Short answer. We pulled Russia's official banned-sites registry (Roskomnadzor's EAIS) plus the extended operator blocklists and analyzed them. The official registry holds 131,361 domains; extended operator lists exceed 1.47 million entries. The headline: almost 70% of blocked domains are international (.com and other zones), not .ru. One in five of the world's top sites (19%) is unreachable in Russia — Meta, X, LinkedIn, BBC, Medium, WordPress.com and more.
How we counted
Methodology (fully open, see footer): we aggregated the same sources public block-checkers use — the public export of the official EAIS registry (registered domains) and extended blocklists covering subdomains and operator lists. Then we overlaid a sample of the world's top sites. Every figure is reproducible — the engine sits behind our RKN block checker. Snapshot date: June 2026.
Scale of the registry
| Source | Domains | What it is |
|---|---|---|
| Official EAIS registry | 131,361 | official registry, registered domains |
| Extended blocklists | 1,470,647 | + subdomains, operator lists, IP blocks |
131k domains in the official registry — not counting the IPs, subdomains and network ranges that push the real number of unreachable resources into the millions.
Headline: it's not "the Russian web" being blocked — it's the global one
The registry's TLD distribution breaks the "they block Russian sites" stereotype:
| Zone | Domains | Share |
|---|---|---|
| .com | 35,077 | 27% |
| .ru | 39,286 | 30% |
| .net | 5,150 | 4% |
| .top | 4,525 | 3% |
| .org | 4,038 | 3% |
| .online | 3,765 | 3% |
| other (550+ zones) | ~39,500 | 30% |
Only 30% of blocked domains are in.ru. The other 70% are international —.com,.net,.organd 550+ more TLDs. The RKN registry is first and foremost a filter on the global web, not the Russian segment.
One in five of the world's top sites is blocked
We sampled the most-visited sites worldwide and checked them against the registry. The result — 19% (13 of 67) are unreachable in Russia:
- Social: Facebook, Instagram, X (Twitter), LinkedIn, Tumblr, Bluesky
- Media: BBC, Medium, SoundCloud, Dailymotion
- Platforms: WordPress.com, Patreon
Many of these still "resolve" — DNS returns an IP, the host pings — but traffic is cut at the operator's DPI layer. So a plain ping or nslookup won't reveal the block: you need a registry check.
Why "the site works for me but my users can't reach it"
A classic case: the owner sees the site up, while some users report it down. Causes:
- IP / subnet block. RKN blocks an IP, but a CDN (Cloudflare, Akamai) hosts thousands of unrelated sites on it — one "neighbour" lands in the registry, everyone on that IP suffers.
- Domain-level DPI block — the operator cuts by SNI/Host while DNS stays "clean".
- Regional variation — operators roll out blocks unevenly.
# Official check for a domain or IP — the Roskomnadzor unified registry form:
# https://eais.rkn.gov.ru
# See whether the site shares its IP with others (block-by-neighbour risk)
dig +short example.com # → check the resulting IP in the same registry formThe other side of the measurement: 4,255 domains owners checked themselves
Everything above describes the registry itself — what is in the list. There is a second source, collected in a completely different way: our blocking checker. Between 24 March and 26 August 2026 it ran 9,031 checks across 4,255 domains (IP addresses excluded, repeat checks of the same domain collapsed).
One caveat first, without which the numbers below mislead: this is not a sample of the Russian web. These are domains somebody was worried enough to check, usually because they already suspected a problem. The sample is skewed toward trouble by construction. Comparisons between zones remain valid, because the same skew presses on all of them equally.
The result: one in five checked domains is blocked — 846 of 4,255, or 19.9%.
A .ru domain is blocked half as often as a .com
By composition, .ru is the largest zone in the registry at 30% of entries — which invites the conclusion that Russian domains are most at risk. But composition and probability are different quantities: roughly five million .ru domains exist, so a share of the list says nothing about the odds for any single address.
An owner needs the conditional question instead: "I am checking my domain — what are the chances?" That flips the picture:
| Zone | Domains checked | Blocked | Rate |
|---|---|---|---|
| .io | 66 | 19 | 28.8% |
| .org | 208 | 54 | 26.0% |
| .com | 1,258 | 307 | 24.4% |
| .net | 188 | 45 | 23.9% |
| .site | 102 | 24 | 23.5% |
| .ru | 1,064 | 140 | 13.2% |
| .top | 302 | 39 | 12.9% |
| .online | 77 | 9 | 11.7% |
| .casino | 50 | 2 | 4.0% |
Only zones with at least 50 domains are listed; below that a single entry moves the rate by whole percentage points. Smaller zones follow the same shape: .cc 30.0% (30 domains), .info 25.0% (32), .me 24.1% (29), .su 23.5% (34) — and at the other end .xyz 2.6% (39), .space 3.1% (32).
The spread falls into three groups with sharp boundaries:
- 24–29% — legacy gTLDs and repurposed country codes: .io, .org, .com, .net, .site, .info, .me. This is where the foreign services most blocking decisions are aimed at actually live.
- 13% — .ru, half the .com rate. A Russian domain answers to a Russian registrar: a disputed resource is more often delegated away or made to remove the material, and never reaches the registry at all.
- 3–13% — cheap new gTLDs: .top, .online, .casino, .xyz, .space.
The third group is the surprise. Zones with a spam reputation are blocked least often, and .casino — whose name states an activity restricted in Russia outright — sits at 4.0%, six times below .com. The explanation consistent with both halves of the data: the registry is filled by decisions against specific resources, not by the suspiciousness of a zone. A cheap domain lives a few months and dies on its own, well before any legal procedure concludes; what gets blocked is what is visible enough and durable enough to be worth the paperwork.
The practical takeaway: the zone itself neither protects you nor sinks you. It only correlates with the kind of projects that live in it.
Blocks are lifted almost as often as they are imposed
641 domains in the sample were checked more than once — which people usually do while waiting for something to change. Over five months the verdict changed for 44 of them, and the two directions came out nearly even:
| What happened | Domains |
|---|---|
| Clean → blocked | 20 |
| Blocked → lifted | 22 |
| Lifted, then blocked again | 2 |
Two conclusions. First: a block is not a life sentence. Among owners who kept watching, slightly more saw a block lifted than saw a new one imposed — consistent with the procedure, which removes an entry once its grounds are gone: content taken down, decision reversed, domain sold.
Second, less comfortable: 44 changes across 641 domains is 7% over five months. In the overwhelming majority of cases the status does not move for months. If your domain is in the registry and you do nothing, waiting for it to resolve itself is statistically pointless — see what a site owner should do.
What these numbers do not say
Publishing a measurement without the limits of its applicability means passing a sample off as reality, so here they are plainly.
- This is not the Russian web. The 19.9% applies to domains that were brought in to be checked. It cannot be extrapolated to "one in five sites in Russia": for a random domain the odds are far lower.
- Zones are comparable to each other; the absolute rates are not. The skew is identical across zones, so the 13.2%-to-24.4% ratio is meaningful while the levels themselves are inflated.
- Small zones are noisy. At 28–39 domains a single entry shifts the rate by 2.5–3.5 points, which is why they sit outside the main table.
- The window from 25 May to 28 June 2026 is excluded. During it our own registry source broke, and for three weeks the tool answered "not found in the registry" for practically every domain: across weeks 23–25, 599 checks returned a 0.0% block rate against roughly 11% on either side. The source was repaired on 26 June. Leaving that period in would have published a broken instrument as a fact about the Russian web.
That last point is worth dwelling on. The failure did not look like an error — it looked like an ordinary answer, confident and unqualified, which is exactly why it survived three weeks. Since this analysis the tool checks the age of its own data: if the registry has not refreshed for more than three days it no longer answers "not blocked" but reports honestly that the check could not be completed. A negative answer now has to rest on fresh data; a positive one remains valid on older data.
A note on framing. A resource's status in the registry is public information, and an owner is entitled to check and report it (Federal Law 149-FZ). Everything here concerns diagnosing your own site and the lawful route to having an entry removed, on the owner's application once its grounds are gone. Circumventing blocks, and the tools for doing so, are outside its scope; the official check and the removal procedure both live at eais.rkn.gov.ru.
What a site owner should do
- Check the domain and IP with the RKN block checker — registry status and network signals in 10 seconds. The authoritative answer comes from the Roskomnadzor unified registry form at eais.rkn.gov.ru, which also names the grounds for the entry and the authority that ordered it.
- If the site is on a shared CDN IP, that's "block-by-neighbour" risk; consider a dedicated IP.
- Put the domain on availability monitoring so you learn about a registry hit first, not from your customers.
Where does the data come from?
The primary source is the Roskomnadzor unified registry (eais.rkn.gov.ru), which also gives the official answer for any single domain. For the bulk counting its form cannot do, we use the public export of that same registry and extended blocklists — the same sources independent block-checkers use. We publish aggregate statistics only, not the lists themselves.
Why 131k in the registry but "millions of blocks"?
131k is the registered-domain count of the official registry. With subdomains, IPs and network ranges (extended lists exceed 1.47M entries) the real number of unreachable URLs is an order of magnitude higher.
The site "pings" but users can't open it — is it blocked?
Possibly. A domain/SNI DPI block doesn't affect DNS or ping — the host "resolves" but traffic is cut. Check against the registry, not with ping.
Can I check my own site?
Yes — the RKN block checker uses the exact same sources as this study.
Methodology. The TLD distribution is computed over the official registry. The "world's top" overlay is a 67-site sample of the most-visited sites by category. Date: June 2026. Reproducible via our RKN checker.
