The report analyzed a number of major intentional internet shutdowns and social media blockages for the year 2025. The authors calculated that 212 serious cases of restrictions occurred in 28 countries, with the total duration of communication disruptions exceeding 120,000 hours, which is 70% more compared to 2024. Approximately 798 million people were affected by these restrictions.
It is important to note that this is not just about complete internet shutdowns. The report highlights complete outages, blockages of individual platforms, and significant traffic slowdowns to levels close to 2G, where access to voice calls and SMS remains, but the normal operation of modern services becomes impossible. In 2025, there were 55,700 hours of complete outages, 54,000 hours of social media blockages, and 12,700 hours of slowdowns. The latter category accounted for about 10% of all hours of intentional disruptions, and according to the authors, it was actively used for censorship again.
According to the report, the largest economic losses were incurred by Russia: $11.9 billion. Venezuela follows with $1.91 billion and Myanmar with $1.89 billion. The authors describe in detail the waves of restrictions in Russia that began in May, characterized not just by complete shutdowns but by a series of technical interventions. One of the most notable methods mentioned in the report is called the "16 KB curtain": access to a number of resources, including those hosted by Cloudflare, was restricted so that only the first 16 kilobytes of data were loaded. Although the connection was maintained, most modern websites became practically unusable. The authors acknowledge that such measures are difficult to quantify, so the damage from the "curtain" remains approximate.
When considering blockages by platform, the most time in 2025 was attributed to X (Twitter) — 18,354 hours. Telegram takes second place with 16,990 hours, and TikTok is third with 14,646 hours. The text provides examples of how prolonged blockages in certain countries "accumulate" hours, as well as how authorities sometimes choose not to impose immediate bans but rather a gradual deterioration of service quality. For instance, in Russia, certain functions of WhatsApp, including calls, were initially restricted, and by the end of the year, the service became practically non-functional for many users.
The authors note an interesting trend in the motives for shutdowns in 2025. For the first time since observations began, the main reason became restrictions related to "information control," which turned out to be both the most costly and the longest-lasting. At the same time, exam-related shutdowns, although occurring frequently, had smaller financial consequences as they usually lasted for a short time. Some countries began to act more selectively and gently, reducing collateral damage; however, the overall trend of the year is described by the authors as a deterioration of "internet freedom" and an increase in the share of preventive restrictions.
The counting methodology is based on the NetBlocks COST tool and includes data from the World Bank, ITU, Eurostat, and the U.S. Census Bureau. The authors emphasize an important rule: social media blockages lasting more than 365 days are not counted in the report as "permanent censorship," since users gradually migrate to alternative platforms, and the economic effect changes. Therefore, old long-standing bans in Russia do not add new "hours" to the annual statistics, even if they remain in force.
The main conclusion of the report is that internet shutdowns are becoming not only more frequent but also more technological. Instead of crude shutdowns, there is an increasing situation where the formal availability of communication conceals the real unavailability of services. For businesses and users, this means that the risks of digital failures are turning from rare emergencies into a factor that must be prepared for in advance.
