Standfirst. On 2026-08-18, webhosting.today published Natalia Nowak's market-by-market read of who answers DNS: Cloudflare is present everywhere and leads fewer than half of the markets it appears in, because national markets are held by national companies. That finding is right. We read the same markets on a different instrument — and the finding survives every altitude we can take it to. Then it goes two floors further down, and gets stronger.
External figures on this page (dotted underline) are webhosting.today's own, from that article as read on 2026-08-18, and are theirs throughout — they count resolving domains. Our figures come from a different population, the active-domains book — domains carrying a live business signal — and every one of them lives in the live-figures panel at the foot of this page, recomputed from the currently served dataset. The two instruments are never mixed in one number.
Their headline mechanism holds. Cloudflare sits in the top eight of all 115 markets they track and leads 46 of them; meanwhile 43 markets are led by companies present in four countries or fewer. A worldwide presence and a stack of national moats are two different maps of the same industry, and their piece is the clearest public statement of that difference we have seen.
The moats at the top of their leadership tiers are real. Their strongest national lead is hoster.by in Belarus, ahead of its nearest rival by 40.4 points on their reading — a figure measured in Belarus, on their instrument, and belonging to that market alone.
Where the two instruments read the same market, they agree. They read OVH at 36.0 percent of French resolving domains. Our instrument, counting only the active French book, reads OVHcloud at just over 36 percent — the exact share and the book it divides are in the panel below. Two crawls, two population definitions, the same answer to within a fraction of a point. That agreement is worth stating out loud, because it is what licenses the next two rungs: when the instruments differ below, it is the question that changed, not the quality of either reading.
The word "leader" in their chart names a brand. Ownership is a different column — and it is a column their own dataset carries. So the fold below is, first of all, their data joined against their data: their chart of national leaders, joined with their own group-brands pages. Where their group data is silent, we added a row from our press-validated acquisition ledger, with the primary source named. Thirteen charted leaders, one row each, every share theirs:
| Market | Their charted leader (their share) | Owner after the fold | Class | Fold source |
|---|---|---|---|---|
| Belarus | Hoster.by — 49.6% | — | no fold row in either dataset | independent as far as both datasets can see — a floor, not an assertion |
| Bulgaria | SuperHosting.bg — 47.5% | team.blue | acquired brand of a pan-European consolidator | their group-brands data |
| Slovakia | Websupport — 44.4% | team.blue | acquired brand of a pan-European consolidator | their group-brands data |
| Norway | Domeneshop — 42.9% | Miss Group | acquired brand of a pan-European consolidator | their group-brands data; our acquisition ledger corroborates |
| Estonia | Zone.eu — 41.3% | your.online | acquired brand of a pan-European consolidator | our registry only — your.online's own announcement, effective 2026-08-03 |
| Lithuania | Serveriai — 38.7% | Miss Group | acquired brand of a pan-European consolidator | our acquisition ledger only — iv.lt's own announcement, 2020 |
| Russia | Reg.ru — 35.6% | Runiti | a domestic group's brand or flagship | their group table — their own top-ten group |
| Italy | Aruba — 35.6% | Aruba S.p.A. | a domestic group's brand or flagship | their group table — a group's own flagship |
| Iceland | ISNIC — 34.5% | — | ccTLD registry | the country's own registry operator — its own class, not an independent host |
| Sweden | Loopia — 33.8% | team.blue | acquired brand of a pan-European consolidator | their group-brands data |
| Japan | Xserver — 30.2% | Netowl / XServer | a domestic group's brand or flagship | their group table |
| Brazil | NIC.br — 28.4% | — | ccTLD registry | the country's own registry operator — its own class, not an independent host |
| Ukraine | Hosting Ukraine — 25.1% | — | no fold row in either dataset | independent as far as both datasets can see — a floor, not an assertion |
Their chart was read 2026-08-18; their group-brands data is our snapshot of their market-insights pages, last modified 2026-07-24 — two vintages of one source, stated separately. Our two added rows carry their own primary sources, cited in the text below.
The counts, our arithmetic on the attributed rows above: nine of the thirteen charted leaders fold into a group. Six of the thirteen belong to three pan-European consolidators — three to team.blue, two to Miss Group, and one to your.online. The rest of the fold is domestic: a Russian group's storefront, a Japanese group's flagship, an Italian group under its own name.
What stands outside the consolidator fold is Belarus, Ukraine, and two national registries.
Norway is the corroboration row — the one market where our shares exist on both sides of the join. Their Domeneshop reads 42.9 percent of Norwegian resolving domains; our instrument reads Miss Group at about 44 percent of the Norwegian active book (the exact share and its base are in the panel). Domeneshop is a Miss Group brand per the press-validated acquisition ledger — an acquisition webhosting.today itself reported in 2024 — and their group-brands data agrees. The numbers agree almost exactly; only the word changes: brand on their axis, owner on this one.
Measured as market power, Norway's hosting-relationship concentration index — computed on the active book with CDN, SaaS-platform and DNS-infrastructure actors removed and the base renormalised — sits well above the 2,500 line that competition regulators call highly concentrated. The index and the book it is computed on are in the panel.
Two rows are what our instrument adds. Lithuania's Serveriai is a storefront for Interneto vizija (iv.lt), acquired by Miss Group in 2020 — per iv.lt's own announcement — and their group pages carry no row for it. Estonia's Zone.eu joined your.online effective 2026-08-03 — per your.online's own announcement — fifteen days before their article's read date. No criticism travels with that row: group membership moves faster than any snapshot, ours included; it is exactly why a fold column needs a dated source per row rather than a memory.
Germany, deliberately, is not a reveal — it is a credit. Their own group-level data already folds IONOS at 18.9 percent and Strato at 14.5 percent under United Internet, and their W3Techs section already compares groups rather than brands. Folded on their own published figures, Germany's leader stands at 33.4 percent — up among their strongest national moats — and our instrument reads United Internet at about 29 percent of the German active book (panel). Their methodology section already makes this point; all we did was apply it to their whole chart.
The second fold is the one their own methodology note points at: their data measures who answers DNS, not who runs the hosting. Our industry-map instrument draws that boundary as a rule of the dataset. Cloudflare appears in it not as a ranked consolidator but as an infrastructure comparator, filed under the instrument's own words: "CDN/DNS infrastructure present for co-tenancy, NOT a hosting consolidator — excluded from the consolidator ranking". Its control-plane co-tenancy book — nearly 69 million domains, the exact count in the panel — is larger than any consolidator's, and that is precisely why it is not ranked: answering DNS in front of other operators' hosting is a different business from operating hosting.
The same separation runs through the network layer: across the groups we profile, the share of a group's web-provisioned book served from network space the group itself operates runs from nothing at all to almost seven in eight (panel). Who answers, who owns, and who operates are three different questions with three different league tables. The full set of those instrument readings — platform concentration, brand-stack cohesion, own-network serving share, and where each measurement stops — is on the measurements page behind the Hosting Anatomy series.
Their piece closes with a warning that a worldwide league table is "the one view that hides the difference". That sentence is true at every altitude. It is true at brand grain, which is their finding: the global row hides the national moats. It is true at owner grain, which is the fold: the national rows hide the multinational owners. And it is true at operator grain: the DNS-answering rows hide the difference between fronting hosting and running it.
One instrument note, offered in the same spirit as their own W3Techs comparison. In the markets both instruments read, our Cloudflare shares run a couple of points above theirs — they read Cloudflare at 3.1 percent of German resolving domains and 5.8 percent of French ones; our readings on the active books of the same two markets are in the panel. The gap has a mechanical explanation: an active-domains book keeps the built-out, trafficked end of the market, which is exactly where Cloudflare over-indexes, while a resolving-domains count keeps every registration that answers. Their own W3Techs section shows the same direction for the same reason — the more a population is filtered toward active websites, the larger Cloudflare reads. Two instruments, two answers, one consistent mechanism. That is corroboration, not correction.
exact figures re-cut from the served release each build; the argument above is written to survive them.
| Figure | Value | Base | As of |
|---|---|---|---|
| the .no market — Miss Group share of the classified active book | 44.1% | of the .no classified active book — the base is its own row below | 2026-08-23 |
| the .no market — Cloudflare share of the classified active book | 6.7% | of the .no classified active book — the base is its own row below | 2026-08-23 |
| the .no market — classified active-domains book (the base the two shares above divide) | 330,627 | active domains in the market resolving to a named control-plane group (control-plane layer, market-structure instrument) | 2026-08-23 |
| the .de market — United Internet share of the classified active book | 29.3% | of the .de classified active book — the base is its own row below | 2026-08-23 |
| the .de market — Cloudflare share of the classified active book | 5.8% | of the .de classified active book — the base is its own row below | 2026-08-23 |
| the .de market — classified active-domains book (the base the two shares above divide) | 4,578,703 | active domains in the market resolving to a named control-plane group (control-plane layer, market-structure instrument) | 2026-08-23 |
| the .fr market — OVHcloud share of the classified active book | 36.2% | of the .fr classified active book — the base is its own row below | 2026-08-23 |
| the .fr market — Cloudflare share of the classified active book | 6.9% | of the .fr classified active book — the base is its own row below | 2026-08-23 |
| the .fr market — classified active-domains book (the base the two shares above divide) | 1,776,550 | active domains in the market resolving to a named control-plane group (control-plane layer, market-structure instrument) | 2026-08-23 |
| the .no market — hosting-relationship HHI (actor-pure) | 3,252 | the actor-pure .no active book: classified control-plane domains with CDN, SaaS-platform and DNS-infrastructure actors removed and the base renormalised | 2026-08-23 |
| the .no market — the actor-pure book that HHI is computed on | 288,210 | see norway_hr_hhi | 2026-08-23 |
| Cloudflare — control-plane co-tenancy book (listed as an infrastructure comparator, excluded from consolidator rankings by the instrument's rule) | 68,673,544 | all DNS-alive domains on the operator's nameservers (control-plane book) | 2026-08-23 |
| live-business share of the control-plane book — minimum (GoDaddy) | 38% | the group's own control-plane book (56,619,310 domains) | 2026-08-23 |
| live-business share of the control-plane book — maximum (hosting.com (fka World Host Group)) | 95.1% | the group's own control-plane book (558,731 domains) | 2026-08-23 |
| GoDaddy — control-plane book | 56,619,310 | all DNS-alive domains on the group's nameservers | 2026-08-23 |
| GoDaddy — live-business domains in that book | 21,528,470 | live-business state of the same control-plane book (one book, four exclusive states) | 2026-08-23 |
| GoDaddy — live-business share of its control-plane book | 38% | GoDaddy's control-plane book, its own row above | 2026-08-23 |
| GoDaddy — parked domains in that book | 9,152,439 | parked state of the same control-plane book | 2026-08-23 |
| GoDaddy — idle domains in that book (a CEILING on idleness per the instrument's caveat: includes real sites not yet page-observed) | 25,517,805 | idle state of the same control-plane book | 2026-08-23 |
| own-network serving share — minimum (Team Internet) | 0% | of the group's web-provisioned book, the share served from network space the group itself operates | 2026-08-23 |
| own-network serving share — maximum (hosting.com (fka World Host Group)) | 86.7% | of the group's web-provisioned book, the share served from network space the group itself operates | 2026-08-23 |
Data version 2026-08-23 · every OUR-side figure of the WHT-response brief; the external (webhosting.today) figures are attributed on the page and are deliberately NOT in this panel — two instruments, two vintages, never mixed. On the own-network rows: a low share never means the group does not host — cloud-served operations are invisible to network attribution (the instrument's own CRITICAL caveat, quoted in full above). Every figure here is derived from the served release by this brief's own derivation script; nothing on this page is hand-typed.
Every our-side figure above is served by the HostingBrain connector: the market shares by the market-structure tool, the concentration index by the market-concentration tool, and the industry-map readings — the live-business splits, the comparator listing and the own-network range — by the consolidation-landscape tool. Ask it:
"Show the control-plane provider ranking for Norway, Germany and France with each market's denominator."
"Show the consolidation landscape: each group's control-plane book, its live-business split, and why Cloudflare is listed as a comparator rather than ranked."
Group-level profiles — including the ownership folds behind national brands — are part of the briefing product.