The hosting industry isn't a cloud-migration story. It's a relationship-control story.

HostingBrain analyst brief · July 2026 · data snapshot 2026-07-06 · aggregate-level

The dominant narrative about hosting is a migration story: everything is moving to hyperscalers, workloads flow freely between clouds, and infrastructure is the battleground. Three separate HostingBrain measurements — where new sites land, how customers actually move, and how owners spread their domains — point somewhere else. Read together, they say the contest was never really about infrastructure. It's about who holds the customer relationship.

One

Traditional hosting still captures most new European sites

If cloud were eating the market, the newest domains would show it first. In the organic European cohort (bulk events screened out), the new arrivals look almost exactly like the installed base — traditional visible-origin hosting holds its ~78% share, and hyperscalers actually under-index among new sites. The displacement story is real in developer mindshare; it is not where European business domains resolve.

new cohort vs installed base by hosting class — Europe, bulk-event screened (%) Traditional79.0 78.4 CDN-fronted12.7 12.2 Hyperscaler8.3 9.3 new cohortinstalled base

Full analysis: Installed infrastructure reflects the past. New websites reveal momentum.

Two

The relationship is stickier than the server under it

People talk as if workloads slide frictionlessly between clouds, CDNs and networks. On the mass-market book they don't. When we decompose a year of movement, one event dominates — the customer changing provider — while the thing the migration narrative obsesses over, the origin server moving while the provider relationship stays put, is essentially a rounding error: 0.08% a year. The provider relationship, not the infrastructure, is the atomic unit that moves.

annualized movement on the confirmed-active book, decomposed — Europe (%) operator switch12.18 IP housekeeping3.15 abandonment1.88 activation0.49 aftermarket rotation0.21 serving move, control-plane stable0.08

Full analysis: How sticky is hosting, really?

Three

Once a customer is captured, they consolidate

The intuition is that experienced owners diversify vendors. The opposite holds. Among owners of two or more live domains, most keep everything in one place — and they keep it with a single brand, not merely a single corporate parent. The relationship, once won, tends to take the whole wallet.

74%of multi-domain owners use a single provider
89%average share of the wallet on the primary provider
99%of loyal owners stay within one brand — chosen, not accidental

Full analysis: How faithful is a hosting customer?

The prize

The valuable question in hosting isn't who has the best infrastructure. It's who owns the control-plane relationship — before SaaS, email, security and AI tools attach to the same customer wallet.

Put the three together. New customers still arrive at traditional hosting relationships. Those relationships almost never come unbundled from their infrastructure — if a customer moves, they move the relationship whole. And once a customer is in, they consolidate the rest of their wallet onto that same relationship. The control plane is the point of leverage: own it early and the attach revenue — workspace email, e-commerce, security, the next wave of AI tooling — has somewhere to land. That is why consolidators buy nameserver books, not data centres.

Method & caveats. Three independent measurements on the confirmed-active business book (verified operating email, live content or platform provisioning; parking and dead pages excluded). Momentum is discovery-paced (first-observed), bulk-event screened. Switching is decomposed and normalised (cloud NS-shard rotation is not counted as a move), annualised from a 26-week window, and is a floor observed at rescan. Faithfulness is measured on tracking-detected ownership (host-independent; certificate-detected owners are deliberately excluded because a shared certificate would overstate loyalty). All domain-level, not revenue-level; aggregate, no named providers.

Reproduce this — or ask it yourself

This brief is a synthesis of three tools. In Claude or any MCP-compatible assistant, one prompt walks all three and reaches the same thesis:

Prompt · paste into an MCP client with HostingBrain connected

“Using HostingBrain, give me three non-obvious insights about the European hosting market by combining: (1) new-site momentum by hosting class, bulk-event screened; (2) the decomposed, annualised switching rates by layer; and (3) how faithful multi-domain owners are to one provider and one brand. Then tell me what the three together imply about where the real leverage in hosting sits.”

Resolves to hosting_momentum, layer_stickiness and customer_faithfulness — all free tier. Named provider detail is Pro/Analyst.

Ask the strategic question, not just the metric. HostingBrain answers inside Claude and any MCP-compatible assistant, with the denominator and caveats attached to every number.

Get free access