21 Enterprises, LLC and the cash-flow test behind local network reliability

Summary

• 21 Enterprises, LLC is publicly evidenced as a United States RIPE NCC local internet registry and resource holder, not as a clearly documented mass-market broadband, cloud, hosting, or managed-network operator. The strongest proof is administrative and routing evidence: a RIPE organisation record, AS60506, two visible IPv4 prefixes, an abuse role, and a parked company domain.
• The live routing footprint is narrow but real. RIPEstat shows AS60506 announced on the review date, with two visible IPv4 prefixes containing 512 addresses and no visible IPv6 route. The same data shows one observed neighbour, AS23352, while the RIPE aut-num entity still contains older import and export policy references to other ASNs. That gap matters because routing records can show intent, history, or administrative policy without proving current capacity or supplier diversity.
• The economics should be tested as a cash-flow problem, not as a label problem. A two-prefix network can support paid workloads, private address customers, research infrastructure, small hosting, security testing, or a specialised connectivity use case, but the public record does not prove subscriber count, recurring revenue, physical facilities, service levels, support staffing, or sales execution.
• The operating risk sits where a small resource holder has the least public margin for error: single-neighbour dependence, scarce IPv4 stewardship, abuse workload, domain and mail outsourcing, RIPE fee obligations, and weak public-market signalling. The upside is control over a compact routable footprint; the downside is that reliability costs arrive whether or not customers pay a visible premium for them.

One dependent account is enough to expose the economics

Start with a modest customer rather than a grand network map. Imagine a small software firm, security lab, remote-office service, or data-handling workload that pays 21 Enterprises, LLC for a connection, a routed address, a server-facing network path, or some other specialised use of its number resources. That customer does not care whether the provider describes itself as a regional ISP, a resource holder, a hosting shop, a private network operator, or a lean infrastructure company. It cares whether packets move, mail reaches the right administrative contact, abuse complaints are answered, and the assigned address space remains reachable.

The monthly fee behind that account has to carry more than bandwidth. It has to pay for the upstream relationship, routing administration, registry membership, domain registration, DNS, abuse handling, router configuration, monitoring, incident response, and the managerial time needed to keep a small network inside the rules of the internet number-resource system. If the service is sold as reliable, it also has to carry redundancy, backup access to skilled operators, replacement hardware, and some cushion for mistakes.

If the fee is too low, the customer receives the appearance of independent addressing while the provider absorbs the real cost base. That is the cash-flow test behind local network reliability.

The public record for 21 Enterprises is unusually useful precisely because it is thin. There is enough evidence to show that the company controls a recognised network identity and appears in the RIPE NCC member universe. There is not enough evidence to prove a large customer base, a retail service catalogue, a robust multi-carrier design, a revenue base, or a full operating team. The analysis therefore should not pretend that an ASN equals a business model. The right question is what kind of business could be supported by the visible assets and what would have to be true for those assets to produce durable cash.

That framing keeps the analysis disciplined. A resource record can be valuable even when the company has no visible storefront. A parked domain can be compatible with a private customer base, an internal network, a legacy project, or a provider that sells through relationships rather than public checkout pages. A single upstream can be rational for a small footprint, especially if the workload is not mission critical. It can also be a warning that the reliability promise is narrower than customers might assume.

What is proven about 21 Enterprises

The strongest evidence identifies 21 Enterprises as a United States organisation in the RIPE NCC system. The RIPE organisation record names 21 Enterprises, LLC, gives the organisation type as local internet registry, records the country as the United States, and includes an Illinois registration number. It also lists a Chicago mailing address and references the maintainer used by the organisation. The related aut-num entity identifies AS60506, names OUTHTECH, links the ASN to the same organisation, and marks the AS as assigned.

This is not incidental evidence. Local internet registry status is a formal role inside the RIPE NCC system. It means the organisation is recognised for managing internet number resources in that registry environment, with duties around database accuracy, resource stewardship, and contactability. It does not mean the organisation sells broadband service to households, operates a data centre, owns fibre, runs cloud compute, or has a particular revenue level. The record establishes administrative standing, not the full commercial surface.

The RIPE member list for the United States includes 21 Enterprises among local internet registries offering services in the United States. That supports the regional classification used by the directory snapshot, but it still has to be read carefully. RIPE NCC membership and service-area listing are number-resource and membership signals. They do not prove a retail access network in a city, a sales team, support desks, installation crews, or a published service level. They show that the company is a recognised entity in number-resource administration.

The domain evidence is lean. The company domain is registered through Gandi, has Gandi nameservers, uses Google mail exchangers, and resolves on plain HTTP to a Gandi parked-domain page. A local HTTPS connection attempt did not produce a working site. That pattern supports a narrow conclusion: the domain is maintained and usable for administrative identity, but it is not being used as a public commercial storefront in the reviewed evidence. That is important because it constrains what can be said about customer acquisition. If customers exist, the public web is not the main evidence of how they are sold.

The abuse role is also real. RIPE records include an abuse contact tied to the organisation. Abuse handling is part of the operating cost of address space. Even a small prefix can attract complaints if a customer host is compromised, if addresses are leased badly, or if a block is used by a third party without adequate controls. A network that cannot process abuse reports can lose supplier confidence, have prefixes filtered, or damage the reputation of the address space. That work is invisible when it goes well, but it consumes time when it goes badly.

The boundary therefore is clear. 21 Enterprises is proven as a legal-named resource holder and RIPE local internet registry with an active AS and associated domain. It is not proven, from the public record reviewed here, as a conventional access ISP, a national carrier, a cloud platform, a hosting brand, a managed-service provider, or a business with publicly visible recurring revenue. The economic analysis has to stay inside that boundary.

The live routing footprint is narrow but real

AS60506 is announced. RIPEstat's AS overview marks it as visible, and the routing-status data shows two IPv4 prefixes and 512 IPv4 addresses in visible announced space on the review date. The announced-prefix data identifies the two visible routes as 81.199.24.0/24 and 81.199.25.0/24 over the two-week window ending on the review date. No visible IPv6 route is shown in that same status view.

For a small operator, two visible /24s are meaningful. A /24 is the minimum practical IPv4 advertisement size on much of the global internet, so two /24s can be independently originated, filtered, protected, or moved under a common ASN. The blocks can support customer addressing, small hosting, private workloads, lab infrastructure, or a compact access customer base. They are also scarce assets. IPv4 address space is costly to obtain, administratively valuable, and operationally sensitive because poor use can poison reputation.

The footprint is also small. Five hundred and twelve visible IPv4 addresses do not by themselves support claims of broad retail scale. They could be enough for a handful of customers with public addresses, a private network use case, a specialised server operation, or a low-density access service. They could also be heavily underused. The public routing table does not disclose assignment density, paying customer count, address utilisation, revenue per address, or how many addresses are reserved for infrastructure.

RIPEstat sees one neighbour for AS60506: AS23352. IPinfo and bgp.tools also show AS23352 as the single visible upstream in their current summaries. That is strong evidence of current single-neighbour dependence in the observable topology. It does not prove that only one contract exists, because private backup arrangements, dormant sessions, out-of-band management, or low-visibility routes may not appear in every public view. But for a customer judging public internet reachability, the visible topology looks concentrated.

The RIPE aut-num entity introduces a useful inconsistency. It contains import and export policy references to AS2856 and AS13335, rather than only the currently observed neighbour. Those records may reflect older relationships, policy templates, dormant arrangements, or entries that have not been kept aligned with live routing. They should not be treated as proof that British Telecommunications or Cloudflare is actively carrying the network today. They are better read as evidence that registry policy data and observed BGP data can diverge. In a small network, that divergence is not rare, but it is still a diligence point.

Third-party datasets add another caution. Some databases list only the two RIPEstat-visible /24s. One public IP intelligence source associates a third /24, 45.140.245.0/24, with 21 Enterprises or AS60506, while other current routing evidence does not show it in the same way. That discrepancy should not be forced into a simple story. It may reflect historical routing, delegated space, geolocation database lag, suballocation, a low-visibility route, or stale enrichment data. The prudent conclusion is that the visible current routing base is two /24s, while some data vendors preserve broader or inconsistent address associations.

That distinction matters commercially. If the company sells reliability, the current visible route set is the deliverable. If it sells address administration, historical or delegated address relationships may matter. If it sells private services, the public table may understate the business. A buyer or customer needs to ask not merely what records exist, but which routes are live, who can change them, who pays for them, and who is responsible when a path disappears.

Number-resource evidence is not proof of monetised service

The easiest mistake is to convert registry records into a service claim. An ASN with two prefixes is evidence of routing capability. It is not proof of retail internet service. A local internet registry record is evidence of number-resource administration. It is not proof of last-mile network ownership. A domain and abuse mailbox prove contact surfaces. They do not prove sales, uptime, support capacity, or customer satisfaction.

That separation is central to 21 Enterprises. The company may have paying users, but the public materials do not show a tariff book, product page, installation process, cloud offering, data-centre location, support terms, customer count, case studies, speed plans, status page, or public service-level agreement. A public storefront may not be necessary if the company serves a small private set of customers. It is still an evidence gap. Without a sales surface, outside observers cannot infer pricing, margin, contract length, or product mix.

The business model could be several things. It could be a compact network serving a small number of customers that need routable IPv4. It could be a private infrastructure operation maintaining addresses for internal or related-party workloads. It could be a resource-administration vehicle. It could lease, sponsor, route, or otherwise support address space for third parties. It could have a legacy operating history that is not visible through the current domain. The evidence does not select one model decisively.

Each model has different economics. A retail broadband model needs customer acquisition, support, field operations, access circuits, and churn management. A hosting model needs facility or facility-supplier contracts, power, hardware, remote hands, and security response. A resource-administration model needs compliance, documentation, abuse handling, and counterparty controls. A private-network model needs only enough uptime to satisfy the internal workload, but it still pays fixed registry and transit costs.

The danger is not that the company lacks value. The danger is overclaiming the kind of value. A compact routable footprint can be valuable if it is attached to sticky demand. It can be a liability if it is attached to low-paying customers, one upstream, high abuse, or weak documentation. A public route can support a serious workload; it can also sit mostly idle while fees continue. Revenue quality, not resource labels, decides which is true.

The revenue question is contribution per usable address

No public financial statements were found in the reviewed material, so revenue has to be analysed by unit economics rather than reported totals. The relevant unit is not only the address. It is the paying workload or account that uses the address, plus the monthly contribution left after upstream, support, registry, domain, compliance, and incident costs.

A two-prefix IPv4 footprint creates 512 visible addresses before reserving space for network infrastructure, management, customer gateways, monitoring, and unusable or held-back addresses. If every usable address were monetised as a bare address rental, monthly revenue would depend on address price, abuse risk, payment reliability, and the cost of managing counterparties. If the addresses support hosted workloads, the address revenue is only part of the bundle; compute, storage, power, support, and facility costs dominate.

If the addresses support access customers, address scarcity may be less important than last-mile cost and monthly subscriber churn.

For a small resource holder, the cash-flow test is unforgiving. Fixed costs do not shrink just because the visible routing table is small. RIPE membership fees and administrative obligations arrive regardless of how many addresses are used. Domain registration and mail service are small but recurring. Upstream connectivity, router capacity, monitoring, and incident response require either vendor payment or skilled labour. Abuse complaints can turn one low-margin customer into a high-cost account. If the company sells to customers that need public IPv4 but do not pay for support, the margin is fragile.

The upside is scarcity. IPv4 availability remains constrained across the ARIN and RIPE environments, and public market reports continue to treat IPv4 as a priced asset rather than a free utility. A company with clean, routable space has optionality: it can number its own services, provide customers with public addressing, move workloads between suppliers, or potentially participate in transfer or leasing arrangements subject to policy and contract limits. That optionality is economically real. It becomes operating profit only when attached to paying demand and disciplined controls.

A useful internal metric would be monthly gross contribution per routed /24. A second useful metric would be abuse-adjusted revenue per customer or per assigned address. A third would be the cost of keeping the AS reachable per month, including upstream, registry, monitoring, and labour. If those metrics are not known, the company may be preserving an asset without knowing whether the service around it earns enough to justify the attention it requires.

The public record does not tell us whether 21 Enterprises earns strong contribution, weak contribution, or no external revenue. It tells us that the company has a compact, active routable footprint. The economic judgment must therefore stay conditional: the footprint is valuable if customers pay for dependable use of it, but public evidence does not show that they do.

Cost base: the small network still has grown-up obligations

Small networks can look cheap because the route table is small. That view misses the fixed nature of internet operations. The first routed prefix needs much of the same administrative machinery as the tenth: registry records, abuse contact, routing policy, upstream coordination, monitoring, DNS, security hygiene, and someone accountable when the route fails. Those obligations scale down imperfectly.

The RIPE cost base is not the largest expense, but it is visible and non-discretionary for a member that wants to remain in good standing. The RIPE NCC 2026 billing material sets annual membership and resource-related fees. A small network can absorb that if it has even modest recurring revenue. It becomes more meaningful if the business is dormant, experimental, or attached to only one or two accounts. A fee that is trivial for a national carrier is a real hurdle for a tiny resource holder.

Upstream connectivity is the larger operating dependency. If AS23352 is the only visible neighbour, 21 Enterprises depends on that relationship for current reachability. The public data does not show price, capacity, committed information rate, burst terms, contract length, outage credits, maintenance windows, or whether the connection is physically diverse. A network can survive with one upstream if its customers understand the risk and the price is aligned. It cannot honestly sell high resilience on that basis unless there is undisclosed backup that performs under stress.

Operational labour is the hidden cost. Someone has to keep route filters current, respond to abuse, maintain contact records, renew domain services, handle supplier tickets, and understand the difference between a routing issue, a DNS issue, a customer compromise, and a registry problem. For a lean company, that knowledge may sit with one person. If so, resilience is not just a network-topology question; it is a staffing and documentation question.

Abuse is especially important. Address space used for hosting, proxying, scanning, or loosely controlled customer workloads can create disproportionate operating burden. A single problem customer can generate complaints, blacklists, supplier pressure, and manual cleanup. Strong terms, identity checks, monitoring, and fast suspension rights protect the address asset. Weak controls convert scarce IPv4 into reputational debt.

The domain posture also shows outsourcing. Gandi nameservers and Google mail exchangers are sensible choices for a small operation; they reduce the burden of running everything in-house. They also show that the company's public administrative surface depends on outside providers. That is not a weakness by itself. It is part of the cost and supplier map.

Supplier dependence is the central operating risk

For 21 Enterprises, supplier dependence begins with AS23352. The observed topology shows one neighbouring AS. If that neighbour withdraws service, filters the routes, suffers a regional problem, changes commercial terms, or requires urgent abuse cleanup, 21 Enterprises has limited public evidence of an alternative path. A small customer may tolerate that if the service is priced as best effort. A customer paying for reliability should not.

The RIPE aut-num references to AS2856 and AS13335 complicate the picture. If those relationships are inactive, the database should not be read as current diversity. If they are standby or historical, a buyer needs to know the conditions under which they can carry traffic. If one of them is a private or tunnel-based arrangement not visible through ordinary observation, it should be documented for customers that rely on it. Public routing evidence and registry policy evidence need reconciliation.

The domain stack adds two more supplier categories: Gandi and Google. Gandi appears in the registration and nameserver evidence, while the mail exchanger records point to Google mail infrastructure. That is a normal small-business architecture. It means administrative reachability depends partly on registrar, DNS, and mail-provider continuity. Those services are generally more resilient than a self-run setup, but account control, payment, and access recovery become operational controls.

If the company offers customers any kind of service, its supplier map may also include data-centre space, virtual servers, remote hands, hardware vendors, payment processors, and support tools. None of those were visible in the public record. The absence is not proof they do not exist. It means the current outside view cannot assess supplier concentration beyond the network and domain layers.

Supplier dependence can be managed. The company could maintain a second visible upstream, keep current route objects and authorisations, publish basic network information, test failover, and document who can access registrar and RIPE accounts. Those controls do not require a large marketing site. They require disciplined operations. In a compact network, discipline may matter more than scale.

Customer concentration is unknowable from the public surface

The biggest commercial unknown is not whether the company has 512 visible IPv4 addresses. It is who pays for them. A small footprint can be healthy if it is tied to a few creditworthy customers with clear contracts and low abuse. It can be risky if it depends on one customer, one related-party workload, short-term renters, or accounts that require intensive support. The public record does not show which case applies.

Customer concentration changes the downside distribution. If one customer accounts for most of the routed use, losing that customer can leave the company paying registry and upstream costs without matching revenue. If many customers use small allocations, the company needs support processes and abuse controls. If the network serves internal workloads, the revenue may be implicit rather than external, and the relevant question becomes whether those workloads justify continued number-resource expense.

The parked domain suggests little public customer acquisition. That can be rational for a private network or referral-based provider. It also reduces market visibility. There are no public plans, no service description, no published service levels, no status history, and no clear customer endorsements in the reviewed set. The company may not need them. But without them, outside analysis cannot give high marks for commercial proof.

The customer-side value proposition would likely be one of four things: public IPv4 access, routing independence, a small specialised hosting or network path, or continuity of a local or private workload. Each is a different product. Public IPv4 customers pay for address access and reputation. Routing-independence customers pay for control. Hosting customers pay for operational reliability. Private workloads pay indirectly by avoiding dependence on a larger cloud or carrier. A company can serve more than one of these, but the costs and risks differ.

The fee must match the risk. A customer that only needs an inexpensive address should not expect custom restoration. A customer that depends on the route for revenue should pay for redundancy and documentation. A provider that accepts high-risk workloads must charge enough for abuse work. The public evidence does not show whether 21 Enterprises makes those distinctions.

Competition is broader than local broadband

The public category places the entity in a regional ISP frame, but the competitive set should be wider because the public proof is number-resource and routing oriented rather than access-network oriented. The substitutes include wholesale transit, cloud networks, managed hosting, address leasing or transfer options, registrar and DNS platforms, DDoS-protection networks, and larger service providers that can bundle connectivity with security and support.

If a customer needs only compute reachable from the internet, the large cloud platforms and hosting companies are the most obvious alternatives. They offer global capacity, automated provisioning, billing, support tiers, and security tooling. Their weakness is loss of control over address identity, location, and sometimes cost predictability. A small independent network can win where control, special routing, address reputation, or personal service matters more than scale.

If a customer needs network-layer protection or global ingress, Cloudflare-style Magic Transit and similar services are substitutes for parts of the value chain. They do not replace the need for an origin network, but they can absorb attacks, announce customer space, and provide a managed front door. For a small operator, such services can be either competitors or suppliers. They reduce the need to build every defensive capability in-house, but they add recurring cost and another dependency.

If a customer needs IPv4 space, the substitute is not simply another ISP. It may be the transfer market, a sponsoring local internet registry, a larger hosting provider that includes addresses, or a cloud platform that charges for public addresses as part of the service. IPv4 scarcity gives a resource holder bargaining power, but it also disciplines customer willingness to pay. Customers can compare the price of direct addressing against managed alternatives.

If the customer needs local access reliability, the substitutes are different again: a second access provider, a mobile backup, a fixed wireless link, satellite, or private transport. The public evidence does not show 21 Enterprises owning local last-mile infrastructure, so it would be speculative to compare it directly with household broadband providers. The safe comparison is between a compact autonomous network and the larger suppliers that can provide internet reachability or infrastructure without the same local identity.

The competitive conclusion is mixed. Scale providers have better public proof, broader support, and stronger redundancy. A compact network has flexibility, control, and potentially lower bureaucracy. The winner depends on the customer's value of independence. A customer that wants commodity uptime will choose the larger platform. A customer that values a specific routable footprint may pay a smaller operator if the controls are credible.

Regulatory and geopolitical exposure sit inside the registry relationship

The company is United States-based in the RIPE records, but its number-resource relationship sits in the RIPE NCC environment. That creates a cross-border administrative context. It is not inherently problematic; many companies operate across registry regions for historical, commercial, or network-design reasons. It does mean the company must stay attentive to RIPE policy, RIPE billing, due diligence, transfer restrictions, sanctions screening, and database accuracy.

RIPE transfer policy matters because IPv4 is scarce and transferable under rules. A resource holder cannot treat addresses as an ordinary asset detached from registry obligations. Transfers must be reflected in the database, scarce-resource restrictions can apply, and sanctions screening can affect requests. The economic value of the footprint depends not only on market demand but on clean documentation and policy-compliant control.

The United States corporate layer matters too. The RIPE organisation record includes an Illinois registration number, but the reviewed evidence does not provide a clean public status report from the Illinois database for the company. A buyer, customer, or supplier should verify current good standing directly. If the legal entity behind the resource record is not current, contract enforceability, supplier onboarding, and registry due diligence become harder.

Sanctions and abuse controls are not abstract. RIPE membership processes and transfer evaluation include checks tied to EU sanctions. U.S. counterparties may also face U.S. sanctions and export-control obligations depending on customers, destinations, and services. Nothing in the reviewed evidence indicates that 21 Enterprises is sanctioned. The point is narrower: a small cross-border resource holder needs customer screening and documentation because address space can be used by counterparties in ways that create legal and reputational exposure.

Data sovereignty is also a customer question. The RIPE and domain records identify the company as U.S.-based, while third-party geolocation and routing records can associate addresses with different places or suppliers over time. Geolocation databases are imperfect, but customers that care about locality, jurisdiction, or compliance should not rely on address labels alone. They need actual facility, route, counterparty, and data-handling representations.

Unofficial market signals are weak and noisy

The unofficial signals around 21 Enterprises are not strong enough to support a broad quality claim. Some IP intelligence pages classify the ASN as ISP, business, or data-centre related, but those labels are vendor taxonomies. They are useful for search and risk triage, not proof of what the company sells. BGP data shows reachability. It does not show customer experience.

The parked domain is a negative signal for public-market visibility but not necessarily for operations. Many small infrastructure companies avoid public e-commerce because they sell by contract, support related workloads, or use the domain only for administrative contact. Still, a parked domain makes it harder for potential customers to inspect terms, support scope, and operating history. That raises diligence cost.

Third-party address datasets are mixed. The strongest current view shows two announced /24s. Some enrichment sources add a third associated range or show geolocation details that differ from the current RIPEstat-visible picture. Such mismatches are common in IP intelligence. They should be treated as reasons to ask for live route, ROA, delegation, and assignment evidence, not as settled facts.

The absence of public reviews is also ambiguous. A consumer ISP with no reviews would raise one kind of question; a private resource holder with no reviews raises another. For this company, the public evidence leans toward a quiet resource-administration and routing footprint rather than a public consumer brand. The lack of chatter therefore mainly confirms that the market surface is private or narrow.

The facts that would change the judgment

The judgment would improve quickly with a few concrete disclosures. First, the company could show current customer or workload categories without naming sensitive accounts: access, hosting, address administration, private infrastructure, or internal use. Second, it could publish or provide under diligence the active upstream list, physical or logical redundancy design, committed capacity, and failover test results. Third, it could reconcile RIPE aut-num policy references with currently observed BGP neighbours.

Revenue evidence would matter more than labels. Recurring revenue by product, churn, address utilisation, top-customer concentration, gross margin after upstream and support, and abuse tickets per month would show whether the footprint funds itself. A network with modest but stable revenue and low abuse can be attractive even at small scale. A network with opaque revenue, one customer, and high incident load is fragile.

Resource hygiene would also change the view. Current route authorisations, accurate contact records, maintained role entities, documented assignments, and clear abuse procedures would support a higher evidence score. Stale records, inactive contacts, unresolved abuse, or inconsistent address claims would lower it. The small size of the footprint means each record matters more, not less.

Legal verification is another swing factor. Current Illinois good standing, authorised signatories, and alignment between the legal entity and the RIPE organisation record would reduce counterparty risk. If the legal entity has changed, merged, or lapsed, the resource records would need clean documentation.

Finally, a public service statement would help. It does not need to be a marketing site. A concise network page naming AS60506, the service boundary, published contact points, peering or transit policy, abuse handling, and customer eligibility would materially improve trust. It would also reduce the risk that customers infer a broader reliability promise than the company intends to sell.

Verdict: a compact routable asset, not a proven broad ISP

21 Enterprises is best understood as a compact U.S.-based RIPE resource holder with an active autonomous system and two visible IPv4 /24 routes. That is a real operating asset. It gives the company control over a small routable footprint and potential leverage in a market where public IPv4 remains scarce. It can support specialised customer value if the customers need address control, routing identity, or a private infrastructure relationship.

The public evidence does not justify describing the company as a broad regional access ISP or cloud-service provider. There is no visible retail service catalogue, no public revenue base, no customer count, no facilities proof, no published service levels, and no confirmed multi-upstream live topology. The company may have private operations that are stronger than the public surface suggests. They cannot be assumed.

The core economic question is whether paying accounts fund the full reliability burden. Who pays is likely a small set of private workloads or customers, if any external customers exist. Who benefits is any party that receives reachable, controlled IPv4 service without having to manage its own registry and upstream stack. Who carries the downside is the operator if fees are too low, customers if reliability is sold without redundancy, and the address asset itself if abuse or poor documentation damages reputation.

The judgment is therefore cautious. The resource evidence is credible; the commercial evidence is incomplete. 21 Enterprises can be economically interesting if it has sticky demand, clean customers, documented controls, and at least a credible backup plan for current single-neighbour dependence. Without those facts, the network should be valued as a small, administratively real footprint whose reliability claim still has to pass the cash-flow test.