The registrar's invoice is small enough to look uninteresting. A wholesale .be registration or annual renewal is listed by DNS Belgium at EUR 5.50 excluding VAT, while the same table puts .brussels and .vlaanderen at EUR 20.00 for comparable core transactions (https://www.dnsbelgium.be/en/registrars/become-registrar/fees-registrars). On a single domain, the difference between EUR 4.00 in an older price environment and EUR 5.50 from 2025 is not the sort of cost that breaks a web agency, a hosting company or a brand owner. On a portfolio of tens of thousands of names, it becomes a real line item. On the registry's side, it is the recurring annuity that funds the boring part of the Belgian internet.
That is why DNS Belgium is more interesting than its size suggests. The organisation is not a venture-backed software company trying to grow at all costs. It is a Belgian not-for-profit association established in 1999, responsible for .be and later for .brussels and .vlaanderen, with a governance structure that includes a general assembly, a board, a finance committee and a remuneration committee (https://www.dnsbelgium.be/en/about-dns-belgium/organisation). The point of the model is not maximal novelty. It is continuity. A registry is paid when names are created, renewed, transferred, restored or protected, and it is judged most harshly when users notice it.
The strategic question is whether DNS Belgium is pricing a commodity database entry or a bundle of trust. The evidence leans toward the second reading, but not without limits. In 2025, there were 207,776 new .be registrations and 1,701,999 active .be names by year-end, according to DNS Belgium's annual reporting (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). Yet the total .be base still declined by 16,117 names, or 0.94 percent, while .brussels and .vlaanderen fell faster in percentage terms (https://www.dnsbelgium.be/annual-report-2025). Mature registry economics can therefore feel paradoxical: the product remains essential, the brand remains trusted, new creations can recover, and still the installed base can soften.
The result is a business where "boring" is not a lack of ambition. It is the service promise. DNS Belgium's own statistics page places .be in a long historical arc: introduced in 1988, transferred to DNS Belgium in early 2000, liberalised later that year, then lifted by steady post-2000 growth, a 2005 promotion, the pandemic-era online rush and a post-2023 break in trend (https://www.dnsbelgium.be/en/statistics). That history matters because the registry's current problem is not proving that a Belgian domain can exist. It is proving that it remains worth renewing when businesses can live on platforms, app stores, marketplaces, social profiles, search listings and generic domains.
The fee is visible; the service behind it is intentionally quiet
The cleanest way to see DNS Belgium's economics is through the registrar price table. The registry charges registrars, not end users directly, for most domain transactions. A .be registration, renewal, transfer to another registrar, or transfer to another holder carries the same EUR 5.50 wholesale fee. Reactivation or transfer from quarantine costs EUR 10.00. Domain Guard is EUR 80.00 per year, while Domain Shield is listed without charge (https://www.dnsbelgium.be/en/registrars/become-registrar/fees-registrars). The core model is therefore recurring and transaction-linked, with modest optional protection fees layered around a large installed base.
The 2025 contract and billing changes made this more visible. DNS Belgium told registrars in November 2024 that from 1 January 2025 they would pay EUR 5.50 for a new domain, a renewal, a registrar transfer or a holder transfer, while quarantine transfer would drop to EUR 10.00 from EUR 40.00 (https://www.dnsbelgium.be/en/news/essential-changes-registrars-2025). The same notice also described changes to billing, credit and invoicing. That combination matters. A registry fee increase is not merely a price point. It changes how registrars fund prefunded accounts, how they explain wholesale costs to resellers, and how much margin remains when .be is bundled into hosting, email, site-builder or agency packages.
The arithmetic is small per name and meaningful in aggregate. At 1,701,999 active .be domains at year-end 2025, the current EUR 5.50 annual fee maps to about EUR 9.36 million of gross wholesale renewal surface if the stock is treated as a one-year base before churn, new adds, timing and VAT (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025 and https://www.dnsbelgium.be/en/registrars/become-registrar/fees-registrars). At the older EUR 4.00 reference point, the same stock would imply about EUR 6.81 million. That simple comparison does not replace DNS Belgium's accounts, but it shows why a EUR 1.50 uplift is strategically material: across the visible .be stock it is a roughly EUR 2.55 million annualised swing before the effect of new registrations, transfers and non-renewals.
For the registrar, the fee can feel like rent because the registry has no direct like-for-like substitute inside the .be namespace. If a Belgian bakery wants example.be, it cannot buy the same address from another registry. For DNS Belgium, the fee can be defended as insurance because the organisation is expected to maintain availability, data integrity, abuse response, compliance, registrar support and policy governance across the national domain. Neither side is making an absurd claim. The economic bargain depends on whether the registry keeps the wholesale fee low enough that registrars can compete and high enough that the trust layer does not decay.
The retail market shows the spread between wholesale and customer-facing pricing. EuroDNS listed a .be annual offer at EUR 10.39 against a regular EUR 19.00 renewal, while showing transfer, trade and renewal prices around EUR 19.00 and bundling extras such as DNS, email and SSL services (https://www.eurodns.com/domain-extensions/be-domain-registration). Those retail numbers are not DNS Belgium revenue. They show the channel economics around the registry: registrars turn a registry fee into a package, discount acquisition, recover margin through renewal or attached services, and compete on convenience as much as on the bare domain.
The registry's advantage is that its revenue base renews annually. Its constraint is that registrars and registrants can reduce domain portfolios, move projects to other TLDs, stop defensive holdings, or compress low-value names when the economy weakens. DNS Belgium's 2025 annual report shows both sides: new .be registrations rose, but total .be domains fell; renewal trends slipped modestly; and concentration among the largest registrars increased (https://www.dnsbelgium.be/annual-report-2025). That is the profile of a mature annuity, not a growth engine.
DNS Belgium is a national registry, not just a website brand
The identity facts are unusually clear. DNS Belgium says it is a Belgian not-for-profit organisation established in 1999. It was formerly known as DNS.be and changed its name at the end of 2012 because it was taking responsibility not only for .be but also for .brussels and .vlaanderen (https://www.dnsbelgium.be/en/about-dns-belgium/organisation). It names Philip Du Bois as general manager and describes the board and assembly structure of the association. That places the organisation closer to a public-interest infrastructure operator than to a conventional domain retailer.
The historical frame is also important. DNS Belgium's statistics page says .be began in 1988 under EARN and Eunet, with Professor Pierre Verbaeten as manager. DNS Belgium took over in early 2000, and liberalisation later that year opened the namespace on a first-come, first-served basis (https://www.dnsbelgium.be/en/statistics). That transition from academic-era stewardship to open commercial registration is the origin of today's annuity. The registry inherited a national technical role and then built a channel model around accredited registrars.
The channel model makes DNS Belgium powerful but indirect. Registrants usually experience the registry through a registrar interface. They search a name, pay a retailer, renew through a retailer, receive validation notices, change holder details through a retailer and learn about redemption through a retailer. DNS Belgium sits underneath that journey, setting rules, fees, system availability and security procedures. This is why public trust in .be can be high even when few registrants can name the registry that keeps it running.
The organisation also has a network presence independent of its public website. PeeringDB lists DNS Belgium VZW networks for AS21239, labelled as the registry network for .be, .vlaanderen and .brussels, and AS199670, labelled as anycast (https://www.peeringdb.com/org/16297). PeeringDB's AS199670 page lists a non-profit network type, four IPv4 prefixes, four IPv6 prefixes and a Europe geographic scope (https://www.peeringdb.com/net/12676). bgp.he.net shows the same AS199670 with eight originated prefixes, all RPKI-valid at the time observed, and 183 BGP peers (https://bgp.he.net/AS199670). Those are not brand claims. They are operating clues.
The RIPE membership evidence aligns with the assignment record. The RIPE member directory has a DNS-Belgium VZW member page for Belgium (https://www.ripe.net/membership/member-support/list-of-members/be/dnsbe/), and public routing references identify ORG-DA132-RIPE as DNS-Belgium VZW with country BE (https://bgp.tools/as/199670). This does not mean the article entity is an ASN, a prefix or a routing table. It means the organisation operating the registry also appears in public internet resource records in ways consistent with critical DNS infrastructure.
The distinction matters because a registry's value is partly institutional and partly technical. DNS Belgium has to persuade registrars that policy is predictable, persuade government that national internet dependency is managed, persuade security partners that abuse response is serious, and persuade network peers that DNS traffic will remain available. That bundle is difficult to see in a domain fee, but it is what the fee funds.
The annuity is mature, and the growth story has narrowed
DNS Belgium's own 2025 data shows a registry with a resilient core and a less forgiving market. New .be registrations rose to 207,776, around 7,500 more than in 2024 (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). That is not a dying namespace. It suggests Belgian businesses, sole traders, associations and private users still see value in a local address. DNS Belgium connected the increase to new business creation and the continued importance of owning a web presence.
Yet total .be domains fell to 1,701,999 at the end of 2025, and DNS Belgium attributed part of the decline to one large foreign food-delivery platform operator removing more than 5,000 inactive domains from its portfolio (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). That detail is revealing. Domain portfolios can inflate a registry in good years and deflate it when a holder rationalises names. A single corporate cleanup does not change national demand by itself, but it reminds investors and registrars that registry volume includes defensive, inactive and campaign-driven holdings, not only live businesses.
The 2025 annual report adds more pressure points. DNS Belgium recorded a 0.94 percent decline for .be, a 2.51 percent decline for .brussels and a 6.50 percent decline for .vlaanderen (https://www.dnsbelgium.be/annual-report-2025). It also described a slight negative renewal trend: down 1.00 percent for .be, 2.00 percent for .brussels and 1.00 percent for .vlaanderen. A small renewal deterioration is financially important because renewals, not new registrations, are the stabilising base of the registry annuity.
The weak performance of .brussels and .vlaanderen also matters. These TLDs extend DNS Belgium's role beyond the Belgian country code, but they have narrower natural demand. The 2025 annual report notes that .brussels and .vlaanderen are used above all by respective governmental authorities, and for .vlaanderen the Flemish government chose to redirect its .vlaanderen names to an existing .be website (https://www.dnsbelgium.be/annual-report-2025). That is a blunt signal: regional identity can be politically meaningful without producing a large commercial namespace.
Competition does not have to take .be's exact place to cap its pricing power. A Belgian company can use .com, .eu, .net, a new generic extension, a marketplace storefront or a social commerce profile. DNS Belgium's statistics page itself notes market fragmentation after the arrival of new generic TLDs such as .shop and .app (https://www.dnsbelgium.be/en/statistics). This makes the .be value proposition less automatic than it was in the early years of liberalised national domains.
But .be retains a trust advantage that competitors cannot easily copy. A local ccTLD signals place, language fit and regulatory familiarity. EuroDNS markets .be as open to anyone but aimed at Belgium, and says local-language website content and a local extension can raise trust for Belgian customers (https://www.eurodns.com/domain-extensions/be-domain-registration). That is a registrar's sales copy, not a neutral academic proof. Still, it reflects a channel view that .be remains a credible retail hook.
The annuity therefore depends on staying boring enough to be trusted and active enough to remain relevant. DNS Belgium cannot simply harvest a mature base. Its 2025 "growing together" initiative asked registrars for ideas on awareness, startup activation, renewals, predictive analysis and registration experience (https://www.dnsbelgium.be/en/news/working-together-grow-be-domain-names). That is the behaviour of a registry that knows the installed base cannot be taken for granted.
Benchmarks make the .be price argument harder
The strongest way to judge DNS Belgium's fee is not to call EUR 5.50 high or low in isolation. It is to compare the price with neighbouring registry scale and cost pressure. SIDN says the basic annual .nl charge to registrars is EUR 4.38 excluding VAT in 2026, after EUR 4.25 in 2025, and frames that price as competitive against an average basic price of about EUR 7 across other registries (https://www.sidn.nl/en/our-prices and https://www.sidn.nl/downloads/2uj5dECp1ctM88wFtpJqDT/6e515ea196e6493209303ebf6b65779c/SIDN_Annual_Report_2025.pdf). DNS Belgium's EUR 5.50 is above .nl's published price, but below SIDN's stated wider European average. That makes .be neither a bargain basement extension nor an obvious outlier.
The .nl comparison is especially useful because SIDN is facing the same mature-registry equation at larger scale. SIDN's 2025 annual report says .nl fell to 6,059,392 domains, the first contraction since SIDN was created, while turnover rose by EUR 0.3 million to EUR 25.9 million because the registry fee went up despite the shrinking zone (https://www.sidn.nl/downloads/2uj5dECp1ctM88wFtpJqDT/6e515ea196e6493209303ebf6b65779c/SIDN_Annual_Report_2025.pdf). SIDN also says the higher fee was needed because of inflation and investment in ICT systems, and that another 2.9 percent increase in 2026 helps cover expected one-off and structural costs from the Hello Registry migration. That is almost the same argument DNS Belgium has to make, except DNS Belgium has a much smaller .be base over which to spread fixed obligations.
Nominet shows the other side of the benchmark. Its FY25 annual report says the .UK register fell to 10.2 million domains from 10.7 million, group revenue slipped to GBP 55.9 million, registry revenue fell by GBP 1.0 million and the renewal rate was 77.0 percent (https://nominet.uk/wp-content/uploads/2025/07/FY25-annual-report-and-accounts.pdf). Yet Nominet's registrar fee schedule still lists a member wholesale registration and renewal price of GBP 3.90 per annum for .uk, with the same annual price in place since January 2020 (https://registrars.nominet.uk/uk-namespace/managing-account/payments/fee-schedule/). A larger registry can hold a lower per-domain fee longer because it has scale, reserves and a broader cost base. DNS Belgium does not have that cushion at the same magnitude.
EURid and DENIC sharpen the contrast further. EURid said its .eu portfolio stood at 3.79 million registrations at the end of 2025 with an 80.5 percent renewal rate, and that its abuse-warning system identified more than 56,000 suspicious registrations and led to nearly 30,000 suspensions after verification checks (https://eurid.eu/fi/news/eurid-annual-report-2025-strengthening-europes-dig/). DENIC's 2025 activity report puts .de at 17,663,886 domains, with 2,167,487 held by foreign owners, while its direct customer price list shows EUR 79.00 per year for DENICdirect administration rather than exposing the cooperative member economics that sit behind ordinary registrar retail offers (https://www.denic.de/en/2025-activity-report/ and https://www.denic.de/en/price-list/). These are very different models, but they prove the same point: scale, channel structure and abuse spending shape fee politics.
Afnic's .fr is the optimistic counterexample. Its 2025 review reported 4,319,120 .fr domains at year-end, 853,000 new create operations and 2.4 percent growth (https://www.afnic.fr/en/observatory-and-resources/news/2025-review-of-the-fr-a-new-all-time-record-for-create-operations/). That makes .fr a reminder that a European national namespace can still expand when domestic demand, creation flows and retention line up. DNS Belgium's .be is not in that position in 2025. It has better new-registration momentum than the headline decline suggests, but the active stock is slipping. That turns the fee discussion from growth monetisation into margin defence.
The benchmark conclusion is uncomfortable for both sides of the channel. Registrars can say, fairly, that .be is more expensive than .nl and .uk at the published registry-fee level. DNS Belgium can answer, also fairly, that .be has less scale than .nl, .uk, .eu, .fr and .de, while carrying the same family of security, compliance, registrar-support and infrastructure burdens. The right question is therefore not whether EUR 5.50 is high in the abstract. It is whether DNS Belgium can keep showing that the extra cents above the cheapest peers buy measurable abuse reduction, resilient DNS, smoother compliance, better registrar tools and sovereign supplier optionality.
Cost is shifting from registration volume to resilience work
The clearest cost signal in DNS Belgium's public reporting is people and infrastructure, not a line-by-line product margin. The 2024 financial report says DNS Belgium had 40 employees at year-end, with 34 full-time and six part-time, one departure and three hires (https://www.dnsbelgium.be/en/Financial-report-2024). It also says the organisation bought its own office building in Leuven during 2024. Those are fixed-cost choices. A registry with a slightly declining domain base still needs security staff, registry platform engineers, support, legal capability, financial controls, communications and governance.
The same report's balance-sheet section lists total assets of EUR 8,135,127 for 2024, down from EUR 9,103,232 in 2023, while capital and reserves were EUR 1,563,579 in 2024, down from EUR 2,504,014 (https://www.dnsbelgium.be/en/Financial-report-2024). It also lists suppliers at EUR 739,426, taxes, remuneration and social security costs at EUR 1,221,367, advances received from registrars at EUR 1,166,438 and accrued charges and deferred income at EUR 3,357,174. These figures are not enough to produce a full margin model, but they show the balance-sheet mechanics behind a prepaid registrar channel and a staff-heavy infrastructure organisation.
Buying an office building can look unrelated to domain pricing. DNS Belgium framed it as a strategic balance-sheet choice that keeps the organisation tied to Leuven, where it originated (https://www.dnsbelgium.be/en/Financial-report-2024). For a not-for-profit registry, that type of asset decision changes financial risk differently than a commercial growth bet would. It can reduce landlord exposure over time, but it also ties capital to a physical asset while domain volumes soften. The important point is that registry trust is financed through a relatively small organisation with real fixed commitments.
Operational resilience has become a more explicit part of the cost base. DNS Belgium described a setup with local nameservers at major Belgian internet operators Telenet, VOO, Proximus and Orange, intended to keep .be available even during cyberattacks or other incidents (https://www.dnsbelgium.be/en/news/strong-internet-infrastructure). It said these nameservers are geographically distributed copies that provide identical information for .be, .vlaanderen and .brussels websites and email. That is a good example of the hidden cost behind a cheap domain. Users do not pay extra because their ISP has a local fallback path; the registry pays to make failure less visible.
Security work is also becoming more expensive in quality terms. DNS Belgium upgraded the signing algorithm for .be, .vlaanderen and .brussels in January 2026 from RSA/SHA-256 algorithm 8 to ECDSA Curve P-256 with SHA-256 algorithm 13 (https://www.dnsbelgium.be/en/news/dns-belgium-modernises-algorithm-security-be-vlaanderen-and-brussels). The organisation said the newer algorithm cuts package size roughly in half while providing equal or better brute-force protection, with different computation tradeoffs. This is not a marketing add-on. It is infrastructure maintenance that keeps the zone aligned with current security practice.
DNSSEC also illustrates the difference between zone security and registrant adoption. DNS Belgium's DNSSEC explainer says DNSSEC links digital signatures to DNS data so resolvers can check authenticity, while noting that it does not solve typosquatting or phishing (https://www.dnsbelgium.be/en/secure/dnssec). A 2023 DNS Belgium article based on a sample of .be domains said only 30 percent were equipped with DNSSEC at that time (https://www.dnsbelgium.be/en/news/be-safe-there-always-room-improvement). The registry can modernise the zone, but registrant-side adoption and registrar support remain uneven. That means the cost of trust is shared, but the registry often absorbs the blame when the public sees failure.
Abuse handling has become a product feature
The biggest change in registry work is that abuse response has moved from exceptional cleanup to day-to-day product quality. DNS Belgium reported that malicious .be revocations fell by 20 percent in 2025, from 1,679 in 2024 to 1,346 in 2025, and that around 200 websites were blocked in collaboration with the government (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). The registry said the decline did not mean abuse disappeared; it reflected earlier detection before suspicious sites went live.
That early detection has a name in DNS Belgium's public material: RegCheck. DNS Belgium said in December 2025 that RegCheck predicts whether a new registration looks suspicious based on historical registration data, and that it was developed with SIDN, KU Leuven researchers and DNS Belgium's R&D team (https://www.dnsbelgium.be/en/news/blocking-suspicious-domains-they-go-live). The tool had been in production since March 2024, replacing a more manual rules system. DNS Belgium said the model lets staff verify registrations more selectively instead of spending time on low-risk cases.
The economics are subtle. A stricter pre-activation filter can reduce fraud but increases false positives, registrar friction and human review. A looser filter protects conversion and registrar throughput but lets more bad registrations resolve. DNS Belgium's RegCheck article says a wide net can block roughly 30 percent of registrations and find about 80 percent of fraudulent registrations, while warning that strictness creates more innocent cases needing review (https://www.dnsbelgium.be/en/news/blocking-suspicious-domains-they-go-live). That is the registry's pricing problem in miniature: protection is valuable until it slows legitimate commerce.
Abuse work extends beyond new registrations. DNS Belgium, SIDN and the Irish registry studied more than 28,000 domains involved in phishing across .be, .nl and .ie (https://www.dnsbelgium.be/en/news/international-study-phishing-using-cctlds). The study found that 80 percent of phishing attacks used compromised websites rather than newly registered domains, and that phishing across the studied ccTLDs spanned 78 countries and 114 market segments. It also said 75 percent of fraudulently registered .be names were addressed at DNS level, with DNS Belgium acting directly in 49.6 percent of cases.
That last point changes the registry's role. If most phishing uses compromised existing domains, registry action at the point of registration is not enough. The registry has to collaborate with registrars, security researchers, government bodies and hosting providers, while deciding when DNS-level action is appropriate. DNS Belgium's June 2026 article on suspicious-domain data says blocklists and abuse feeds are useful but insufficient, and that organisations need to combine multiple data sources to detect bad registrations early (https://www.dnsbelgium.be/en/news/which-data-helps-you-identify-suspicious-domain-names). The registry is becoming a data-analysis and judgement organisation, not merely a zone file publisher.
Neglected domains add another cost line. In September 2025, DNS Belgium said it had identified 1,300 vulnerable .be domains out of more than 1.7 million because of dangling records, where outdated DNS settings could be exploited by criminals (https://www.dnsbelgium.be/en/news/warning-abuse-neglected-domain-names). The registry personally warned administrators and then publicised the problem. This is not directly a registration-fee event. It is a safety campaign caused by configuration mistakes among holders. The registry's brand absorbs the reputational risk even where the holder caused the exposure.
The policy documents reinforce the same pattern. Enom's .be support page says DNS Belgium can require extra validation for suspicious registrations, can temporarily delegate names to DNS Belgium nameservers pending proof, and can revoke a domain if data is not corrected within set periods (https://support.enom.com/support/solutions/articles/201000065307--be-domain-policies). OpenSRS gives similar operational guidance for .be renewals, deletion windows and revocation after inaccurate details (https://support.opensrs.com/support/solutions/articles/201000063468--be-domain-policies). These registrar support pages are market signals: the registry's abuse and data-accuracy rules show up as operational obligations in reseller systems.
Regulation is turning verification into shared infrastructure
NIS2 has made registrant data a regulated operating surface. DNS Belgium's NIS2 guidance says Belgian registrars and resellers, plus those with a legal EU representative in Belgium, fall under the Belgian NIS2 Act, and that entities offering domain registrations face specific duties around keeping, verifying, publishing and providing lawful access to registrant contact details (https://www.dnsbelgium.be/en/getting-started-nis2-registrar-and-reseller). It also says registries, registrars and resellers had to register with Belgium's national cyber security authority by 18 December 2024.
For DNS Belgium, this creates a two-sided burden. It must comply as critical internet infrastructure and help the registrar channel comply without destroying the ease of .be registration. DNS Belgium's .be registrar agreement update said version 6.3 would enter force on 1 January 2025 and added future registrar audits, mandatory abuse contact sharing in WHOIS, cooperation with partial bulk transfers and NIS2-linked contact verification changes (https://www.dnsbelgium.be/en/news/update-be-registrar-agreement-6-3). This is a concrete example of European regulation converting a registry's public-interest role into contract language.
The outside policy context confirms why this matters. The DNS Research Federation's NIS2 explainer says Article 28 of NIS2 requires TLD registries and domain registration service providers to collect and maintain accurate and complete registration data with due diligence, and to maintain verification procedures (https://dnsrf.org/blog/nis2-and-domain-names---an-explainer). It also notes that registration data duties affect registries, registrars, resellers and proxy providers. That means DNS Belgium's policy changes are not isolated Belgian bureaucracy. They sit inside a wider European attempt to make domain ownership data useful for security while preserving data protection limits.
The commercial effect lands on registrars first. A registrar has to collect more accurate data, explain verification to customers, deal with blocked or deactivated names, fund account credit and absorb support tickets when a domain cannot be activated. DNS Belgium's 2025 changes put some duties into the registrar contract while saying the registry wants to lead where possible on verifying, maintaining and providing access to registrant data (https://www.dnsbelgium.be/en/getting-started-nis2-registrar-and-reseller). That is good channel politics, but it also raises the registry's operating burden.
This is where the EUR 5.50 fee becomes a governance question. If DNS Belgium underprices the namespace, it risks underfunding verification, support, legal review and security engineering. If it overprices, registrars can push customers toward alternatives, reduce promotional spend or let low-value names drop. NIS2 shifts the equilibrium upward by making accurate data a regulated expectation. The fee is no longer only about adding a row to a database; it is about making that row defensible.
There is also reputational asymmetry. If verification is too heavy, registrars and registrants complain that .be is harder to buy than another domain. If verification is too weak, victims, government bodies and security researchers complain that .be is unsafe. DNS Belgium's public positioning tries to occupy the middle: keep registration smooth, add smarter screening, keep humans involved and communicate with registrars. That is a narrow path, and it is more expensive than the old assumption that a registry can remain mostly passive.
Supplier dependence is now part of the trust product
DNS Belgium's most strategically revealing move is not the 2025 fee increase. It is the decision to leave AWS and reduce dependence on non-European technology. In December 2025, the organisation said it wanted to remove critical services from AWS and migrate to a European cloud provider, with the transition starting in 2027 and expected to complete in the second half of that year (https://www.dnsbelgium.be/en/news/dns-belgium-leaves-aws). It stressed that its nameserver infrastructure had never run on AWS and that the migration would concern systems such as the registration platform, not domain resolution itself.
The reason was political and strategic rather than technical failure. DNS Belgium said its registration system had been located in AWS European data centres since 2017 and remained aligned with GDPR and NIS2, but geopolitical realities made non-European dependence riskier (https://www.dnsbelgium.be/en/news/dns-belgium-leaves-aws). The registry wanted to avoid being forced into an unplanned move if cloud access, policy or tariffs changed. For a registry, this is not normal procurement talk. It is a redefinition of resilience from uptime to jurisdictional optionality.
The supplier review began earlier. In April 2025, DNS Belgium said it had started mapping non-European products and services in 2024 and was considering European alternatives because of geopolitical tensions and changing trade relations (https://www.dnsbelgium.be/en/news/supplier-analysis-driven-geopolitical-changes). It said there was no current problem with DNS security or availability, but that price, performance, security and world-order risk all belonged in supplier assessment. This is the kind of statement that would have sounded unusual for a domain registry a decade ago. It now reads like mainstream European critical-infrastructure governance.
The PostgreSQL migration made the strategy concrete. In June 2026, DNS Belgium said it had migrated the database behind its registration platform from Oracle to PostgreSQL, reducing dependence on non-European technology and making the future cloud move safer by separating the database conversion from the later provider migration (https://www.dnsbelgium.be/en/news/dns-belgium-swaps-oracle-open-source-alternative-postgresql). It also said Oracle licensing would be less flexible and potentially more expensive in a sovereign European cloud or private infrastructure. That is a pure cost-base insight: sovereignty can raise procurement complexity unless the technical stack is changed first.
This supplier strategy is financially ambiguous. Moving away from AWS, changing databases, testing platform migrations and reviewing vendors cost money and management attention. The benefits are risk reduction, bargaining power and political legitimacy. For a not-for-profit registry serving a national namespace, that tradeoff can be rational even if it would not maximise short-term surplus. The registry is trying to make trust visible before a crisis forces the issue.
It also strengthens DNS Belgium's argument to registrars. A registrar may dislike a wholesale increase, but the registry can point to concrete projects: local nameserver deployment, algorithm modernisation, RegCheck, NIS2 readiness, supplier review, PostgreSQL migration and the planned European cloud transition. Those projects are not generic overhead. They are the machinery behind a national domain that continues to resolve and continues to be acceptable to government, banks, SMEs and civil society.
The risk is execution. Cloud migration can create outages, registration interruptions, cost overruns or technical debt if poorly managed. DNS Belgium has already told stakeholders that domain resolution should continue and that registration interruptions, if any, would be planned and communicated (https://www.dnsbelgium.be/en/news/dns-belgium-leaves-aws). That makes 2027 an important test. A registry that sells boring trust must make the supplier transition boring too.
The registrar channel is both customer and constraint
DNS Belgium works with around 350 registrars for .be, according to the 2024 and 2025 annual reports (https://www.dnsbelgium.be/en/annual-report-2024 and https://www.dnsbelgium.be/annual-report-2025). That number is large enough to make channel governance complex. Registrars vary from local Belgian hosting companies and web agencies to international domain platforms. They compete with one another, bundle .be differently, carry different support costs and face different exposure to NIS2.
The channel is becoming more concentrated. DNS Belgium's 2025 annual report says the largest registrar group held 36.33 percent of .be names, up from 34.88 percent in 2024 (https://www.dnsbelgium.be/annual-report-2025). Concentration can help operational coordination because a few large registrars can implement changes quickly. It can also create bargaining power. If the largest registrars dislike pricing, validation rules or technical changes, they can influence retail promotion and customer experience.
DNS Belgium appears aware of that tension. The 2025 call for .be growth ideas asked registrars to contribute to awareness, activation, renewal, data analysis and user journey improvements (https://www.dnsbelgium.be/en/news/working-together-grow-be-domain-names). In March 2025, InternetNews reported that DNS Belgium opened applications for its registrar forum, with up to 20 members, three-year mandates and twice-yearly in-person meetings (https://www.internetnews.me/2025/03/04/dns-belgium-opens-call-for-registrar-forum/). The forum is a market signal: registrars are not just sales outlets; they are participants in product governance.
The channel's visible friction is practical. Enom tells .be customers that renewal is through auto-renewal only, that there is no renewal grace period, that redemption lasts 40 days and that certain contact changes require a trade with a new one-year term (https://support.enom.com/support/solutions/articles/201000065307--be-domain-policies). OpenSRS similarly describes one-year renewal terms and the absence of a renewal grace period (https://support.opensrs.com/support/solutions/articles/201000063468--be-domain-policies). Realtime Register's .be knowledge base walks through validation, DNSSEC, transfers and registry contact procedures (https://kb.realtimeregister.com/article/7-be). These pages are not complaints, but they show the operational specificity that registrars must support.
This is why fee perception matters. A registrar's visible retail price includes the registry fee, payment cost, support, compliance handling, billing risk, tax, customer acquisition and margin. If the wholesale registry fee rises while support and compliance burden also rises, the registrar may not simply pass through the increase cleanly. It may cut discounts, raise renewal prices, promote alternative TLDs, or reserve .be for customers with clear Belgian intent. That would not destroy .be, but it would weaken impulse and defensive registrations.
DNS Belgium's answer is to make the registry more valuable to registrars than a bare namespace. The updated registrar report, described in 2022, gave registrars market-share, renewal, revoke, bounce, infection and support-case data so they could compare themselves with peers and improve performance (https://www.dnsbelgium.be/en/news/updated-be-registrar-report). The registrar forum, growth initiatives and security data all fit the same strategy: give the channel better intelligence and make compliance feel like shared infrastructure rather than unilateral burden.
The danger is that not all registrars benefit equally. Large registrars can integrate APIs, run validation checks, manage credit, monitor portfolios and use registry reports. Small registrars may experience the same system as administrative drag. A mature ccTLD needs both groups. Large registrars bring volume; small local providers bring trust, language fit and SME reach. DNS Belgium's channel strategy has to keep both in the game.
Network evidence supports the boring-utility thesis
The public routing footprint does not make DNS Belgium a telecom carrier, but it does support the thesis that the organisation is a real infrastructure operator. PeeringDB lists AS199670 as DNS Belgium VZW Anycast, with a non-profit network type, Europe geographic scope, four IPv4 and four IPv6 prefixes, and low disclosed traffic levels (https://www.peeringdb.com/net/12676). It also lists AS21239 as DNS Belgium VZW Registry, with .be, .vlaanderen and .brussels described as the registry context (https://www.peeringdb.com/net/12675). The organisation page ties both networks to DNS Belgium VZW (https://www.peeringdb.com/org/16297).
bgp.tools provides more live-routing texture. It lists AS199670 as DNS-Belgium VZW, registered to be.dnsbe, with registration dating to 12 October 2012 and with internet exchange points including R_iX, BelgiumIX, AMS-IX, NL-ix, LINX LON1, BNIX and MIX-IT (https://bgp.tools/as/199670). Hurricane Electric's bgp.he.net page shows seven internet exchanges, eight originated prefixes, all valid under RPKI at the time shown, and visible peers across IPv4 and IPv6 (https://bgp.he.net/AS199670). These measurements are third-party observations, not DNS Belgium publicity.
The geographic spread is consistent with anycast logic. A registry wants DNS answers close to users and resilient to path failure. DNS Belgium's own March 2025 article about local nameservers at major Belgian ISPs describes a domestic layer for Telenet, VOO, Proximus and Orange customers (https://www.dnsbelgium.be/en/news/strong-internet-infrastructure). The peering pages show broader European connectivity. Together they suggest an operator that uses both local distribution and exchange presence to keep resolution dependable.
The traffic levels on PeeringDB are low, which is not a weakness by itself. Authoritative DNS traffic for a national TLD can be strategically important without being massive in bandwidth terms. The economic implication is that DNS Belgium's value is not measured by raw traffic volume. It is measured by the availability and integrity of small answers that enable millions of websites and email addresses. The cost of failure is asymmetric: a tiny DNS response can be worth little per query and enormous when unavailable.
RPKI validity is another small but meaningful signal. bgp.he.net shows AS199670 originated prefixes as RPKI-valid and zero invalid for the listed set (https://bgp.he.net/AS199670). That does not prove perfect routing security. It does suggest that the operator's public BGP posture is maintained in line with current routing hygiene expectations. For a registry whose own annual material stresses integrity and resilience, this consistency matters.
The network evidence should not be overstated. PeeringDB entries are maintained by network participants, BGP views vary by collector, and routing pages can change daily. But the pattern is coherent across DNS Belgium's own statements, RIPE membership context, PeeringDB, bgp.tools and bgp.he.net. DNS Belgium is not just a legal holder of a namespace. It has the visible network fabric of an organisation that must keep national domain infrastructure boring under pressure.
Market dependence is local, but not only local
The .be market is Belgian in trust logic but international in holder and registrar structure. DNS Belgium's 2025 reporting says new .be registrations rose even in a difficult Belgian economy with high bankruptcies and record new business creation (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). That connects the namespace to SME formation, local visibility and digital autonomy. It is a national demand story.
At the same time, DNS Belgium noted that a large foreign food-delivery platform operator removing more than 5,000 inactive domains contributed to the year-end decline (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). That is an international portfolio story. Defensive registrations, brand campaigns, platform landing pages and localised marketing portfolios can make a national ccTLD more volatile than its domestic business base alone would suggest.
The renewal study adds more nuance. DNS Belgium said it participated with nine other European ccTLD registries in a study of more than 40 million domains, and that for .be, registrants located in Belgium renew more often than .be holders outside Belgium (https://www.dnsbelgium.be/en/news/european-study-renewal-behaviour). It also said older domains renew more often than one-year-old domains. This fits ordinary domain economics: speculative or campaign names churn, while established names become embedded in email, search, signage, invoices and identity.
For DNS Belgium, local renewal strength is both a moat and a ceiling. The more Belgian a .be holder is, the stronger the trust signal and renewal habit. The more foreign or portfolio-driven the holder, the easier it is to cut names. If .be growth depends heavily on Belgian SMEs, public services, associations and local professionals, the namespace will track Belgium's business cycle and digital behaviour. If it depends heavily on foreign defensive portfolios, it will be vulnerable to corporate cleanup.
This dependence also shapes product messaging. DNS Belgium's 2025 public commentary emphasised owning a domain as a way to avoid dependence on large platforms and algorithms, but the article also showed that total volume can fall even when new registrations improve (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). The registry is making a strategic case for digital autonomy because the market no longer assumes that every business needs its own domain first.
Registrar retail pages turn that argument into sales language. EuroDNS says .be is unrestricted, aimed at Belgian audiences, and associated with local customer trust, while placing .be alongside .brussels, .vlaanderen, .gent, .lu and .nl (https://www.eurodns.com/domain-extensions/be-domain-registration). That competitor set matters. A Belgian business can signal place with .be, city or regional identity with other TLDs, Benelux reach with .lu or .nl, or global aspiration with .com. DNS Belgium's pricing power is strongest where .be is identity, weaker where it is merely one domain among many.
The market is therefore dependable but not captive. DNS Belgium's job is to keep .be worth renewing, not just available for registration. That requires security, reputation, registrar cooperation and a narrative that a Belgian domain is still a durable piece of digital infrastructure.
The competitive ceiling is reputation, not technical exclusivity
DNS Belgium has technical exclusivity over .be, .brussels and .vlaanderen, but the user's choice is broader than that. The first substitute is another TLD. The second is not owning a domain at all. The third is using a domain only as a redirect to a platform storefront or social page. Each substitute weakens the registry's ability to treat its fee as pure rent.
The .be brand still has strong advantages. It is short, familiar, national and old enough to feel safe. It is not a novelty extension whose meaning has to be explained. DNS Belgium's 2025 annual material ties .be to Belgian digital economy patterns, municipality-level density and local holder behaviour (https://www.dnsbelgium.be/annual-report-2025). The DNS Belgium statistics page says .eu had no measurable impact on .be in 2005, which suggests the Belgian ccTLD has historically resisted at least some regional competition (https://www.dnsbelgium.be/en/statistics). But the same page later notes fragmentation from new generic TLDs.
The more serious competitive issue is not .eu or .shop alone. It is whether a small business sees a domain as the primary home of its identity. DNS Belgium's own growth initiative asked registrars how to increase awareness, activate startups, encourage renewals and optimise the user journey (https://www.dnsbelgium.be/en/news/working-together-grow-be-domain-names). That is a concession that awareness can fade. A domain registry cannot rely forever on the idea that every new venture begins with a domain search.
Security reputation is therefore a competitive asset. DNS Belgium said .be ranks among the safest extensions in the world in its 2026 article on falling malicious revocations, while explaining that earlier interception lowered revocation counts (https://www.dnsbelgium.be/en/news/belgian-internet-continues-grow-nearly-208000-new-be-domain-names-2025). That claim is partly promotional, but the underlying strategy is credible: a ccTLD that becomes visibly associated with phishing, counterfeit shops or neglected configuration loses its local trust premium. Abuse handling protects demand.
The open registration policy creates a second tension. .be is unrestricted, so it can attract global registrants and international brand use (https://support.enom.com/support/solutions/articles/201000065307--be-domain-policies). Openness supports volume and retail simplicity. It also broadens abuse risk and weakens the automatic link between holder and country. NIS2, RegCheck and validation measures are ways to keep openness from damaging reputation.
Regional TLDs are a different competitive lesson. .brussels and .vlaanderen give DNS Belgium additional mandates, but their 2025 declines show that identity TLDs can remain symbolically useful without becoming broad commercial annuities (https://www.dnsbelgium.be/annual-report-2025). The registry cannot assume every managed string will behave like .be. The Belgian country code is the economic centre; the regional extensions are narrower public and identity instruments.
This is why the registry's best competitive defence is not a price war. It is making .be the low-friction, high-trust default for Belgian identity. That means staying affordable enough for registrars to promote, strong enough for security professionals to respect, and predictable enough for businesses to renew for years.
Non-official signals show a practical, not emotional, channel mood
The non-official market signals around DNS Belgium are not dramatic. That is itself informative. Registrar support pages describe .be as a rules-heavy but ordinary ccTLD: one-year terms, no renewal grace period, validation for suspicious registrations, holder-change trades, redemption or quarantine periods, and registry checks for inaccurate data (https://support.enom.com/support/solutions/articles/201000065307--be-domain-policies and https://support.opensrs.com/support/solutions/articles/201000063468--be-domain-policies). The tone is practical. Registrars are not selling a crusade; they are teaching customers how not to lose a name.
Retail pricing also looks pragmatic. EuroDNS packages .be as a Belgian trust signal with discounts, renewals and bundled services (https://www.eurodns.com/domain-extensions/be-domain-registration). It does not sell .be as exotic. It sells it as obvious for Belgium. That is good for DNS Belgium because obvious products renew. It also means the registry must avoid creating surprises that turn an obvious purchase into a support event.
Domain industry commentary around the registrar forum is more telling than ordinary retail pages. InternetNews described the forum as a way for DNS Belgium to present upcoming policy and technical changes and get feedback from channel partners, especially in a climate of growing regulation such as NIS2 and changing technical conditions (https://www.internetnews.me/2025/03/04/dns-belgium-opens-call-for-registrar-forum/). The framing is not anti-registry. It treats registrar consultation as increasingly necessary because domain policy is becoming more complex.
The market signal from PeeringDB and BGP tools is similarly practical. DNS Belgium's anycast network is present at several exchanges with low traffic and broad peer visibility (https://www.peeringdb.com/net/12676 and https://bgp.tools/as/199670). Network operators appear to treat DNS Belgium as a normal critical network that should be reachable. That is not a retail sentiment signal, but it supports the boring-utility thesis more than any slogan would.
There are also warning signals. DNS Belgium's own reporting about neglected domains, compromised sites and phishing shows that bad use of .be does not always start at registration (https://www.dnsbelgium.be/en/news/warning-abuse-neglected-domain-names and https://www.dnsbelgium.be/en/news/international-study-phishing-using-cctlds). If registrants fail to maintain DNS records or attackers compromise legitimate websites, the registry's preventive tools have limits. This can create pressure for more scanning, more warnings and more cooperation, all of which cost money but do not directly create new registrations.
The absence of loud registrar rebellion over the 2025 fee change is not proof of satisfaction. Domain registrars often respond through pricing, promotion and portfolio management rather than public argument. The better read is that the .be fee remains low enough to be absorbed, but the cumulative burden of NIS2, validation, billing changes and support can still affect retail enthusiasm. DNS Belgium's decision to invite growth ideas and maintain a registrar forum is therefore economically rational.
Market chatter should be read as a set of constraints, not as established fact. Registrar pages show where rules create customer friction. Retail pages show how much margin and packaging exist above wholesale. Industry commentary shows consultation demand. Routing databases show network seriousness. Together they support a view of DNS Belgium as a competent infrastructure organisation whose annuity is defensible, but whose channel goodwill must be renewed alongside the domains.
What would change the judgment
The positive case for DNS Belgium would strengthen if the .be base returned to stable growth while malicious-domain revocations stayed low for reasons tied to earlier interception rather than underreporting. New-registration growth alone is not enough. The stronger evidence would be rising active domains, healthy renewal rates by Belgian holders, lower abuse per million domains and registrar satisfaction after the 2025 contract changes. DNS Belgium already publishes annual and statistics pages, so future reports should make this visible (https://www.dnsbelgium.be/en/about-dns-belgium/annual-reports and https://www.dnsbelgium.be/en/statistics).
The negative case would strengthen if renewals kept sliding despite new registrations, especially if the decline spread from foreign portfolios to Belgian SME and association holders. A mature ccTLD can tolerate churn in campaign names. It cannot easily tolerate erosion in its local identity base. The 2026 renewal-behaviour study already points to local holders renewing more often (https://www.dnsbelgium.be/en/news/european-study-renewal-behaviour). If that pattern weakened, .be's trust moat would look thinner.
Supplier execution is another swing factor. The PostgreSQL migration appears strategically sound because it reduces the complexity of the later cloud move (https://www.dnsbelgium.be/en/news/dns-belgium-swaps-oracle-open-source-alternative-postgresql). The larger test is the planned AWS exit in 2027 (https://www.dnsbelgium.be/en/news/dns-belgium-leaves-aws). A clean migration would validate the argument that sovereignty and resilience can be bought without hurting service. A messy migration would make registrars question whether geopolitical risk management created operational risk.
NIS2 implementation could also change the economics. If verification duties become smoother, with registrars supported by registry-led tooling and clear data access rules, DNS Belgium's trust premium rises. If verification becomes slow, inconsistent or costly, registrars may treat .be as more burdensome than substitutes. DNS Belgium's updated registrar agreement and NIS2 guidance show the intended direction (https://www.dnsbelgium.be/en/news/update-be-registrar-agreement-6-3 and https://www.dnsbelgium.be/en/getting-started-nis2-registrar-and-reseller). The question is operational performance, not policy intent.
Abuse response metrics will be equally important. RegCheck's early result, a claimed 30 percent reduction in malicious registrations in the seven months after rollout, is a strong signal if sustained (https://www.dnsbelgium.be/en/news/blocking-suspicious-domains-they-go-live). But attackers adapt. If abuse shifts to compromised domains and dangling records, registry prevention must be matched by registrar and holder hygiene. Future data on blocked pre-activation names, false positives, DNS-level interventions and compromised-site response would make the trust economics clearer.
Finally, pricing remains the visible pressure point. EUR 5.50 is still low in the context of retail .be pricing, security work, staff, infrastructure, compliance and supplier migration. But small fees compound in large portfolios, and registrars can express dissatisfaction quietly. If DNS Belgium raises prices again without matching transparency on outcomes, the rent argument becomes stronger. If it ties fees to measurable resilience, lower abuse, better registrar tooling and successful European infrastructure migration, the cheap-insurance argument wins.
The present judgment is therefore constructive but conditional. DNS Belgium is a mature, credible registry with a defensible annuity and unusually clear evidence of operational investment. Its challenge is not to become exciting. It is to keep .be boring in a world where boring costs more: fewer easy domain-growth years, more security judgement, more data regulation, more supplier politics, and more pressure from registrars that must turn a wholesale trust product into a retail experience. The registry fee is not just rent, but it must keep proving that it buys more than permission to use two letters.