Summary
- NOVATEL is not merely a holder of Internet number resources. Magyar Telekom's 2025 annual report identifies it as a wholly owned Bulgarian wholesale telecom subsidiary, while the company reports a national DWDM network with 39 points of presence and cross-border connections to five neighbouring countries.
- Its network has observable operating depth: RIPE routing data showed 29 originated prefixes and 23,040 announced IPv4 addresses on 10 July 2026, while PeeringDB listed regional traffic of 50-100 Gbps, public exchange connections in Sofia and Frankfurt, and six interconnection facilities.
- Bulgarian regulator data validate demand but also expose weak bargaining power. NOVATEL grew national wholesale leased-line volume by 31.3% and revenue by 28.2% in 2024, yet its share of wholesale leased-line revenue slipped from 26.4% to 25.1% as Neterra advanced to 33.7%.
- The longer comparison is colder. NOVATEL held 43.0% of Bulgarian wholesale leased-line revenue in 2020 but 25.1% in 2024. The market expanded, yet its implied annual revenue from that segment was roughly BGN 6 million in both years.
- The latest detailed company figures show the difference between growth and value creation. For 2024, operating revenue was BGN 25.0 million, operating profit BGN 1.1 million and net profit BGN 625,000, equal to an operating margin of about 4.3% and a net margin of about 2.5%.
- The strategic conclusion is explicit: NOVATEL has a defensible regional operating franchise, not a demonstrated economic moat. Its best chance to improve returns is to sell scarce routes, long-lived capacity and accountable managed service around the network; adding undifferentiated bandwidth or project revenue without committed demand would leave it a price-taker with a larger cost base.
Relevance has to be earned before scale can be claimed
NOVATEL's economic incentive is simple. It must keep enough customers dependent on its routes, response times and technical accountability to prevent its network from becoming a wholesale commodity. The task is harder than adding capacity. Capacity becomes cheaper. Customer traffic grows, but the cost per transmitted bit tends to fall faster. Public cloud services shift applications away from customer premises, which increases the need for dependable connectivity while also making it easier to run those applications over standard Internet access. A carrier can therefore carry more traffic and still capture less value from each unit.
This is the distinction between revenue growth and value creation. A new circuit produces revenue. It creates value only if the expected contract contribution exceeds the incremental cost of access, optical equipment, upstream capacity, support, failures and capital employed. A municipal technology project may make annual sales look stronger. It creates durable value only if the company earns an acceptable project margin, avoids a working-capital squeeze and converts the installation into recurring support or connectivity. A new international route may look strategic. It is economic only when enough capacity is pre-sold, committed or otherwise protected from the price erosion that follows new supply.
NOVATEL has more substance than a small resource holder. It also operates far below the scale at which a hyperscaler can spread software, procurement, security and data-centre investment across global demand. Its survival question is consequently not whether it can imitate cloud scale. It cannot. The question is whether it owns or controls enough locally scarce access, trusted execution and route diversity to be paid for outcomes that a large remote platform cannot deliver alone.
The evidence gives a qualified answer. There is differentiated demand. The 2024 regulator data prove it. There is not yet strong evidence that NOVATEL retains enough of the resulting economic surplus.
The operating boundary is a Bulgarian carrier inside a much larger group
The legal and ownership boundary matters because the brand can make a small subsidiary appear to have the balance sheet and reach of its ultimate group. Magyar Telekom's 2025 annual report identifies NOVATEL EOOD as incorporated in Bulgaria, 100% owned and consolidated, and describes it as a wholesale telecom service provider. It lists the main activities as domestic and international leased lines, international Internet access, IP VPN services, roaming services, infocommunication solutions and international call termination. Deutsche Telekom in turn controls Magyar Telekom, but NOVATEL remains a distinct Bulgarian limited liability company with its own local assets, staff, customers and obligations.
The company's Bulgarian identity is independently visible. Its published fixed-telephony conditions give unified identification code 131301643. The RIPE NCC member page uses the same current Sofia contact details as the company site. The company says it was established in 2004. Its current chief executive, Kamen Georgiev, took the role in March 2025 after a long sales career at NOVATEL. This continuity suggests the strategic turn toward managed ICT services was developed locally rather than imposed by a newly arrived parent representative.
Control should not be confused with unlimited support. Full ownership can lower procurement costs, open group routes, supply process standards and strengthen customer confidence. It may also give NOVATEL access to international capacity that an independent Bulgarian operator would have to assemble contract by contract. But a parent relationship does not make every subsidiary investment rational. Magyar Telekom has an incentive to demand a return, consolidate platforms and allocate capital to uses with better risk-adjusted economics. Group affiliation is therefore a supplier and credibility advantage, not proof that NOVATEL can sustain low margins indefinitely.
The practical operating boundary is narrower than the promotional one. NOVATEL controls a Bulgarian network and sells regional connectivity and technical services. It can use group coverage and external partners beyond that footprint. It does not own Deutsche Telekom's global network, the public clouds to which customers connect, the data centres where it colocates, every access tail required for a multinational customer, or the Black Sea cable that has been proposed under the Kardesa project. Whenever a service leaves its own fibre, facilities or staff, part of the customer payment must be shared with an upstream carrier, landlord, equipment vendor, cloud platform or local access partner.
That boundary determines who carries the downside. NOVATEL carries the local service promise even where it depends on another provider. The customer buys one accountable service; the company purchases several inputs. Integration is valuable precisely because the customer does not want to coordinate those inputs. It is also dangerous because the integrator can end up guaranteeing performance while owning only part of the delivery chain.
Number resources prove operational standing, not economic power
The RIPE NCC's member record confirms that NOVATEL is a Local Internet Registry serving Bulgaria. That status matters. It allows the company to administer number resources and support customer assignments in the RIPE service region. It is evidence of an operating network and governance obligations. It is not evidence that the company sells a particular service, has a certain quality level or earns excess returns.
The direct cost of membership is too small to create a meaningful entry barrier. Under the RIPE NCC 2026 charging scheme, one LIR account costs EUR 1,800 a year, with separate charges for certain independent resources and AS numbers. The scarce value lies not in paying the fee but in the accumulated address holdings, routing competence, customer trust, clean operational history and ability to make those resources useful across a network. Even IPv4 scarcity does not automatically accrue to the operating business. Addresses can support service revenue, but they can also be leased, transferred or stranded in low-value uses. Their balance-sheet value and earning value are different questions.
The live routing evidence is substantial enough to reject the idea that NOVATEL is a paper LIR. On 10 July 2026, the RIPEstat announced-prefixes result for AS41313 returned 29 originated routes: 24 IPv4 and five IPv6. The companion routing-status result counted 23,040 addresses in its announced IPv4 space and 175 observed neighbours. The first observed route under the autonomous system dated to November 2006. This is a long-running routing operation, not a recently activated shell.
The larger customer cone is visible in public route observation. Hurricane Electric's record for AS41313 showed hundreds of announced customer routes and direct or indirect adjacency to many Bulgarian networks. Those customer routes should not be described as NOVATEL-owned address space. They are evidence that the network carries reachability for other networks. That distinction matters: a carrier can announce a large customer cone without owning the underlying addresses, and the revenue depends on whether those customers buy transit, transport, peering access or a bundled service on durable terms.
Resource-holder status therefore supports the business model but does not settle the economic question. It lets NOVATEL participate. It does not protect price. The return comes from the contract attached to the route, not the database entry attached to the address.
The physical network is real, but the useful measure is sold capacity
NOVATEL says its network is a wholly owned national DWDM system with 39 points of presence in major Bulgarian cities and international connections toward Turkey, Greece, North Macedonia, Serbia and Romania. It describes one of its roles as providing dark fibre and leased capacity between Western Europe, Turkey, the Caucasus and the Middle East. The claim is commercially plausible when set beside the observed routing and interconnection footprint, although the exact fibre kilometres, duct ownership, lit capacity, utilisation and replacement age are not publicly disclosed.
PeeringDB provides a second operational view. The AS41313 profile, updated in June 2026, classifies NOVATEL as a regional network service provider with self-reported traffic of 50-100 Gbps and an open peering policy. It lists public connections at BIX.BG and T-CIX in Sofia, DE-CIX in Frankfurt, NetIX and several other exchange fabrics, plus facilities in Sofia and at Equinix FR5 in Frankfurt. Published port capacities include a 100G connection at T-CIX, two 40G BIX connections and a 30G DE-CIX connection. Those entries are self-maintained, so they should be read as declared operating data rather than audited traffic measurements. They still show where the company has chosen to spend money: local exchange density, international reach and proximity to major content and transit ecosystems.
The RIPE routing declaration for AS41313 names large upstreams and peers, including Lumen, Arelion and RETN, and exchange or content relationships involving Google, Meta, Amazon, Akamai and Cloudflare. A routing declaration does not prove that every listed session is active at every moment. It does show the intended architecture: buy global reach from multiple carriers, peer directly where traffic volume justifies it, and carry a customer set behind the NOVATEL autonomous system.
That architecture has two economic advantages. First, direct peering can lower paid transit cost and improve latency for popular destinations. Second, multiple upstreams and border paths make resilience saleable to enterprises and smaller Internet providers that cannot build the same diversity efficiently. The company can charge for the engineering and accountability around the route, not only for raw Mbps.
It also has a cost that traffic statistics conceal. Each point of presence needs space, power, backhaul, monitoring, spares and intervention. Each exchange port and remote facility adds recurring fees and equipment. Each optical upgrade creates another capital decision. The company's international leased-line page advertises 10, 100 and 400 Gbps options over DWDM. Supporting those products before demand is committed can produce an impressive network and an unimpressive return.
The economically useful network measure is not PoPs, prefixes or theoretical capacity. It is contribution from sold capacity after access, upstream, colocation, support and capital costs. NOVATEL does not publish that measure.
Wholesale demand validates the franchise and reveals the erosion
The strongest public evidence comes from Bulgaria's Communications Regulation Commission, because it reports the market rather than the company story. The CRC's 2024 annual report put the whole Bulgarian electronic communications market at BGN 4.021 billion, up 5.7%. Data transfer and Internet access generated BGN 2.176 billion. Leased lines were only BGN 28.633 million, or 0.7% of the market. The segment in which NOVATEL is most visible is important to business continuity but economically small beside consumer and mobile data.
Wholesale leased lines generated BGN 23.856 million from 2,687 lines in 2024. NOVATEL held 20.6% by line count and 25.1% by revenue. Neterra led with 25.9% of lines and 33.7% of revenue; Vivacom had 15.0% and 18.3%, respectively. These figures establish a real market position. NOVATEL was not a marginal participant. Its revenue share above its line share also suggests a higher average revenue mix than the market as a whole, although line capacities and destinations differ too much for revenue per line to be treated as a clean price measure.
The year-on-year movement is more instructive. NOVATEL's count of national wholesale leased lines rose 31.3% in 2024. Its revenue from those national lines rose 28.2%. This is genuine demand growth in the recurring transport base. Yet its overall wholesale revenue share fell from 26.4% in 2023 to 25.1% in 2024, while Neterra's jumped from 17.7% to 33.7%, helped by international wholesale lines. NOVATEL grew and lost relative economic ground at the same time. That is exactly why revenue growth cannot be used as a proxy for value creation.
The four-year comparison is harsher. The CRC's 2020 annual report gave NOVATEL 30.1% of wholesale lines and 43.0% of wholesale leased-line revenue. By 2024 those shares were 20.6% and 25.1%. Applying the rounded shares to each year's reported segment total implies revenue of about BGN 6.3 million in 2020 and BGN 6.0 million in 2024. The exact amounts should not be over-read because the percentages are rounded, but the direction is clear: the market expanded without delivering comparable nominal expansion to NOVATEL in this narrow segment.
There is a positive interpretation. NOVATEL endured, remained one of the three leading providers and accelerated national lines in 2024. A pessimistic interpretation is more persuasive for returns: infrastructure that held 43% of segment revenue in 2020 held only a quarter four years later, and the absolute revenue indicated by the regulator tables was broadly flat. Inflation, wages and equipment costs did not stand still. A stable nominal revenue pool can mean a shrinking real contribution.
The company is not powerless. A 25.1% share gives it customer relationships, operational learning and procurement volume. But the regulator data do not support the claim that network ownership alone confers pricing power. They support a narrower claim: NOVATEL owns a relevant position in a competitive market where rivals can take the incremental economics.
The revenue mix says transport lasts while projects swing
NOVATEL does not publish a detailed investor presentation. Secondary compilation of its filed Bulgarian accounts nevertheless allows the shape of the business to be tested. SeeNews reports 2024 operating revenue of BGN 25.021 million, operating expenses of BGN 23.918 million, operating profit of BGN 1.077 million and net profit of BGN 625,000. Net sales were BGN 24.942 million. The resulting operating margin was about 4.3%; the net margin was about 2.5%.
Those figures align with the unit-level equity and income amounts reproduced in Deutsche Telekom's 2025 financial statements, which list BGN 16.669 million of equity and BGN 625,000 of net income for NOVATEL. They should still be treated as the latest detailed subsidiary figures rather than group-segment disclosure. The parent does not break NOVATEL out as a separately reportable business.
The 2023 comparison demonstrates volatility. Operating revenue was BGN 29.535 million and net profit BGN 3.186 million. In 2024, revenue fell about 15.3% and net profit about 80.4%. A business that looked comfortably profitable in one year returned to a thin margin in the next. Without contract-level disclosure, it is not possible to say whether 2023 contained an unusually profitable project, asset right, settlement or other item. It is possible to say that the 2023 result did not establish a stable earnings level.
The service split supports that view. SeeNews reports that data-transfer net sales increased from BGN 11.756 million in 2023 to BGN 13.100 million in 2024, an increase of roughly 11.4%. Internet revenue increased from BGN 1.366 million to BGN 1.698 million, and support edged up from BGN 371,000 to BGN 388,000. By contrast, IT revenue fell from BGN 8.122 million to BGN 6.788 million, and the broad other category fell from BGN 6.603 million to BGN 1.992 million. The recurring network core strengthened while project-like or less clearly defined revenue contracted.
That is not a failure of strategy. It is a warning about its economics. Transport revenue tends to recur but faces persistent unit-price pressure. IT integration can earn better gross profit when NOVATEL supplies scarce design skill and captures ongoing support, but it can also be hardware resale with low margins and heavy working-capital demands. Smart-city installations can create a visible order book, but public procurement is episodic and can concentrate execution risk in a few contracts. Combining all three activities makes the top line larger; it does not automatically make the company more valuable.
The strongest version of the model uses each layer to protect the others. Connectivity gets NOVATEL into the account. Security, SD-WAN, systems work and support make switching harder. The network gives the integrator operational credibility. A multiyear service relationship turns a one-off installation into recurring revenue. The weak version is a collection of low-margin circuits and projects sold under one brand, each competed separately and none with enough attachment to defend price.
The reported margin says the burden of proof remains with the strong version.
Thin profit exposes a cost base that cannot be marketed away
At BGN 25.021 million of operating revenue, every percentage point of operating margin is only about BGN 250,000. A delayed customer payment, emergency fibre repair, failed project acceptance, router refresh or adverse contract repricing can absorb several points. The 2024 result left little room for execution error.
The balance sheet offers support but not immunity. SeeNews reports BGN 34.871 million of total assets and BGN 16.669 million of equity at the end of 2024, an equity ratio close to 48%. That is not the profile of a company funded only by short-term debt. It also reports negative working capital of BGN 3.672 million. Negative working capital can be manageable when recurring customer cash arrives before supplier bills, especially inside a well-funded group. It becomes dangerous when project milestones slip, receivables age or equipment must be paid for before customer acceptance.
The cost base has six main buckets.
First is the physical network: fibre leases or owned-route maintenance, ducts, rights of way, points of presence, power, cooling, optical amplification and field response. Even owned fibre is not free fibre. It must be monitored, repaired and upgraded, and some routes depend on third-party property or infrastructure.
Second is active equipment. The company markets 10G, 40G, 100G and higher-capacity services. Optical shelves, coherent modules, routers, security appliances and customer equipment age before the fibre does. Capacity upgrades can reduce unit cost, but only after traffic fills them. Until then they raise depreciation and support complexity.
Third is the external network. International transit, exchange ports, remote facilities, local access tails and partner circuits convert network reach into a variable and semi-fixed supplier bill. Multiple upstreams improve resilience and negotiating leverage, but duplication has a price.
Fourth is labour. The 2024 accounts compiled by SeeNews show 57 employees. That is a small team for a national network, round-the-clock operations, enterprise sales, project delivery, cyber security and systems integration. A compact workforce can be efficient. It can also create key-person dependence and force the company to buy specialist help at short notice. The broad current leadership page, which names separate heads for operations, projects, ICT, the network operations centre and telecom services, indicates how many disciplines the company expects those 57 people to cover.
Fifth is vendor and project cost. NOVATEL lists partners including Microsoft, Palo Alto, Fortinet, IBM, Dell, Lenovo, Schneider and Siemens. These relationships broaden the offer but place much of product economics in the hands of global suppliers. If the customer can solicit the same hardware from several certified integrators, NOVATEL earns a premium only for architecture, delivery speed, financing, support or the attached network.
Sixth is compliance and resilience. Bulgaria requires public communications providers to notify the CRC, submit annual reports and comply with the Electronic Communications Act and general requirements. The CRC states that providers above the revenue threshold pay an annual regulatory fee of 0.291% of relevant gross communications revenue after specified deductions. Cybersecurity, privacy, incident response and service-quality obligations add staff and technology costs that do not disappear when bandwidth prices fall.
These costs explain why more traffic is not enough. NOVATEL must raise gross contribution per customer, lower avoidable supplier spend or improve asset utilisation. A strategy that merely adds products raises complexity faster than earnings.
Customers pay for continuity, but contract durability is uneven
Who pays NOVATEL? The public evidence points to four groups: Bulgarian enterprises buying dedicated Internet and private networks; smaller carriers and Internet providers buying transit, leased lines or dark fibre; multinational or regional customers needing a Bulgarian access and support partner; and public-sector organisations buying integration or smart-city work. Each group values a different part of the same operating base.
The dedicated Internet offer promises uncontended symmetric bandwidth, static IPv4 or IPv6 addresses, redundant international routes, service levels and round-the-clock support. The MPLS and IP VPN offer sells traffic segregation, quality of service and managed connectivity between offices and data centres. These customers pay to avoid the cost of building network expertise and coordinating faults across suppliers. A small enterprise benefits from one accountable party. NOVATEL benefits from a recurring monthly relationship. The downside sits with NOVATEL when a failure originates in a supplier circuit but the service commitment remains its own.
The dark-fibre offer has a different economic profile. NOVATEL offers either an indefeasible right of use or a lease. A long-lived right can bring upfront cash and committed utilisation but gives away future pricing upside on the fibres covered. A lease preserves flexibility and recurring revenue but leaves renewal and vacancy risk. For a customer with high traffic, lighting its own fibre can be cheaper than continuing to buy managed capacity. The product is therefore both a monetisation tool and a route through which the most bandwidth-intensive customer can take active-service margin away from NOVATEL.
International leased lines can be durable because changing a cross-border production circuit is operationally risky. NOVATEL advertises protected and unprotected point-to-point connections with performance monitoring and service levels. But contract duration, minimum commitments, renewal rates and termination rights are not disclosed. An older company page promoted the absence of long-term contracts as a customer benefit. That may have helped a flexible challenger win business, yet flexibility transfers demand risk to the carrier. A network with long-lived assets and short-lived customer commitments is structurally exposed.
Managed IT and security can deepen durability when the company understands the customer's architecture, operates devices and responds to incidents. The information-security offer spans vulnerability testing, audits, risk analysis, incident response and infrastructure design. Those services can carry higher value than bandwidth because the customer buys judgement and accountability. They can also remain project work if the engagement ends after assessment or installation.
Customer concentration is the largest undisclosed commercial risk. The company says it connects more than 150 customers across more than 50 countries, but it does not disclose revenue concentration, churn, contract duration or the share tied to its parent group and affiliates. The fall in broad other revenue between 2023 and 2024 shows how one or several large projects can move the total. It does not identify the customer. Any claim that NOVATEL is diversified would therefore be premature.
The test is simple: how much gross profit is secured for the next three years before a salesperson makes another call? The public record does not answer it.
Upstream diversity lowers outage risk, not dependency
NOVATEL's observed network relationships are an advantage because a regional provider that depends on one international carrier is selling a fragile service. Multiple global upstreams, local exchanges and direct content paths reduce that fragility. They also create choice in procurement. If one transit provider raises price or suffers a failure, traffic can move.
Dependency remains. Global reach still requires other networks. Direct peering reaches only the destinations available through the relevant exchange or bilateral session. Content platforms can change peering policy, traffic engineering or cache placement. A data-centre landlord can raise cross-connect and power charges. A router vendor can extend delivery times. A local access provider can fail outside NOVATEL's own footprint. The company can diversify each dependency; it cannot eliminate the category.
The economics are especially unforgiving because upstream prices tend to fall. BEREC's 2024 report on IP interconnection found that prices and costs for transit, peering and related services continued to trend downward because of technology and competition. TeleGeography reported that weighted median 100 GigE transit prices across selected global cities fell at a 12% annual rate between the second quarters of 2022 and 2025. Lower input cost helps NOVATEL, but customers and rivals see the same benchmark. Procurement savings are competed away unless the company adds something scarce.
That scarce element can be local access, route diversity, fast repair or a bundled service level. It is rarely generic international transit. PeeringDB's open policy also makes sense in this context: exchanging traffic directly can reduce cost and improve performance. Yet an open policy means the act of peering is not itself exclusive. NOVATEL must be paid for reaching the customer, operating the route and solving the failure.
Its supplier strategy should therefore be judged by contribution and resilience together. The cheapest upstream is not necessarily the best if it shares the same physical route as another. The most diverse upstream is not necessarily economic if customers will not pay for the protection. NOVATEL needs route-level cost and failure correlation, not a long list of carrier names.
Cloud demand is an opportunity that also commoditises the carrier
Cloud competition presses on NOVATEL from both sides. It creates traffic because applications, backups, security services and collaboration tools move off customer premises. It also moves the higher-margin computing layer to global providers. A Bulgarian enterprise may need better connectivity after adopting cloud services, but most of its new technology budget can accrue to the software and infrastructure platform rather than the access carrier.
The company acknowledges the substitution. Its SD-WAN offer combines dedicated Internet, broadband and LTE or 5G underlays, chooses paths by application and provides central monitoring. This is commercially sensible. It lets NOVATEL sell control and support even when some traffic travels over cheaper Internet access rather than a private MPLS circuit. It also cannibalises the premium that traditional private networking once earned. If NOVATEL does not offer the substitute, another integrator will. If it does, it must replace lost circuit margin with software, security and service margin.
The realistic alternatives for a customer are broad. A multi-site enterprise can buy MPLS from NOVATEL, Neterra, Vivacom or A1. It can buy commodity Internet from more than one provider and overlay SD-WAN. It can move applications to a public cloud and rely on encrypted Internet access. It can contract a systems integrator that sources connectivity from several carriers. A large network can lease or acquire dark fibre and operate its own equipment. The customer's switching cost depends less on the carrier name than on how deeply the service is integrated into sites, security policy, addressing, monitoring and incident response.
Bulgaria offers both runway and a demand ceiling. Eurostat reported that only 17.8% of Bulgarian enterprises used paid cloud services in 2025, compared with 52.7% across the EU. Adoption barely moved from 17.5% in 2023. The optimistic reading is that a large future migration can drive connectivity, security and integration demand. The colder reading is that NOVATEL serves a market where many firms are small, price-sensitive and slow to buy advanced services. The opportunity may arrive gradually while the carrier continues to fund current network capability.
Cloud scale also changes customer expectations. Global platforms make computing capacity appear elastic and measurable. A regional carrier cannot match their unit economics, automation budget or product breadth. It can beat them where the physical and organisational problem is local: connecting a factory, repairing a fibre, navigating a municipal site, designing route diversity, integrating old equipment or taking a phone call in Bulgarian when the network is down.
That is the defensible boundary. Strategy should concentrate resources there. Calling every adjacent product an ICT opportunity would be marketing, not allocation.
Bulgaria is competitive enough to keep pricing pressure permanent
NOVATEL competes in a market with a strong incumbent, national mobile groups, specialist carriers and many local providers. The CRC counted 825 active electronic communications undertakings in 2024. Only 18 supplied wholesale leased lines, but 621 supplied retail Internet access or data transfer. The addressable market is broad; the supplier field is crowded.
The 2024 wholesale leased-line table identifies the closest economic alternatives. Neterra has greater current share and a broad international infrastructure proposition. Vivacom can combine fixed access, mobile, enterprise service and national reach. A1 can do the same in business connectivity. Smaller carriers can be aggressive on selected routes. Global providers can sell transit at major hubs. NOVATEL's group relationship is helpful, but it does not remove these alternatives.
Low retail prices reinforce buyer expectations. The European Commission's 2024 broadband price study found that fixed broadband prices across the EU fell 5% from 2023 and that Bulgaria and Romania had some of the lowest purchasing-power-adjusted prices for double- and triple-play services. NOVATEL is focused on business and wholesale products rather than consumer bundles, so the baskets are not directly comparable. The commercial culture still matters. Customers accustomed to cheap high-speed access will challenge a premium unless resilience, service levels or integration are concrete.
The company's strongest local alternative is not always another carrier. It is customer acceptance of lower resilience. Many SMEs can operate on business broadband plus mobile backup rather than dual dedicated routes. Many applications can tolerate an Internet VPN. A procurement team can accept slower repair in exchange for a lower monthly price. NOVATEL earns a premium only when downtime, latency, compliance or operational complexity has a cost the buyer recognises.
This is why the company should not pursue undifferentiated volume. The 2024 increase in national lines is encouraging only if the contracts contribute enough after service cost. Winning twice as many low-price circuits can consume ports, field time and support capacity without raising economic profit. The right metric is return on the incremental route and customer relationship, not market share in isolation.
Kardesa could create scarcity, but a memorandum is not contracted return
The Kardesa Black Sea cable is the most credible path to a different economic position because routes can be scarcer than bandwidth. In January 2026, Bulgartransgaz announced a memorandum among the NOVATEL-Bulgartransgaz association, Apollo Submarine Cable System Limited and Vodafone Ukraine covering work on the Aheloy-Sofia segment. The official announcement said Vodafone and NEQSOL Holding were investors in a project exceeding EUR 100 million, that Bulgartransgaz would contribute existing assets rather than new financial resources, and that the parties would seek completion by mid-2027.
The Bulgarian innovation ministry described the proposed system as roughly 1,700 kilometres of submarine cable with an Aheloy landing station and an Aheloy-Sofia terrestrial connection using almost 1,200 kilometres of new fibre infrastructure. NOVATEL says it will build and operate the Bulgarian terrestrial infrastructure. The project aims to connect Bulgaria and Georgia, with a Turkish branch and a planned Ukrainian connection, as part of a wider route toward Central Asia.
The economic attraction is clear. A landing station and terrestrial route can create a control point. NOVATEL could earn from construction, operation, capacity, backhaul, colocation, protection paths and enterprise connections that attach to the new corridor. The route could diversify traffic away from existing paths and make the company's Bulgarian fibre more valuable. External investors and Bulgartransgaz's existing ducts could limit the capital NOVATEL must supply relative to building the whole system alone.
The missing facts are more important than the announced scale. A memorandum sets a framework; it is not a final construction contract, capacity purchase or revenue guarantee. The public disclosures do not state NOVATEL's required equity, construction exposure, minimum revenue, ownership of the landing station, share of capacity, operating fee, take-or-pay commitments or liability for delay. They do not show how much of the 1,200 kilometres is new cable in existing ducts rather than a wholly new physical route. They do not disclose presales.
Kardesa can also create its own price pressure. New submarine supply lowers transport cost. That benefits the route owner only if demand fills capacity and the owner retains scarcity in landing, terrestrial backhaul or protected alternatives. Otherwise the system can reduce prices for every carrier, including NOVATEL. TeleGeography's observation that new submarine systems accelerate price declines in previously expensive markets is relevant: new capacity produces social and customer value more reliably than owner returns.
Geopolitics raises both demand and cost. A route across the Black Sea and toward Ukraine can command value for diversity, but it faces permitting, maritime, security and repair risk. The EU's 2025 cable-security work emphasises prevention, surveillance, repair capacity and deterrence because subsea systems are exposed to accidental and hostile damage. Those are not abstract policy costs for a landing operator. They become monitoring, insurance, redundancy, spares and response obligations.
Kardesa should therefore be treated as an option on differentiated demand, not as current proof of it. The project improves the thesis only when the commercial commitments and NOVATEL's capital burden are known.
Regulation can raise trust while consuming the margin
NOVATEL operates at the intersection of telecom regulation, cybersecurity, privacy, public procurement and critical infrastructure. This can favour an established operator. Compliance history, certifications, group standards and a staffed operations centre make it harder for a lightly organised reseller to compete for sensitive work. They can also turn a thin-margin provider into the party that bears legal and operational responsibility for several suppliers.
The EU's NIS2 Directive is especially relevant. Article 3 treats medium-sized providers of public electronic communications networks or publicly available communications services as essential entities. With 57 employees and 2024 revenue above EUR 10 million, NOVATEL appears to fall within the medium-sized range, although final scope depends on the legal classification and Bulgarian implementation. Essential-entity treatment raises expectations around risk controls, incident handling, supply-chain security and leadership oversight.
Implementation timing adds uncertainty rather than relief. The European Commission's Bulgaria status page records that it sent a reasoned opinion in May 2025 over failure to notify full transposition. That page does not establish NOVATEL's current individual status. It does show that Bulgarian providers have had to prepare against an EU obligation while national details were still moving. Delayed clarity can increase advisory and implementation cost because firms must plan before every local rule is settled.
The company also sells security and public-sector systems. That creates a useful revenue attachment and a conflict of economics: better controls require skilled people, redundant systems and continuous testing, while customers often compare bids on price. A low-cost award can leave the integrator with years of support exposure. Public-sector payment schedules can also worsen working capital.
Bulgaria's adoption of the euro on 1 January 2026 changes NOVATEL's functional currency to EUR according to Magyar Telekom's annual report. The lev had long been fixed to the euro, so this is not a sudden removal of open currency risk. It should reduce conversion and accounting friction for a company buying international capacity and equipment in euros. It does nothing about dollar-priced hardware, interest cost, supplier concentration or the economic risk of selling long-term capacity too cheaply.
Regulation is therefore a modest barrier and a permanent cost. It can make NOVATEL more trusted. It cannot rescue a weak contract.
Market signals support the direction, not the return
The non-financial signals point to an active shift from carrier identity toward infrastructure and integration. The refreshed company site puts Kardesa, system integration and smart-city work beside telecom services. Its current home page says more than 150 customers are connected across more than 50 countries. Its LinkedIn activity highlights the Black Sea route and municipal street-lighting projects. A German Embassy Sofia post reproduced there credited NOVATEL's collaboration with municipalities in installing more than 35,000 efficient lighting fixtures, including more than 4,000 in Pernik.
These are signals, not audited order-book evidence. They indicate that NOVATEL can win public projects and has institutional visibility. They do not disclose contract value, gross margin, payment timing, maintenance responsibility or repeat revenue. Street lights demonstrate execution outside classic telecom. They do not prove that smart-city services earn better returns than transport.
The company's own network counts also illustrate why promotional numbers need care. The dedicated network page says 39 points of presence, while the current about page refers both to more than 39 and to more than 40. An older LinkedIn description says 35. The likely explanation is network expansion and uneven page updates, not deception. The economic lesson is that counting sites is less useful than publishing utilisation, protected-route availability and customer contribution.
Public routing evidence is a better signal because it is harder to create without operating activity. The live autonomous system, IPv6 routes, exchange ports and customer reach corroborate the underlying carrier. The market-share and filed-account figures then constrain the conclusion: the network is real, but real networks can still earn thin returns.
The downside rests with the party that adds capacity before commitment
Who benefits from NOVATEL's strategy is easier to answer than who earns from it. Enterprise customers benefit from another national carrier, route diversity and one local party for connectivity and systems. Smaller providers benefit from upstream choice and access to international reach without building it themselves. Municipalities benefit when a technology contractor can combine communications, devices and ongoing support. Magyar Telekom benefits from a Bulgarian wholesale presence and an option on a new east-west corridor. Cloud platforms benefit when reliable local connectivity allows more Bulgarian workloads to move online.
Who pays is also clear. Customers pay recurring circuit and support charges or project fees. External investors are expected to fund the large submarine element of Kardesa. Bulgartransgaz contributes an existing physical route. NOVATEL pays for its local staff, network operation, sales effort, supplier commitments and whatever construction or equipment obligations its final agreements assign.
Who carries the downside depends on contract sequence. If customers commit before NOVATEL spends, risk is shared. A dark-fibre right, take-or-pay capacity agreement, multiyear managed service or funded public contract can anchor investment. If NOVATEL builds first in expectation of demand, the company and ultimately its shareholder carry utilisation risk. If it promises an end-to-end service using third-party inputs, it carries service risk while suppliers retain their contracted fees. If it accepts a fixed-price integration project while hardware and labour costs move, it carries delivery risk.
The 2024 margin means NOVATEL cannot absorb much unpriced downside. Its parent can provide patience, but patience is not value creation. The rational allocation rule is severe: capacity should follow committed demand; new services should attach to existing customers or assets; public projects should be priced for working capital and support; and Kardesa exposure should be limited by external funding, presales or contractual operating income.
Anything looser is a bet that scale will repair economics after the spending. Below cloud scale, that bet usually belongs to the customer and supplier, not the regional carrier.
Seven facts would change the judgement
The present evidence supports a defensible franchise with inadequate proof of durable returns. Seven disclosures could move that conclusion.
First, 2025 and first-half 2026 accounts showing that data-transfer growth lifted operating cash flow and restored operating margin would demonstrate that 2024 was a transition year rather than the normal earning level. Revenue alone would not be enough.
Second, contract data showing low churn, multiyear minimum commitments and no excessive top-customer exposure would turn the physical network into a more credible annuity. The crucial measure would be gross profit under commitment, not signed contract value.
Third, a bridge from BGN 625,000 of 2024 net profit to recurring cash earnings would clarify whether depreciation, finance cost, project timing or one-off items obscure stronger underlying economics. Maintenance capital spending must be included. Excluding the cost of keeping the network competitive would flatter the answer.
Fourth, route-level utilisation and return data would show whether 39 PoPs and the international footprint are productive assets or partly idle overhead. Capacity sold under long-term rights should be separated from capacity available for future sale.
Fifth, final Kardesa contracts disclosing NOVATEL's capital obligation, ownership, operating fee, pre-sold capacity and delay liability could transform the project from strategic narrative into an investable corridor. A funded operating role with take-or-pay demand would materially improve the thesis. An open-ended build obligation would weaken it.
Sixth, evidence that security, SD-WAN and integration increase customer contribution after labour and vendor cost would validate diversification. Growth in hardware-heavy revenue without recurring support would not.
Seventh, stabilisation or recovery in the CRC wholesale revenue share would show that NOVATEL can retain value while adding lines. Continued volume growth alongside falling share and flat real revenue would confirm price-taker economics.
NOVATEL is relevant, but relevance is not yet a moat
NOVATEL has enough differentiated demand to remain a serious Bulgarian infrastructure provider. The regulator data, live routing footprint, national optical network and group ownership make that conclusion firm. It has something worth defending: local routes, wholesale relationships, operational accountability and a credible role in a proposed Black Sea corridor.
It does not yet have enough public evidence of differentiated demand to claim robust value creation. A 2.5% net margin, volatile project revenue and a fall in wholesale leased-line revenue share from 43.0% to 25.1% outweigh the promotional force of more PoPs, more products or a larger future route. The company is not a pure price-taker in local access and service. It is close to one in generic capacity, hardware and any integration work that can be rebid among certified suppliers.
The right strategy is selective, not expansive. NOVATEL should concentrate capital on routes and services where it controls the customer outcome, secure commitments before adding capacity, use SD-WAN and security to defend relationships rather than merely widen the catalogue, and accept Kardesa risk only where external funding and contracted demand protect the return. Its resource holdings and parent affiliation make this strategy possible. They do not make it inevitable.
The conclusion is not diplomatic: NOVATEL has built an operating franchise, but the latest economics do not show a moat. Until recurring cash return rises and wholesale value share stops eroding, the burden of its infrastructure remains heavier than the pricing power it has proved.