Web Hosting Statistics Updated November 2026

Human Written & Fact Checked · Updated

100 Web Hosting Statistics for 2026

You cannot measure the web-hosting industry by counting cloud revenue alone. Start with the question: who hosts a website, which server responds, how many names are registered, or how quickly content reaches the browser? Each requires a different denominator.

This hub keeps those measurements separate. Provider disclosures describe businesses with several products; independent surveys describe detected technologies; performance datasets describe requests and user experiences. Publication dates accompany every observation, while each section identifies the measurement period.

Headline figures

  • Global sampled websites: Shopify: 5.4% website usage. (W3Techs / Q-Success, Nov 2026)
  • Global sampled websites: Hostinger: 5.2% website usage. (W3Techs / Q-Success, Nov 2026)
  • Global sampled websites: Wix: 4.2% website usage. (W3Techs / Q-Success, Nov 2026)
  • Global survey responses: Surveyed sites: 1,470,435,351. (Netcraft, Sep 2026)
  • Global survey responses: Cloudflare busiest- million share: 29.10%. (Netcraft, Sep 2026)
  • Global domain registrations: Total domain registrations: 411.6 million. (DNIB / Verisign, Oct 2026)
  • Global domain registrations: .com registrations: 169.8 million. (DNIB / Verisign, Oct 2026)
  • GoDaddy global business: GoDaddy customers: 20.5 million. (GoDaddy, Oct 2026)
  • IONOS global business: IONOS customers: 6.91 million. (IONOS Group, Aug 2026)
  • Eligible sampled web pages/sites: Desktop good TTFB: 55%. (HTTP Archive / Google CrUX, Jan 2026)
  • Eligible sampled web pages/sites: Mobile good TTFB: 44%. (HTTP Archive / Google CrUX, Jan 2026)
  • HTTP Archive sampled websites and requests: Mobile HTML requests served through CDNs: 35%. (HTTP Archive, Jan 2026)

Browse web hosting statistics

Which companies host the websites that W3Techs detects?

Provider usage measures a technical relationship, not the value of a customer account. Hosted site builders belong in this dataset alongside infrastructure companies, so its ranking answers a different question from a comparison of hosting-only revenues.

Which larger providers appear in the website-usage snapshot?

Data period: 9 November 2026. These are worldwide website-usage observations from the dated W3Techs snapshot. They identify providers linked to sampled websites; they do not count paid subscriptions. A site-building platform can therefore appear alongside a general infrastructure provider. Treat the denominator consistently when comparing companies.

How widely used are specialist and regional hosting groups?

Data period: 9 November 2026. These entries retain W3Techs' company-group labels. A group can contain several brands, and a brand's commercial positioning does not change the unit being counted. The percentages describe observed website usage, instead of market value, server capacity or the number of people visiting each provider's customers.

What does the smaller-provider part of the usage table show?

Data period: 9 November 2026. A small observed percentage does not indicate that a service is unsuitable for a particular workload. This table describes adoption in a measurement population. Evaluating a provider for a project would require separate evidence about requirements, contracts and operations; none of those characteristics can be inferred from usage share alone.

How large is the measured web and which servers answer requests?

The visible server layer can be different from the company hosting the underlying application. Compare server observations only within the same denominator: sites, active sites, popular sites, computers and domains answer different questions.

What do total responses and all-site server shares show?

  • Surveyed domains: 323,147,961. (Netcraft, Sep 2026)
  • Web-facing computers: 15,463,120. (Netcraft, Sep 2026)
  • Cloudflare all-site share: 18.10%. (Netcraft, Sep 2026)
  • nginx all-site share: 15.33%. (Netcraft, Sep 2026)
  • Apache all-site share: 11.56%. (Netcraft, Sep 2026)
  • OpenResty all-site share: 9.02%. (Netcraft, Sep 2026)

Data period: September 2026. Netcraft's survey counts several units separately. A domain can expose multiple sites, and a computer can serve multiple domains. These figures describe responses and detected software in the survey. They are not interchangeable measures of hosting accounts or distinct organisations purchasing hosting services. The all-site inventory can include pages with little independent content. Its size is useful for understanding the measured surface, but should not be read as the number of active publishing businesses or separately purchased hosting plans.

How do active sites and the busiest sites differ?

  • Cloudflare active-site share: 25.57%. (Netcraft, Sep 2026)
  • nginx active-site share: 15.93%. (Netcraft, Sep 2026)
  • Apache active-site share: 12.14%. (Netcraft, Sep 2026)
  • LiteSpeed active-site share: 5.70%. (Netcraft, Sep 2026)
  • nginx busiest- million share: 19.65%. (Netcraft, Sep 2026)
  • Apache busiest- million share: 15.01%. (Netcraft, Sep 2026)
  • LiteSpeed busiest- million share: 3.66%. (Netcraft, Sep 2026)

Data period: September 2026. The active-site and busiest-site tables use narrower populations than the all-site table. Their shares should be read within those populations, not added together. The presence of an intermediary can also influence the server identified by an external request, making these observations different from a catalogue of origin-hosting suppliers. Popularity and activity filters change the population before the shares are calculated; neither table is a revenue ranking.

Which server technologies appear across computers and domains?

  • nginx computer share: 44.00%. (Netcraft, Sep 2026)
  • Apache computer share: 17.70%. (Netcraft, Sep 2026)
  • Microsoft computer share: 5.41%. (Netcraft, Sep 2026)
  • OpenResty domain share: 17.91%. (Netcraft, Sep 2026)
  • nginx domain share: 17.44%. (Netcraft, Sep 2026)
  • Cloudflare domain share: 14.25%. (Netcraft, Sep 2026)
  • Apache domain share: 13.94%. (Netcraft, Sep 2026)

Data period: September 2026. Computer shares and domain shares place different weight on infrastructure. A single machine can serve many domains, while a service can distribute traffic across machines. Keeping the denominator visible avoids turning a technical deployment table into an unsupported ranking of hosting revenue or customer demand. Externally visible software can also reflect a proxy in front of the application. Identifying that proxy does not reveal the origin supplier.

What do domain registrations reveal about demand for web infrastructure?

A registration can support a website, email or a reserved name. Domain totals provide infrastructure context, but they cannot be converted directly into a count of hosted websites or paying web-hosting customers.

How did the domain base and the established extensions change?

Data period: Q3 2026. DNIB combines registry information and its domain-industry analysis. Registration totals are stocks at the reporting date, while growth rates describe changes over different intervals. A quarterly comparison and an annual comparison therefore answer separate questions. The figures do not establish whether the registered names resolve to active websites.

How do new registrations, country codes and renewals compare?

Data period: Q3 2026; Q3 2025. New registrations are flows during a quarter; the installed base is a stock. Renewal percentages describe another part of the domain lifecycle and may be preliminary. Country-code and generic-extension categories should remain distinct. DNIB credits registry and zone data alongside ZookNIC inputs; these are domain-industry observations, not a global hosting-revenue estimate.

How often do websites meet loading-performance thresholds?

Hosting is only one part of a page's loading path. Separate the arrival of the first byte from the appearance of visible content before attributing a slow page to a server.

How do desktop and mobile time-to-first-byte results compare?

Data period: 2025 dataset. Time to First Byte includes connection and network overhead as well as server processing. The device-specific distributions describe eligible observations in the dataset, not a controlled comparison of hosting suppliers. A good result is a category in the report instead of a guarantee for every visit. Rounded categories may not sum exactly to the whole, and differences between device groups should not be interpreted as the causal effect of the hosting service.

How do first-contentful-paint results differ by device?

Data period: 2025 dataset. First Contentful Paint measures when visible content first appears. Its timing includes work after the server has started responding, so it cannot isolate hosting performance. Desktop and mobile observations represent different usage conditions. The measured result belongs to the complete delivery and rendering path, including resources and browser work. A provider ranking would require a matched workload and controlled comparison that this population distribution does not supply. Correlation is insufficient.

How do CDNs participate in delivering hosted content?

An edge service can deliver content without being its origin host. Keep request-level CDN adoption separate from website hosting shares and from laboratory timing measurements.

Which resources and popular sites use CDNs?

  • Mobile subdomain requests served through CDNs: 52%. (HTTP Archive, Jan 2026)
  • Mobile third-party requests served through CDNs: 71%. (HTTP Archive, Jan 2026)
  • Top 1,000 mobile sites using CDNs: 71%. (HTTP Archive, Jan 2026)

Data period: 2025 dataset. A page can obtain its document, first-party assets and third-party resources through different delivery paths. Request-level percentages consequently do not describe a single hosting decision. The popularity bands use a site population instead of a resource population, so those shares must also remain separate. Detection depends on visible DNS and response signals. A CDN can be present alongside another origin host, and a crawl can miss details hidden by a provider's routing or a site's use of several networks. Categories remain distinct.

What did popular-site adoption and mobile DNS measurements show?

  • Top 10,000 mobile sites using CDNs: 70%. (HTTP Archive, Jan 2026)
  • Median mobile CDN DNS resolution: 176ms. (HTTP Archive, Jan 2026)
  • Median mobile origin DNS resolution: 217ms. (HTTP Archive, Jan 2026)

Data period: 2025 dataset. These are simulated-browser medians from the archive's test environment. They are not promises about the speed of a future visitor's connection. DNS timing depends on resolver behaviour, caching, network conditions and the route to the answering service. The dataset compares observed delivery categories, instead of assigning identical workloads randomly to providers. A difference between medians therefore cannot establish the causal benefit a particular website would obtain by purchasing a CDN. The origin and edge categories also remain separate from company-level hosting market shares.

Glossary

CDN
Content delivery network, an intermediary delivery layer.
TTFB
Time to First Byte.
FCP
First Contentful Paint.
TLD
Top-level domain; ccTLD denotes a country-code extension.
EBITDA
Earnings before interest, tax, depreciation and amortisation; adjusted definitions vary.

Methodology and sources

Every figure on this page comes from the organisation that produced it: regulators, statistics offices, public bodies, research institutes, and companies reporting their own data or surveys. Each one links to the original report or dataset and carries its release date. Survey results keep their population and fieldwork period, and financial figures keep the measure the producer used. Figures from different producers are not added together or ranked against each other. The page is reviewed monthly and updated when a producer releases newer data.