a clean portfolio of stateless web applications. They have accounting systems, local databases, file shares, access-control servers, camera archives, mail habits and small vendor applications that were not designed for a sudden move to a hyperscale cloud. For these customers, Adamant's VPS, colocation and support products can function as a practical halfway house: more controlled than an office closet, less abstract than a foreign cloud account, and close enough for local engineers to intervene. That demand may not be fashionable, but it can be resilient because the cost of rewriting the application stack is higher than the cost of renewing the hosting contract.
The demand is not unlimited. A modern enterprise that has already standardised on cloud identity, SaaS collaboration, managed endpoint security and remote-work workflows can treat local hosting as a legacy exception. A highly regulated buyer may require a more formal certification and audit trail than Adamant's public pages disclose. A fast-growing software company may want multi-region automation that a regional data centre cannot provide on its own. Adamant's best-fit market is therefore neither the most basic consumer nor the most cloud-native multinational. It is the Kyiv organisation that needs continuity, local support and enough infrastructure control to make a nearby provider useful.
Exact private demand metrics would settle much of the uncertainty: renewal rate by product, average revenue per business access line, colocation cabinet occupancy, VPS churn, number of customers buying both access and hosting, percentage of revenue with support or administration add-ons, share of tickets resolved remotely versus on site, and revenue concentration in the top 20 customers. Without those figures, the article can only infer stickiness from product breadth and switching friction. Those inferences are reasonable, but they should remain bounded.
Competition Is Stronger Than The Price Sheet Shows
Adamant competes on at least five axes. The first is larger Ukrainian carriers. They can offer national sales teams, bigger capital programmes, broader mobile assets and, in some cases, integrated fixed-mobile backup. The second is other Kyiv ISPs and data centres with their own fibre footprints. The third is mobile substitution. The fourth is satellite backup. The fifth is cloud migration.
Kyivstar is the obvious scale comparator. Public reporting on the December 2023 Kyivstar cyberattack showed the operational stakes of national-carrier dependence; Wired reported that a cyberattack on the operator cut communications for millions and affected Kyiv's air-raid warning system, citing statements from Ukrainian computer-emergency authorities and Kyivstar leadership (https://www.wired.com/story/ukraine-kyivstar-solntsepek-sandworm-gru). Reuters also reported the attack as a major wartime telecom incident (https://www.reuters.com/world/europe/ukraines-top-mobile-operator-hit-by-biggest-cyberattack-war-2023-12-12/). The lesson for Adamant is double-edged. A large carrier may have more resources, but scale itself can become a systemic target. A smaller provider may offer diversity from a dominant operator, but it may have fewer resources after a severe incident.
Ukrtelecom is another useful comparator because it represents national fixed-line depth. The public record around Ukrtelecom's wartime operation and restoration work shows that large fixed operators can keep broad networks running, invest in optical expansion and repair damaged hubs, but it also shows that repairs are labour-intensive and capital-consuming. Adamant does not need to beat such operators nationally. It needs to be good enough in its chosen Kyiv operating surface that a buyer values local responsiveness over broader national scale.
Mobile backup changes the pricing floor. A small office can add a mobile router and keep working when the fixed line fails, especially if most workloads are cloud-hosted. That reduces the willingness to pay a high premium for fixed-line resilience. But mobile backup is often best as a second path, not a full substitute. When power cuts hit base stations, when many users move onto mobile data at once, or when indoor signal is poor, the mobile path can degrade. The same is true of satellite backup. It is valuable when terrestrial paths fail and was widely used in Ukraine during the war, but it requires terminal power, clear placement, subscription budget and acceptance of a different reliability model. A buyer that uses mobile or satellite as a backup may still renew Adamant as the primary line.
Cloud-only architecture is the most strategic substitute. If a Kyiv enterprise removes local servers, moves identity and files to a global platform, and trains staff to work from any connection, the local ISP becomes less central. The buyer can spread access across fibre, mobile and satellite. That is the case against Adamant's hosting bundle. The case for Adamant is that many organisations do not make that transition cleanly. Legacy applications, local data needs, budget limits, security concerns, staff habits and procurement rules keep part of the workload local. For those buyers, an integrated regional ISP and data-centre provider remains relevant.
Competition is therefore not a single market-share chart. It is a question of which failure the buyer fears most. If the feared failure is one local fibre cut, the buyer may want a second carrier or mobile backup. If the feared failure is office power, the buyer may want Adamant's data centre. If the feared failure is a provider cyber incident, the buyer may want carrier diversity and off-provider backups. If the feared failure is cloud lock-in or foreign-platform billing, the buyer may want a local provider. Adamant's renewal risk rises when customers decide that the most likely failure is Adamant-specific and falls when they decide that switching creates more risk than it removes.
Starlink-style backup sits in a different category from the carrier substitutes because it is not mainly a price competitor for everyday fibre. It is a continuity option for the moment when terrestrial routes, building power or local access fail. That makes it powerful in procurement meetings: a manager can argue that a satellite terminal reduces the need to overpay for terrestrial resilience. Yet satellite backup still requires power, installation discipline, traffic prioritisation and a decision about which systems are allowed to run when capacity is constrained. For an enterprise that needs predictable latency, fixed public addressing, local hosting, large file movement or low-friction support, satellite is a complement rather than a full replacement.
The larger carrier option has the opposite profile. It may be a direct replacement for Adamant as the primary line, especially if the carrier can reach the building quickly and bundle mobile failover. Its weakness is concentration at national scale. The Kyivstar cyberattack comparator matters because it shows that a large operator can be both better resourced and more systemically exposed (https://www.reuters.com/world/europe/ukraines-top-mobile-operator-hit-by-biggest-cyberattack-war-2023-12-12/). The Adamant customer does not have to believe a smaller provider is safer in every respect; it only has to believe that diversity across providers lowers the risk of all communications failing at once.
Mobile backup is the cheapest perceived substitute and often the least understood. It can be bought quickly, tested by non-specialists and justified to finance teams as a modest recurring cost. But it moves the bottleneck to radio coverage, battery-backed base stations, SIM management, traffic caps, indoor antennas and failover configuration. A poorly configured mobile backup can create false confidence. Adamant's renewal defence is stronger if it can integrate fixed access with a realistic customer backup plan, rather than pretending the fibre line is sufficient by itself.
Public-Sector Continuity Is A Plausible Angle, Not A Proven Revenue Pillar
Adamant's routing record includes peerings or customer-style entries that name public institutions, including DSTSIP, Rada, Cabinet of Ministers-related shorthand, the Ministry of Economy of Ukraine and the National Security and Defence Council staff in the RIPE AS8788 record (https://stat.ripe.net/data/whois/data.json?resource=AS8788). Those entries should be handled with caution. RIPE records can include historical, policy and routing relationships that do not equal current paid revenue, and some labels are shorthand rather than contracts. But the presence of public-institution names in the routing record is consistent with Adamant's positioning around secure access, G-Cloud and local infrastructure.
The continuity logic is straightforward. Public institutions and contractors need connections that survive imperfect conditions. They may need Ukrainian location, local technical support, predictable routing to domestic networks, and hosting options that do not require shipping equipment abroad. Adamant's secure internet and G-Cloud pages speak to that market (https://adamant.ua/en/internet/secure-internet-site and https://adamant.ua/en/data-center/g-cloud). The company's data-centre access controls, 24/7 security, physical access rules, diesel-generator input, UPS options and network interconnection claims also fit institutional procurement concerns (https://adamant.ua/en/data-center).
The proof gap is revenue materiality. Public evidence does not show what share of Adamant's revenue comes from public-sector, government-adjacent or critical-infrastructure customers. It does not show contract terms, churn, service-level credits or renewal rates. A conservative judgement should say that public-sector continuity is a credible angle, not a proven pillar. It can explain why Adamant would keep a sticky customer base; it cannot be used to assert a specific earnings quality.
There is also a political and regulatory dimension. Telecommunications providers in wartime Ukraine operate in a sector where service continuity is socially important and where government, security and regulator expectations matter. That can support demand, but it can also raise compliance burden. Adamant's documents page lists contracts, rules and domain policies, while its services include domains and hosting (https://adamant.ua/en/company/documents). The more Adamant serves institutions, the more it must maintain documentation, access control, security practice and continuity procedures that smaller consumer-only ISPs may not need.
Repair Labour Is The Scarce Input
The most important operating cost may be neither transit nor routers but people. Fibre breaks, building access problems, power-equipment failures and customer-premises faults all require labour. Under wartime pressure, repair labour is harder to schedule because staff safety, travel routes, curfews, power cuts, fuel, spare parts and simultaneous faults can constrain the day. A provider with local technicians can be valuable precisely because the work cannot be fully centralised.
Adamant's corporate-network page promises installation, start-up, testing, documentation, warranty and post-warranty maintenance, and service support for all elements of the system (https://adamant.ua/en/internet/fiber-optic-networks). Its business broadband page includes consultative help in equipment setup, 24/7 technical support and monitoring (https://adamant.ua/en/internet/broadband-access). Its contacts page separates home-network support, general issues, internet for business, technical support and data-centre services, with technical support shown as 24/7 (https://adamant.ua/en/contacts). These details do not prove staffing levels, but they show that support labour is part of the product.
That labour is also the lever that could break the thesis. If repair queues lengthen, if technicians cannot reach buildings, if spare parts become expensive, if customer tickets pile up, or if generator and power issues consume engineering time, the local-support advantage weakens. Larger carriers can throw more resources at some problems; cloud providers can avoid physical local repair for many workloads; mobile and satellite backup can keep a customer partially online while waiting. Adamant's renewal defence is strongest when its local repair performance is visibly better than the customer's alternatives.
The public article cannot verify repair times. Adamant's site has planned technical-work notices and service promotions, but no public monthly uptime dashboard, mean-time-to-repair data, or outage-credit history. Customer forums and social media can reveal frustration, but they are noisy and often skew toward complaints. The defensible use of such signals is as watchpoints: if chatter about repair delays rises across multiple channels and matches visible outage reports, it would weaken the case. If the company keeps promoting optical migration and blackout-oriented access without widespread public complaints, that is a softer positive signal, not proof.
Repair labour also has a queueing problem. A single isolated fault can be solved with one team, one spare optical terminal and one building visit. A citywide power event or a storm of customer-premises failures can turn the same labour pool into a triage system. The economically relevant question is not whether Adamant has technicians; it is how many simultaneous faults it can absorb before response time deteriorates. A national carrier may have more teams but a larger affected base. A smaller local provider may have fewer teams but shorter distances and better knowledge of the buildings it serves. The outcome is local and empirical.
The private metrics that would clarify this section are direct: number of field technicians available per shift, spare optical terminals and routers on hand, median and 90th-percentile repair time by fault type, percentage of faults caused by customer power versus provider plant, number of repeat faults per 100 business circuits, and repair performance during scheduled and unscheduled blackout periods. Procurement teams should also care about escalation rules. A business customer buying hosting and access from the same provider needs to know whether a failure is handled as one coordinated incident or split between separate support queues.
Unofficial market signals can help but should not dominate. A burst of social complaints after a known city outage is not the same as structural service failure. A lack of complaints may reflect satisfied customers, low public visibility or customers who complain in private channels. The useful pattern would be repeated, specific, multi-source evidence: recurring mentions of long repair delays in Adamant-covered buildings, repeated power-related failures despite PON migration, or visible customer movement to larger carriers after outages. In the absence of that pattern, the evidence supports a watchpoint rather than a negative conclusion.
The Margin Story Depends On Bundling And Pass-Through
Adamant's margin mix is likely better when it sells bundled services than when it sells a standalone low-price access line. Access alone is exposed to price competition. Hosting, colocation, dedicated servers, managed support, secure internet and corporate-network buildout allow the company to charge for expertise, power, ports, IP addresses and convenience. The official data-centre pages show this ladder: low-cost hosting, VPS, dedicated servers, colocation, additional ports, administration, monitoring and power terms (https://adamant.ua/en/data-center and https://adamant.ua/en/data-center/colocation).
Power pass-through is a margin defence and a churn risk. If electricity costs rise, a data-centre operator needs to recover them. Adamant's colocation page says customer-equipment electricity can be calculated at the supplier's market price (https://adamant.ua/en/data-center/colocation). That protects the provider from absorbing all power volatility, but it makes bills less predictable for customers. A customer comparing Adamant to cloud-only architecture may see local colocation as attractive when it values physical control and local hands, but unattractive when power costs are hard to forecast.
Transit and peering can also influence margin. PeeringDB's traffic estimate of 5-10 Gbit/s and open peering policy, plus Adamant's presence at UA-IX, DTEL-IX and Giganet, suggests that local exchange participation is part of its cost and performance model (https://www.peeringdb.com/api/net?asn=8788 and https://www.peeringdb.com/api/netixlan?net_id=1755). Adamant's own data-centre page claims 160 Gbit/s of channel capacity and parity channels with large platforms (https://adamant.ua/en/data-center). More local traffic exchange can reduce paid transit and improve user experience for domestic and cache-heavy traffic. But the evidence does not disclose traffic costs, utilisation, port charges or settlement terms, so this remains an operating hypothesis rather than a quantified margin estimate.
Promotions reveal pressure. The "Internet without light" offer prices a new-user promotional plan at UAH 120 per month for seven months and then moves users to another tariff after the promotional period (https://adamant.ua/en/company/news/promotion-internet-without-light-from-adamant). The optical-switch promotion offers existing Ethernet subscribers migration to optics for UAH 1 during the promotion window (https://adamant.ua/en/company/news/extension-of-the-promotion-switch-to-optics-for-uah-1-for-existing-subscribers). These are sensible defensive moves, but they also show competition for subscribers and the need to absorb connection or migration costs to keep the base. The enterprise story is stronger if business and data-centre revenue offsets promotional pressure in consumer or small-office access.
The most revealing market signal is not the advertised promotional price; it is what happens after the promotion. If customers stay after the low-price period, Adamant has probably converted blackout anxiety into durable optical demand. If they churn after the initial discount, the campaign is more defensive and less valuable. The same applies to existing Ethernet customers moved to optics for UAH 1. The migration can lower future support cost and improve outage performance, but only if it reduces trouble tickets and protects average revenue. A cheap migration that merely shifts customers to a lower-margin product without improving retention would be weaker economics.
Private margin metrics would change the judgement quickly: gross margin by business access, home PON, colocation, VPS and dedicated servers; electricity and fuel cost as a percentage of data-centre revenue; support tickets per UAH of recurring revenue; average installation payback period after survey-based connections; proportion of customers with more than one Adamant product; and discount-to-renewal conversion after blackout-related promotions. Public evidence lets us see the product ladder and the pressure points. It does not show whether the ladder produces attractive cash flow.
What Would Change The Judgement
The bullish case is that Adamant is a resilient Kyiv operator with a relevant bundle: PON and corporate fibre for access, local data-centre infrastructure for hosting, AS8788 routing optionality for performance, secure internet for risk-sensitive buyers, and support labour that is harder to substitute than bandwidth. In that case, wartime pressure increases the value of a known local provider because customers renew to avoid migration and repair risk. The public evidence from Adamant's pages, RIPEstat and PeeringDB supports the existence of that bundle (https://adamant.ua/en/internet/for-businesses, https://stat.ripe.net/AS8788 and https://www.peeringdb.com/net/1755).
The bearish case is that the same pressure accelerates substitution. Customers may decide that larger carriers, mobile backup, satellite backup and cloud migration are a better resilience mix. They may keep Adamant only as one access path rather than the main provider. Power pass-through may make local hosting less attractive. Repair queues may expose the limits of a regional operator. Routing diversity may not matter if the failure is local power, building access, or customer equipment.
Three facts would improve the judgement. First, customer-level renewal and churn data split between business access, home access, hosting and colocation would show whether the bundle is sticky or merely broad. Second, measured uptime and repair-time data during blackout periods would show whether Adamant's local support advantage is real. Third, facility-level power autonomy, fuel logistics, battery runtime and generator-use terms would show whether the data-centre promise survives long outages rather than only ordinary utility interruptions.
Three facts would weaken it. First, sustained public evidence of slow repairs or recurring blackout-era service failures would damage the renewal thesis. Second, aggressive price-led offers from larger carriers in Adamant-covered buildings would erode the incumbent advantage. Third, visible migration of local institutional workloads to cloud or national-carrier infrastructure would reduce the value of Adamant's hosting and secure-access pitch.
The best current conclusion is deliberately bounded. Adamant is not a national substitute for Ukraine's largest operators, and public sources do not support a claim that it can withstand every wartime stress. But it is also not just a small access reseller. Its official service pages, AS8788 record, PeeringDB presence, data-centre infrastructure claims and PON migration messaging all point to a Kyiv operator whose economics depend on the practical work of keeping customers connected when power, repair routes and procurement confidence are under strain. That makes the renewal decision more interesting than the monthly tariff. Adamant's connection carries repair risk under pressure, but that risk is also the reason some customers may keep paying for a local provider that knows the route, the cabinet, the rack and the phone number to call when the lights go out.