Home › Guides › Iran blackout in open data
✓ checked 23 October 2026 · sources: Chrome UX Report on BigQuery (CC BY 4.0), Wikimedia pageviews API (CC0), Freedom House press release (22 Oct 2026) · method · caveats · download the data
When Iran cut its connection to the global internet in 2026, the web that Chrome users in Iran visit shrank by 90%: Google’s Chrome UX Report counted 319,590 websites with enough visitors from Iran in February, and 30,794 in March. Over the same month, Persian Wikipedia’s readers in the United States fell by 69%. That is consistent with many “American” Persian readers being people in Iran using VPN exits abroad, though the data cannot prove it.
Chrome-visible sites, Iran: 319,590 → 30,794Persian Wikipedia, US views: 9.54M → 2.96Mback to 321,008 sites by Sep 2026
The shutdown
Freedom House’s Freedom on the Net 2026 report, released on 22 October 2026, says Iranian authorities imposed a nationwide internet shutdown lasting about 88 days, the longest in the report’s history, and restructured access so that only a privileged few could reach global platforms. It ranks Iran’s decline as the largest of the year. We cite the press release only; we did not use Freedom House’s data, which needs permission for commercial use. The numbers below come from two open datasets that let anyone check the effect month by month.
How many websites Chrome users in Iran reached
The Chrome UX Report (CrUX) publishes, for each country, the list of website origins that had enough Chrome users there in a month. It is a rough count of the web people in a country visibly use. For Iran it usually ran between about 280,000 and 350,000 origins a month from 2024 to February 2026. It fell to 30,794 in March 2026, then recovered to 93,595 in April, 187,537 in May and 293,092 in June, reaching 321,008 in September.
Smaller dips show up earlier: 243,966 in June 2025, down from 306,145 in May, and 231,008 in January 2026, down from 350,163 in December. For comparison, Russia’s count fell more slowly, from 856,567 in January 2024 to 661,624 in September 2026, down 23%.
Persian Wikipedia’s readers abroad
Wikimedia publishes monthly page views by country for each Wikipedia, but leaves out the countries on its Country and Territory Protection List. Iran is on the “not published” part of that list, as are Afghanistan, Russia and China, so Persian Wikipedia’s views from Iran itself never appear. What does appear are the views that Wikimedia places in other countries, and those dropped sharply when Iran went offline.
Views of Persian Wikipedia attributed to listed countries fell from 40.62M in February 2026 to 12.09M in March, down 70%. US views fell from 13.89M in March 2025 to 2.96M in March 2026, down 79% on the year. The Persian-speaking diaspora in the US did not shrink that fast. A simpler explanation is that much of this traffic came from readers in Iran whose VPNs exit in the US, Germany or Canada, and who lost access with the shutdown. In April 2026 the US was ranked first, the only month in this series where no unlisted country ranked above it. By September 2026 listed views were back to 32.38M, against 38.25M a year earlier.
| Month | CrUX origins, Iran | Persian Wikipedia views, listed countries | of which United States |
|---|---|---|---|
| Jan 2025 | 309,166 | 59.23M | 14.96M |
| Feb 2025 | 310,821 | 53.53M | 13.81M |
| Mar 2025 | 287,462 | 53.29M | 13.89M |
| Apr 2025 | 310,084 | 42.85M | 11.06M |
| May 2025 | 306,145 | 39.83M | 11.56M |
| Jun 2025 | 243,966 | 36.19M | 8.97M |
| Jul 2025 | 281,325 | 48.91M | 12.85M |
| Aug 2025 | 293,614 | 42.15M | 10.15M |
| Sep 2025 | 305,862 | 38.25M | 9.15M |
| Oct 2025 | 320,860 | 43.78M | 9.82M |
| Nov 2025 | 349,367 | 43.20M | 11.35M |
| Dec 2025 | 350,163 | 48.90M | 14.09M |
| Jan 2026 | 231,008 | 32.09M | 9.18M |
| Feb 2026 | 319,590 | 40.62M | 9.54M |
| Mar 2026 | 30,794 | 12.09M | 2.96M |
| Apr 2026 | 93,595 | 10.84M | 2.49M |
| May 2026 | 187,537 | 16.44M | 4.21M |
| Jun 2026 | 293,092 | 30.82M | 8.17M |
| Jul 2026 | 298,577 | 34.33M | 10.22M |
| Aug 2026 | 316,571 | 32.71M | 10.69M |
| Sep 2026 | 321,008 | 32.38M | 9.90M |
CrUX: distinct origins in chrome-ux-report.country_ir. Wikipedia: sum of views_ceil across listed countries, all access methods.
The same VPN effect is one reason why proxy and VPN traffic from Iran is so hard to measure directly; see our look at Tor Snowflake requests from Iran and at how Wikipedia counts automated traffic.
Method
CrUX: one BigQuery query counting distinct origins in the chrome-ux-report.country_ir and country_ru monthly tables from January 2024 to September 2026, run on 23 October 2026 (a dry run first; 1.22 GB processed). Wikipedia: the Wikimedia pageviews API top-by-country endpoint for fa.wikipedia, all access methods, monthly from January 2024 to September 2026, fetched on 23 October 2026. The API reports rounded figures (views_ceil) and leaves out countries with 100 views or fewer, so the totals are approximate. Percentage changes are rounded to the nearest point.
Caveats
- CrUX counts an origin only if enough eligible Chrome users visited it from that country, so the count falls when traffic falls, even if no site was blocked. It measures the Chrome-visible web, not blocking.
- Both datasets place users by where their connection appears to come from. A VPN user in Iran counts as being in the VPN’s exit country. That is the basis for the reading of the Wikipedia data, and it is an inference, not a measurement.
- Persian is also read in Afghanistan, which is also unpublished, and diaspora readers browse more in some months than others.
- Freedom House’s findings are cited as reported, not checked by us.
Download the data
- Monthly CrUX origins for Iran and Russia, and Persian Wikipedia views by listed country total and US (CSV), January 2024 to September 2026.
Cite as: ProxyPicker, “Iran’s 2026 Blackout in Open Data: Chrome and Wikipedia”, 23 October 2026, https://proxypicker.com/iran-blackout-chrome-wikipedia/. CrUX data © Google, CC BY 4.0. Wikimedia data CC0.
Press & data requests: email [email protected] for interviews, the raw data behind this page, or corrections. See also our contact page.