Dragonfly Holding B.V. and the margin risk below cloud scale

Summary

• Dragonfly Holding B.V. is publicly visible as a Dutch RIPE NCC member and legal holder of Internet number resources, but public company records classify it as a financial holding, so the economic read must distinguish the legal resource holder from Qfast's commercial telecom and ICT offer.
• Qfast's public proposition is an SME-oriented bundle of business internet, fiber, telephony, VPN, 4G backup, network support and managed services. That is a continuity and accountability proposition, not a cloud-scale or mass-access proposition.
• The network-resource evidence is mixed. RIPE records tie Dragonfly Holding B.V. to AS31172 and address resources, yet current routing evidence shows AS31172 with no visible announced prefixes and Dragonfly-linked IPv4 and IPv6 resources originated through Nextpertise AS41960.
• Nextpertise gives a credible route to operational reach through wholesale broadband access, private interconnect and the ability to advertise customer RIPE addresses, but that also makes supplier concentration and wholesale economics central to Dragonfly's margin risk.
• Public evidence does not disclose revenue, gross margin, churn, contract length, customer count, top-customer concentration, capital expenditure or cash generation, so the investment-grade conclusion cannot rest on unverified scale or retention claims.
• The present judgment is conditional and cautious: Dragonfly/Qfast may have a defensible niche when customers pay for integrated continuity, routed IP space, telephony and hands-on support, but the public fact pattern still leaves it at risk of being an infrastructure price-taker below cloud scale.

Management's incentive is relevance without cloud scale

The economic incentive for Dragonfly Holding B.V. is easy to frame and hard to prove. A small network-resource holder in the Netherlands has to remain relevant in a market where customers can buy connectivity from national operators, cable companies, mobile networks, wholesale resellers, cloud platforms, security vendors and managed-service firms. Relevance cannot come from scale alone. Dragonfly does not present itself in public records as a national access network owner, a hyperscale cloud provider, or a carrier with a global autonomous-system footprint.

The plausible incentive is narrower: hold enough technical control, commercial knowledge and local service capacity to make business customers prefer Qfast when generic connectivity alone is not enough.

That position can work. Small and mid-sized enterprises often care about a person answering the phone, a router being configured correctly, a backup link failing over cleanly, a voice platform surviving a local access outage, a building owner being able to split fiber service among tenants, or a branch network being stitched to a datacenter and cloud services without each supplier blaming another. Those needs are not glamorous, but they can carry margin if the provider becomes the accountable integrator. A regional provider does not need to beat KPN, VodafoneZiggo, Odido, Eurofiber or every wholesale platform on raw network ownership.

It needs to own the customer problem.

The danger is that this language can disguise weak economics. "Managed", "business", "custom" and "independent" are common terms in telecom marketing. They do not by themselves prove pricing power. If the provider buys access from larger networks, buys equipment from vendors, relies on a wholesale backbone, and serves customers who can switch to another reseller, the business may be doing real technical work while capturing only thin resale margin. Resource-holder status can support a better proposition, especially where routed IP addresses, failover and network continuity matter.

But resource-holder status is not automatically economic power. It is an input.

The public evidence around Dragonfly/Qfast therefore has to be read as an incentive map rather than a complete financial record. Public RIPE records show a durable resource-holder identity. Qfast's own site shows a commercial service menu aimed at business continuity. Routing records show that the resources are currently operationally tied to another autonomous system, Nextpertise AS41960, rather than being visibly originated by AS31172. Dutch market evidence shows a competitive broadband market dominated by large operators and increasingly fiber-heavy access.

The central issue is whether Dragonfly's customer relationships convert those technical inputs into durable demand.

The answer is not available from a single registration. It depends on five linked tests. First, does Qfast have differentiated demand from customers whose needs are not solved by a basic broadband line? Second, are contracts sticky because Qfast is embedded in voice, VPN, IP addressing, support and continuity? Third, does the company need meaningful cash capital to maintain its role, or can it leverage wholesale infrastructure with limited fixed assets? Fourth, is supplier concentration manageable, especially where current public routing points toward Nextpertise? Fifth, are substitutes realistic and cheap enough to cap margins?

The public record gives partial answers. The missing disclosures matter.

Dragonfly is the resource holder; Qfast is the commercial surface

The first discipline is entity separation. The RIPE NCC member listing identifies Dragonfly Holding B.V. in the Netherlands, with an address in Huizen, a contact email using the qfast.nl domain, a Dutch service area and a phone number that matches the Qfast contact surface. RIPE database records for ORG-DHB3-RIPE identify Dragonfly Holding B.V. as an LIR, with Dutch registration number 32095529, an address at Betuining 33 in Huizen, and creation in February 2012. Those records support the proposition that Dragonfly is a legal holder and registry-facing resource entity.

Company-information records add a different lens. Company.info lists Dragonfly Holding B.V. at the same Huizen address and classifies its activity under financial holding activities. Creditsafe also lists Dragonfly Holding B.V. with the same Dutch registration number and a 2003 incorporation date. That does not negate the RIPE record, but it means the public identity is not a simple retail ISP biography. Dragonfly appears as the holding and resource-holder layer, while Qfast is the brand through which services are marketed to customers.

Qfast's website is the public commercial surface. It describes Qfast as an independent provider of telecom and ICT services for the business market, with more than 20 years of expertise and a compact organization. Its menu includes business internet, business fiber, internet for multi-tenant buildings, 4G backup, business telephony, cloud telephony, SIP trunking, number services, mobile telephony, VPN, international connections, managed services, cloud backup, Microsoft 365 support, network security and 24/7 network support. The language is practical and SME-oriented.

It is built around continuity, advice, customization and support rather than consumer volume.

This distinction matters for valuation. A holding company that owns or controls Internet number resources can support a service company, but its value is not the same as the list value of an address block or the existence of an autonomous-system number. A customer does not pay Dragonfly because ORG-DHB3-RIPE exists. A customer pays because Qfast can solve a communications problem reliably, because Qfast can supply multiple access technologies, because Qfast can manage a voice environment, because Qfast can give the customer a routed subnet, or because Qfast can provide a failover design that avoids operational downtime. The resource record is evidence of capability, not proof of demand.

The Qfast registration page is useful here. Qfast says it is registered with the Dutch Authority for Consumers and Markets under two registration numbers, and it names bodies such as COIN, KPN Wholesale, Telecom Society, RIPE NCC and FIST. That set of references suggests the business sits in the regulated telecom-service layer rather than merely selling office IT on an agency basis. Still, registration is a floor, not a moat. Dutch telecom providers that offer public electronic communications networks or services are required to notify ACM.

Registration says the provider is in the regulated market; it does not say that the provider has unusual scale, high margin or low churn.

The strongest public thesis is therefore not that Dragonfly is a large network company. The stronger thesis is that Dragonfly/Qfast has the ingredients for a regional managed-connectivity role: registry identity, business-service registrations, a service portfolio, network-resource holdings and access to multiple infrastructure suppliers. The weakest public thesis would be to treat those ingredients as if they already proved durable economics. They do not. The gap between technical capability and monetized demand is the whole investment question.

The service menu sells business continuity, not mass-market access

Qfast's own pages make the customer proposition clear. The company sells business internet across DSL, fiber, cable, radio and 4G backup; it describes fiber options up to 100 Gbps; it emphasizes symmetry, reliability and multiple IP addresses; it offers site-to-site and datacenter connectivity; and it combines connectivity with telephony, VPN and support. This is not a pure retail broadband pitch. It is a continuity pitch for companies that depend on communications but do not want to engineer the stack themselves.

The 4G backup page is the clearest example. Qfast explains that primary fixed infrastructure can be affected by excavation damage and that a mobile backup link enters the building through a different physical path. It says the router automatically switches over when the primary line fails and that the same WAN IP address remains available during the outage. This is economically important. A backup line is not just a second commodity SIM if the provider can preserve addressing, routing and service behavior in a way that keeps email, voice, payment systems or remote access working.

The margin opportunity is in the design and accountability around failover.

The multi-tenant building fiber page also shows a specific problem. Qfast markets a building-level fiber solution where capacity can be distributed to multiple tenants, including the use of each tenant's own IP range. That proposition requires more than a generic access line. It involves shared building infrastructure, tenant service separation, addressing, support, billing and potentially the landlord's desire to provide a better telecom environment. If Qfast can own that coordination role, it may have more negotiating room than a reseller quoting a single broadband tariff.

The VPN and business-network pages make the same point in another way. Qfast says its own backbone allows it to design, build and manage networks in-house, with less dependence on suppliers. It offers private company networks over DSL, fiber and radio connections, branch connectivity, cloud and SaaS access, server placement in connected datacenters, security and migrations. The claim is not that Qfast owns every last-mile path. The claim is that it can combine access, addressing, routing, support and datacenter/cloud reach into a managed network.

Telephony adds another layer of stickiness. Qfast offers business telephony, hosted telephony, SIP trunking, number services and fixed-mobile integration. Its support materials describe number portability, scalable hosted accounts, device choices and a framework agreement that allows customers to scale up or down. The FAQ says hosted telephony requires a separate xDSL link to the hosted exchange for guaranteed quality and availability, and it advertises the potential for savings against a KPN business product.

The public tariff page gives 2025 call rates for Dutch fixed and mobile destinations and a setup fee per call, while other telephony economics remain quote-based or bundled.

This service menu is coherent. It targets businesses that might not have an internal network team, but for whom downtime is expensive. It also targets businesses that value one provider across internet, voice, addressing, support and continuity. Dental practices, offices, smaller professional firms, multi-site businesses and multi-tenant buildings fit that pattern. Qfast's reference page and first-party testimonial language point in that direction, although they should be treated as marketing evidence rather than independent proof of retention or revenue.

The counterpoint is just as important. Most items on the service menu have substitutes. Business fiber can be bought from national providers, cable operators, fiber networks, wholesalers and other resellers. Hosted telephony can be bought from cloud voice providers and managed-service firms. 4G backup is available through many routers and mobile operators. VPN and managed firewall services are crowded markets. The service menu only creates economic power if Qfast's integration reduces risk or effort enough that customers accept a premium and renew. Without that customer proof, the menu proves addressable demand, not captured margin.

The resource record gives control, but current routing shows dependency

Dragonfly's strongest hard evidence sits in the Internet number-resource record. RIPE records associate Dragonfly Holding B.V. with an IPv4 allocation, IPv6 resources and AS31172. The IPv4 allocation 87.239.96.0/21 dates back to the 2005 allocation record. The IPv6 records include 2a00:c2c0::/29 as an allocated IPv6 block and 2001:67c:1f0::/48 as an independent assignment. AS31172 carries the name QFAST-AS in RIPE records, with historical creation in March 2004. The aut-num remarks describe Qfast as providing broadband, voice, data, internet, transit and monitoring services for ISPs, ASPs and other companies in the Benelux.

Those records matter because a small provider with its own registry identity can do things a pure agency reseller cannot do as easily. It can assign or route customer address space, maintain a more durable addressing relationship, support business continuity designs, and negotiate with upstream or wholesale partners from a more technical position. It can also build customer trust among clients that understand why IP address continuity and routing control matter. In a market where most small businesses do not care who owns the underlying fiber, the customers who do care can be valuable.

But the live routing picture weakens any simple claim of standalone network control. RIPEstat data for AS31172 currently shows zero announced prefixes and zero visible AS neighbours. BGP.tools likewise shows AS31172 without current announced prefixes in the public routing table. By contrast, RIPEstat and RIPE route objects show Dragonfly-linked IPv4 and IPv6 resources being originated under Nextpertise AS41960. The route object for 87.239.96.0/21 lists origin AS41960, and RIPEstat routing status shows that prefix visible globally with AS41960 as origin. The more specific IPv6 prefix 2a00:c2c1::/32 is also visible with AS41960 as origin, while the broader Dragonfly IPv6 block is not itself seen as a direct origin.

This is not necessarily bad. A small provider can make a perfectly rational choice to have a wholesale or network partner originate its resources. It can reduce operational burden, improve reach, simplify upstream management and give customers stable service without the provider maintaining a full independent routing footprint. Nextpertise's own wholesale pages explicitly describe private interconnect arrangements where Nextpertise can advertise customer RIPE IP addresses using AS41960. That is a plausible mechanism for what the routing data shows.

The economic interpretation, however, changes. If Dragonfly/Qfast's public resources are currently carried through Nextpertise, then resource-holder status is not the same as independent backbone power. It is a control option mediated through a supplier. The company may still benefit from holding the resources, because customers can use stable Qfast/Dragonfly address space while Nextpertise supplies reach. But the margin chain includes a critical wholesale dependency. The value is shared with, and partly constrained by, the upstream network operator.

The RPKI evidence supports operational legitimacy rather than margin strength. RIPEstat validation for 87.239.96.0/21 with AS41960 is valid, and validation for 2a00:c2c1::/32 with AS41960 is also valid. That means the public authorization state is consistent with AS41960 originating those prefixes. It does not tell us how much Dragonfly/Qfast pays, how contracts are structured, whether customers are billed directly by Qfast, or whether Dragonfly has fallback upstream arrangements. In investment terms, the routing evidence confirms a real technical pattern and creates a supplier-concentration question at the same time.

The public aut-num record also contains history that should not be overread. AS31172's RIPE entity includes historical references to multiple peers and upstreams, and current import/export lines referencing AS41960. Old remarks can preserve former topology, former peering intentions or administrative history. The live routing table is the better evidence for current operational exposure. At present, the live picture is not an independent AS31172 footprint. It is a Dragonfly/Qfast resource position whose current public reach appears to be delivered through AS41960.

Nextpertise turns resource ownership into a wholesale choice

Nextpertise is central because it supplies the practical bridge between resource-holder status and service reach. Its wholesale broadband access materials describe a nationwide network, direct interconnection with major Dutch broadband networks such as KPN, Tele2, Ziggo and Eurofiber, and a model that lets partners offer broadband connections to business and consumer customers. It advertises speeds from 1 Mbit/s to 100 Gbit/s, and it gives partners choices over how much of the service stack they want Nextpertise to run. Partners can use the full infrastructure or provide parts such as DNS, SMTP or RADIUS themselves.

That offering fits the economics of a regional provider below cloud scale. A company like Qfast can assemble a business-facing proposition without owning every physical network. It can use wholesale access for last-mile reach, private interconnect for routing, and its own customer support and service design for differentiation. The customer sees a Qfast-managed service. The underlying network may include multiple infrastructure owners and a wholesale aggregator. This model lowers the capital requirement compared with building fiber everywhere, and it lets a smaller provider serve locations it could not economically build itself.

The same model also compresses margins. Wholesale access is not free infrastructure. The larger network operator, fiber owner, mobile provider, datacenter and equipment vendor all take their share before Qfast captures its service margin. If the end customer mainly wants a broadband line and a low price, Qfast has limited room to charge more than the wholesale cost plus support. If the customer wants a bespoke failover design, routed subnets, voice quality assurance, datacenter connectivity and support, Qfast can charge for integration. The question is how much of Qfast's base falls in the second category.

Nextpertise's own private-interconnect language is particularly relevant. It says a partner can connect privately, have Nextpertise advertise the partner's RIPE IP addresses using BGP AS41960, use redundant private peering, use Nextpertise IP space if needed, and combine these services with L2TP, RADIUS, SD-WAN and CPE templates. This is almost a blueprint for a resource-holder reseller or managed-service provider. It preserves customer-facing flexibility while moving much of the network-scale burden to Nextpertise.

For Dragonfly/Qfast, that can be a good operating design. It may reduce cash capital needs and let management focus on customer needs, support and integration. It may also let the company preserve continuity for customers even if a particular access product changes, because the addressing and service wrapper can move through a controlled wholesale network. In a niche where customers value continuity, that is a real advantage.

But it is not a moat unless the supplier terms, customer relationships and technical execution are favorable. If Nextpertise changes pricing, if access suppliers reprice, if another reseller offers the same underlay at lower margin, or if customers move to bundled offers from national operators, Qfast's differentiation narrows. Public PeeringDB data also shows no AS31172 profile, while Nextpertise has a public AS41960 profile with a selective policy and network-service-provider type. That contrast reinforces the idea that the live network gravity sits with the wholesale partner rather than with Qfast's own autonomous system.

This is the pivotal economic trade. Dragonfly/Qfast can use wholesale reach to avoid the cash burden of network ownership, but wholesale reach also means the company must earn its margin in service quality, account control and integration. The resource-holder record helps, because the company can bring address and routing continuity to the wholesale model. It does not eliminate price pressure. It changes the basis of competition from "who owns the road" to "who owns the business outcome."

Pricing power is hidden because public tariffs are thin

The most important missing evidence is pricing power. Qfast publishes some telephony rates, including 2025 Dutch fixed-line and mobile call rates, a per-call setup fee, and hosted-account pricing in support materials. It also describes potential savings against a named KPN business product and points to monthly invoicing and customer cost insight. But there is no public price book for the core business-internet, fiber, VPN, 4G backup, routed-subnet, datacenter-connectivity or managed-service bundle.

This is normal in business telecom. Custom fiber, multi-site VPN, business continuity and voice integration are often quoted rather than listed. Quote-based pricing can support margin if the provider understands the customer problem and can bundle services. It can also hide weak margins if most deals are competitively bid against national operators and other resellers. Public absence of prices is therefore not positive or negative by itself. It simply prevents an outside reader from seeing whether Qfast is paid as a premium integrator or priced as a thin intermediary.

The tariff fragments do reveal something. Hosted telephony priced per account, usage-based call rates, per-second billing, setup charges and device choices suggest an SME service model where recurring revenue is built from seats, lines, devices, support and usage. That model can be attractive if customers stay for years and add services. It becomes less attractive if customers churn, if voice prices continue to commoditize, or if cloud voice competitors capture the higher-margin software layer while Qfast supports the access line.

Business internet pricing is the bigger unknown. Qfast advertises multiple access types and speeds, including DSL, FTTH, business fiber, coax, radio and 4G backup. It says business fiber can be used for internet, telephony, payment terminals, backups, remote applications and site links, and that multiple IP addresses or a subnet can be provided. The margin on such services depends heavily on underlay cost, installation cost, support intensity, contract term and the customer's willingness to pay for resilience. None of those figures is public.

The multi-tenant building product could be economically better if Qfast can aggregate tenants behind a shared building solution and become the default provider. In that case, customer acquisition costs may be spread across multiple tenants, and the landlord relationship may create stickiness. But the public page does not disclose building count, tenant count, average revenue, gross margin or churn. The concept is credible. The scale is unproven.

The same is true of 4G backup. Maintaining the same WAN IP during failover is valuable. Customers that depend on remote access, voice, payment terminals, security systems or hosted applications may pay for it. But the market is crowded with routers, SD-WAN appliances, mobile failover offerings and managed-service providers. Qfast has to turn technical correctness into contract value. Public materials do not show how often customers buy the backup service, what attach rate it has to fiber or voice contracts, or whether it materially increases retention.

This is why the article's conclusion must resist both extremes. It would be too harsh to call Qfast a commodity reseller when its materials show real integration and resource-holder backing. It would also be too generous to infer premium margins from that integration. The public record leaves pricing power hidden. The only defensible conclusion is that the business has plausible premium niches, but the proof would require contract and margin disclosures that are not publicly available.

The cost base is a stack of suppliers, engineers and obligations

A below-scale telecom provider has a different cost problem from a national network owner. It may avoid the enormous capital spending required to build ubiquitous fiber or mobile infrastructure, but it cannot avoid the costs of access, wholesale transport, routing, equipment, support, compliance and engineering. Dragonfly/Qfast's public model appears to lean on supplier breadth rather than physical-network ownership. Qfast says it can work with infrastructure from KPN, Ziggo, Tele2, Eurofiber and other access types, while Nextpertise describes a wholesale platform connected to major Dutch broadband networks.

That supplier stack can make Qfast commercially nimble. The company can choose the available access technology for a customer site, combine fixed and mobile paths, add voice and VPN services, and manage the outcome. It can avoid deploying capital into every street. It can serve customers across the Netherlands through a mix of wholesale and partner infrastructure. For a compact organization, this is the rational way to build reach.

The cost risk is that each layer is controlled by someone else. Last-mile access has a wholesale price. Mobile backup has mobile-network pricing and equipment costs. Business fiber may require installation charges, term commitments and construction delays. Voice platforms and devices have vendor costs. Datacenter services and interconnect have their own fees. Network support requires skilled staff or contractors who can respond when customers experience downtime. Compliance with telecom rules requires process and attention, even if the business is small.

The RIPE NCC cost floor is modest but real. The 2026 RIPE charging scheme lists an annual LIR contribution of EUR 1,800, fees for independent Internet number resources and ASNs, and a sign-up charge for new members. Those fees are not large enough to define the business model, but they show that resource-holder status has recurring administrative cost. The larger costs will be supplier, staffing, support and customer-acquisition costs, none of which are public.

Support is especially important. Qfast offers tiered network support, with a Bronze level during office hours, a Silver level into weekday evenings, and a Gold level with 24/7 telephone outage service. That kind of offer can justify price premium if it prevents downtime or resolves incidents quickly. It also creates labor obligations. If too many customers buy low-margin connectivity but expect high-touch support, the support desk can absorb the margin. If customers buy higher-value bundles and support reduces churn, it becomes an asset.

The physical-diversity story adds another cost dimension. Qfast correctly notes that a backup service is more resilient when the backup path enters through a different infrastructure than the fixed line. That design is useful, but it is not free. It requires router configuration, mobile service, testing, support, customer education and often monitoring. The provider must make sure that failover works under real conditions, not just in a product description. The more accountable Qfast becomes, the more it must invest in operational discipline.

Supplier concentration is the clearest public risk because routing evidence points toward Nextpertise AS41960. If Dragonfly/Qfast uses Nextpertise to originate its resources and support wholesale reach, Nextpertise becomes more than an access option. It becomes part of the core delivery chain. That does not mean the relationship is weak or unfavorable. It means outside readers cannot assess bargaining power without contract details. The fact pattern to watch is whether Dragonfly/Qfast has alternative origination options, multiple wholesale paths, and enough customer demand to negotiate rather than accept terms.

In short, the cost base is probably lighter than an infrastructure owner's but heavier than a simple software reseller's. Dragonfly/Qfast must pay for the privilege of being technical and accountable. Its upside depends on whether customers pay enough for that accountability to cover the supplier stack and engineering load.

Customers buy accountability, but substitutes are abundant

The positive customer thesis is accountability. Small and medium-sized businesses often do not want to coordinate KPN, Ziggo, Eurofiber, a voice supplier, a firewall vendor, a mobile backup provider and an IT consultant when something breaks. They want one provider to diagnose, route, escalate and fix. Qfast's portfolio is built for that demand. It can sell the fixed line, the backup path, the voice service, the VPN, the subnet and the support tier as one managed outcome.

Accountability is valuable in outage moments. If a dental practice loses connectivity, appointment systems, payment terminals, telephony and patient communication may all be affected. If a professional office loses voice or remote access, staff productivity falls quickly. If a multi-tenant building has poor telecom coordination, tenants complain to the landlord. In those settings, a provider that knows the customer's layout and has configured failover can be worth more than the cheapest access line.

The negative thesis is that accountability is hard to defend when substitutes improve. National operators have business service desks, managed routers, fiber portfolios, mobile backup products and cloud voice partnerships. Managed-service providers can bundle internet, firewall, Microsoft 365 and voice. SD-WAN vendors and cloud providers make failover easier. Landlords can be approached by alternative fiber providers. Many customers may not understand or value routed IP continuity until after an outage, and even then they may choose a familiar national brand.

Qfast's independence can cut both ways. The company says it is supplier-independent and can use multiple infrastructures. That can be useful if the best solution depends on location, redundancy or technology mix. It can also mean Qfast does not control the underlying network and must mediate when infrastructure suppliers have faults. The customer may appreciate that mediation, or may decide to buy directly from the infrastructure owner.

The Dutch broadband market makes substitution realistic. ACM has reported that fiber subscriptions overtook cable subscriptions, that high-speed access is widely available, and that KPN and VodafoneZiggo remain the largest fixed broadband players with each in the 35 to 40 percent range in the reported period. ACM has also stated that it sees sufficient competition in fixed telecom services while monitoring acquisitions and market developments. For a small business customer, that means there are multiple credible routes to connectivity.

The same market can still leave room for Qfast. Competition among infrastructure owners can benefit a service integrator. If fiber, cable, DSL, mobile and wholesale options are available, Qfast can select and combine them. If customers are confused by choice or need continuity across sites, Qfast can turn market complexity into a managed-service opportunity. In this model, the value is not scarcity of access. The value is decision-making, configuration and accountability.

The issue is customer segmentation. Qfast likely has stronger economics where customers have multi-site needs, public IP requirements, voice-dependency, 4G failover, building-level distribution, or a willingness to pay for support. It likely has weaker economics where customers only need a standard broadband connection and can compare monthly prices. Public materials show the company addresses both ends of that spectrum. They do not show the mix.

This makes management's incentive sharper. The company should want to move customers away from pure access resale and toward integrated continuity bundles. It should want to attach voice, VPN, backup and support to connectivity. It should want to make its resource-holder capability visible in outcomes customers understand, such as stable IP addressing, cleaner failover and easier multi-site management. It should avoid competing as a generic access reseller where the national network owners set the price.

Contract durability rests on integration rather than access alone

The most plausible source of contract durability is integration. A customer with only an internet line can switch providers at renewal if another offer is cheaper. A customer whose internet line, voice numbers, hosted telephony, routed subnet, firewall, VPN, 4G backup, support procedures and branch connections are all tied into Qfast has more switching friction. The business may still switch, but the risk, project work and downtime become higher.

Qfast's public materials show several ingredients for this kind of friction. Number portability and hosted telephony can embed Qfast in day-to-day communications. Routed subnets and WAN IP continuity can embed Qfast in remote access, payment systems, voice service or security arrangements. VPN and datacenter connections can embed Qfast in applications. Multi-tenant building service can embed Qfast in landlord operations. Tiered network support can make Qfast the first call during incidents.

These are not artificial lock-ins. They are operational dependencies. When delivered well, they create customer value and provider durability at the same time. The customer avoids repeated vendor coordination. The provider earns recurring revenue and learns the customer's environment. Over time, the provider can sell additional services because it already understands the network.

But integration durability is only as strong as execution. If Qfast's support is slow, if failover is not tested, if voice quality is unreliable, if a supplier outage leads to unclear responsibility, or if invoices become harder to understand than alternatives, the same integration can become a reason to leave. Customers that outsource accountability are also quick to blame the accountable provider when things fail. The model requires operational excellence, not just a product list.

The public evidence does not disclose churn, renewal rates or contract terms. Qfast's FAQ mentions framework agreements and the ability to scale hosted telephony up and down. That suggests a recurring relationship rather than a one-time sale, but it does not reveal term length or retention. References and testimonials provide some comfort that Qfast has served real business customers, but first-party references are selected by the company and do not substitute for independent retention data.

Contract durability also depends on customer size. Very small businesses may value Qfast's support but be price-sensitive. Larger SMEs may have more complex needs and higher willingness to pay, but they may also run competitive procurement and demand stronger service-level commitments. Multi-site customers and multi-tenant buildings may be more durable if Qfast becomes part of the operating fabric. The public record does not provide enough customer mix detail to know which group dominates revenue.

The investment conclusion must therefore be conditional. If Qfast has a meaningful base of integrated customers buying internet, voice, backup, VPN, routed IPs and support under recurring contracts, Dragonfly's resource-holder status has economic value. It supports continuity and makes Qfast more than a simple reseller. If most revenue comes from single-service access lines or low-seat hosted telephony, the resource-holder status may be under-monetized. It would still be useful, but not enough to overcome price pressure.

Regulation raises the floor and limits the rent

Dutch telecom regulation gives Dragonfly/Qfast a credible operating floor but not a monopoly rent. ACM's provider registration rules require public telecom providers to notify the authority when they offer public electronic communications networks or services. Business.gov.nl explains that providers must register with ACM unless they are only acting on behalf of an already registered provider. The Dutch Radiocommunications Agency also describes obligations around availability, authenticity, safety, security and integrity for public telecom providers.

Qfast's own registration page states that it is registered with ACM and refers to telecom-industry bodies and supplier relationships. That matters because customers can see that Qfast is not merely an informal IT installer. It operates in a regulated environment and presents itself as a telecom provider. For business customers, especially those depending on telephony and internet continuity, that can support trust.

Regulation also limits easy rent extraction. A registered provider must operate under market rules, security expectations and customer-service obligations. In a competitive market, regulatory legitimacy does not allow a small provider to charge monopoly prices. It allows the provider to participate. The provider still has to win customers from incumbents, other resellers and managed-service firms.

ACM's market commentary is relevant. The authority has described fixed telecom competition as sufficient while still monitoring market structure and acquisitions. It has also reported fiber growth, high-speed availability and the continued dominance of KPN and VodafoneZiggo in fixed broadband market shares. That combination means a provider like Qfast operates in a market where customers have choices, infrastructure is improving, and scale players remain powerful.

The earlier wholesale-access regulatory history is also useful context. In 2018 ACM imposed wholesale fixed-access obligations after finding joint significant market power between KPN and VodafoneZiggo, though the later market context changed. The broader lesson is that smaller providers often depend on access to larger networks. Even when regulation or wholesale products create room for competitors, the economics of those competitors depend on wholesale terms, customer acquisition and service differentiation.

For Dragonfly/Qfast, regulation may help by preserving a market in which alternative providers can operate and customers can buy services across multiple infrastructure owners. It may also impose obligations that raise the professionalism floor. But regulation does not answer the margin question. It cannot turn a thin resale model into a high-margin managed service. It can only create the conditions in which the provider tries to differentiate.

This is why resource-holder status and registration should be seen as enabling assets. They make the company more credible and technically capable. They do not, by themselves, prove pricing power. The public record supports the idea that Dragonfly/Qfast belongs in the Dutch telecom-service ecosystem. It does not prove that the company captures economics beyond a modest provider role.

The market signal is credible niche presence, not disclosed scale

Qfast has a credible niche signal. The site is detailed, not skeletal. It describes multiple products in practical terms, lists support tiers, publishes telephony tariffs, provides contact information, names registration bodies, and presents first-party references and partners. The RIPE and routing evidence connects the commercial surface to real Internet number resources. Nextpertise's wholesale model offers a plausible operating path. This is enough to treat Dragonfly/Qfast as a real niche telecom-service entity rather than a paper-only registration.

What is missing is scale. There is no public revenue line. There is no EBITDA, gross margin, cash-flow or capital-expenditure disclosure. There is no customer count, ARPU, churn, contract duration or top-customer concentration. There is no disclosure of how many services are sold per customer, how much revenue comes from voice versus access versus managed services, how much traffic or address space is active, or what portion of delivery depends on Nextpertise.

Public company profiles do not fill the gap. Company.info and Creditsafe confirm incorporation and registry facts, but detailed financials are not publicly available in the materials reviewed. The SBI classification as a financial holding adds caution, because the legal company named in RIPE records may not map one-for-one to the operating economics of Qfast-branded services. A holding structure can be ordinary, but it means the article should not imply that Dragonfly's public registration record alone reveals operating revenue.

Public routing also argues against overstating scale. If AS31172 were announcing multiple prefixes and maintaining visible peering diversity, one might infer a more independent network role. The current evidence shows no AS31172 announced prefixes and Dragonfly-linked resources originated via AS41960. That pattern can still support customers, but it does not look like a large independent backbone. It looks like a smaller provider using a partner's network scale.

First-party references are useful but limited. Qfast lists customer names and testimonials, including examples from dental and business customers. These references support the market fit for SME continuity services. They do not show renewal rates, margin, contract size or customer concentration. A handful of happy references can coexist with either a healthy niche book or a small low-margin book.

The better public conclusion is therefore: credible niche presence, undisclosed scale. Dragonfly/Qfast has more evidence than a generic dormant company, because the registry, resource and service records align. It has less evidence than a proven compounder, because the financial and customer disclosures are absent and current routing shows dependency on a wholesale partner. The company may be profitable and durable, but the public record does not prove it.

The economic story remains attractive only if the hidden facts break in the right direction. The customer base must value integrated support and continuity enough to pay premium margins. Supplier terms must leave room after wholesale cost. Engineering and support costs must be controlled. Contract retention must be high. The resource-holder position must help win or retain customers rather than merely sit behind a partner-originated route. Without those facts, the fair view is disciplined uncertainty.

Cash capital needs are lower than network ownership, but not trivial

One reason the Dragonfly/Qfast model may work is that it avoids the worst capital burden of telecom. Building a national access network, owning ducts and fiber at scale, or operating a mobile network requires heavy capital, long payback periods and regulatory complexity. Qfast's model appears to use other parties' infrastructure while adding service design, resource control, support and customer management. That can be capital-light relative to network ownership.

Capital-light does not mean cost-free. Customer equipment, routers, voice devices, monitoring tools, support systems, installation coordination, datacenter connections, interconnects and staff time all require cash. If Qfast promises 24/7 support, failover, managed security or site-to-site networks, it must maintain the capability to deliver those services. The provider also needs working capital if it pays suppliers before customers pay, or if installation charges are recovered over contract terms.

The need for engineering capability may be the most important hidden investment. A provider that handles routed IPs, private interconnects, hosted telephony, VPNs, 4G failover and multi-tenant buildings cannot rely entirely on sales scripts. It needs people who understand routing, CPE behavior, voice quality, network monitoring, escalation paths and customer environments. Skilled engineering labor is not cheap, and a compact organization can be vulnerable to key-person risk.

The public records name contacts and directors, but they do not disclose staffing depth. Qfast's "compact organization" language can be positive if it means low overhead and quick service. It can be a risk if customers depend on a small number of engineers or managers. A service provider selling accountability must have enough operational redundancy inside its own team, not only in the networks it resells.

Capital needs also depend on customer acquisition. If Qfast wins customers through referrals, building relationships, and add-on services to existing accounts, acquisition cost may be manageable. If it must compete in paid advertising or procurement processes against larger operators, acquisition cost could rise. The website's first-party references suggest relationship selling, but there is no data on sales efficiency.

The resource-holder aspect has a modest but strategic cash profile. RIPE fees are small relative to telecom infrastructure, but keeping registry resources, route authorizations, security practices and technical administration in order requires discipline. The real value is optionality: Dragonfly can maintain stable IP resources and potentially move or negotiate network delivery if wholesale relationships change. That optionality is worth something, but only if management can exercise it without disrupting customers.

Thus, cash capital needs are likely manageable compared with infrastructure ownership, but the model is not as asset-light as a pure software subscription business. The provider sells operational assurance. Assurance requires staff, processes, supplier management and technical assets. The economic question is whether customers pay enough recurring gross profit to fund that assurance and leave a return.

The conclusion turns on margin, not technical legitimacy

The public record establishes technical legitimacy. Dragonfly Holding B.V. is a Dutch RIPE NCC member and LIR. It is associated with AS31172 and Internet number resources. Qfast presents a coherent telecom and ICT service portfolio. Qfast states that it is registered with ACM. Current routing evidence shows Dragonfly-linked resources operating through Nextpertise AS41960 with valid RPKI state. None of that looks imaginary.

The public record does not establish margin strength. It does not show that Dragonfly/Qfast earns premium gross margins from resource-holder status. It does not show that customers pay materially more because Qfast can supply routed IP continuity or managed failover. It does not show that the company has enough scale to spread engineering and support costs. It does not show that supplier terms are favorable enough to protect margins.

This distinction is important because telecom analysis often confuses technical seriousness with economic power. A small provider can run real services, hold real resources, support real customers and still earn modest returns. Conversely, a company with no national network can earn attractive returns if it owns high-trust customer relationships and keeps supplier costs flexible. Technical legitimacy is the starting line, not the finish line.

For Dragonfly/Qfast, the highest-value customer is probably not the customer buying one cheap line. It is the customer that values continuity and wants one accountable provider across access, backup, voice, IP addressing, VPN and support. That customer may be willing to pay for Qfast's knowledge and responsiveness. The lower-value customer is the one comparing monthly broadband prices and willing to switch for a small saving. The public service menu spans both types, so the unknown mix is decisive.

Current routing evidence also changes the language of control. Dragonfly may hold resources, but Nextpertise appears to provide current public origination for key prefixes. That can be efficient, but it means Dragonfly's resource position is mediated by a wholesale partner. The economic prize is not simply "owning IPs"; it is using IP control to make customer continuity more durable while negotiating supplier terms that leave margin.

The cautious conclusion is that Dragonfly/Qfast has a believable niche but not a publicly proven moat. It can plausibly earn value where service complexity and continuity matter. It remains vulnerable where access is commoditized, supplier costs rise, customers demand direct national-operator service, or support obligations exceed contract value. Resource-holder status improves the company's toolkit. It does not automatically move the company above price-taker economics.

What would change the judgment

The judgment would become more positive if public or verifiable evidence showed that Qfast has a recurring base of integrated business customers buying multiple services per site, with low churn, multi-year terms and meaningful support attach. A disclosed customer count is less important than quality of revenue. One hundred customers buying complex, sticky, multi-service continuity packages could be better than many more customers buying low-margin access lines.

Gross margin disclosure would matter most. If Qfast earns healthy gross margins after wholesale access, mobile backup, voice platform, equipment and support costs, the resource-holder model has economic force. If margins are thin and support-intensive, the model may be operationally useful but financially limited. Segment-level margin would be ideal: business fiber, 4G backup, voice, VPN, managed support and multi-tenant building services likely have different economics.

Supplier exposure would also change the view. Evidence that Dragonfly/Qfast has multiple independent origination options, favorable wholesale agreements, redundancy across access providers and the ability to move resources without customer disruption would reduce the Nextpertise concentration risk. Evidence that most public routing and wholesale reach depends on one partner with limited fallback would increase the risk.

Customer proof would help if it went beyond testimonials. Case studies with specific continuity outcomes, uptime performance, failover events, multi-site deployments, tenant counts, service durations or renewal behavior would show that customers value Qfast's integrated proposition. Independent reviews, procurement wins or public customer references with detail would be stronger than logo lists.

Network evidence could also change the view. If AS31172 resumed visible announcements, maintained diverse upstreams or peers, or carried customer traffic directly, the analysis of control would shift. That would not automatically prove profitability, but it would show a more independent network role. If current AS41960 origination remains the model, the analysis should focus on wholesale terms and customer integration rather than autonomous-system independence.

Finally, financial filings or management disclosure could resolve the holding-company uncertainty. The public record now shows Dragonfly as a financial holding and Qfast as the operating-facing telecom brand. A clearer map of ownership, operating subsidiaries, revenue flows and asset ownership would help determine whether Dragonfly's resource-holder status sits inside a profitable operating model or mainly supports a smaller service operation.

Until those facts are available, the fair answer to the core question is restrained. Dragonfly Holding B.V. has enough public evidence to be treated as a genuine resource-holder behind a business telecom proposition. It does not have enough public evidence to prove that resource-holder status creates durable economic power. The most likely defensible position is a service-integrator niche: valuable when customers need continuity, routing and accountable support; exposed when customers buy access as a commodity. Below cloud scale, the company earns its right to margin one customer problem at a time.