a evaluation of Q1 2026 Web disruptions

Within the first quarter of 2026, government-directed shutdowns figured prominently, with extended Web blackouts in each Uganda and Iran, a stark distinction to the dearth of noticed government-directed shutdowns in the identical quarter a yr prior. This quarter, we additionally noticed various Web disruptions attributable to energy outages, together with three separate collapses of Cuba’s nationwide electrical grid. Navy motion continued to disrupt connectivity in Ukraine and in addition impacted hyperscaler cloud infrastructure within the Center East. Extreme climate knocked out Web connectivity in Portugal, whereas cable injury disrupted connectivity within the Republic of Congo. A technical downside hit Verizon Wi-fi in america, and unknown points briefly disrupted connectivity for purchasers of suppliers in Guinea and the UK.

This put up is meant as a abstract overview of noticed and confirmed disruptions and isn’t an exhaustive or full checklist of points which have occurred throughout the quarter. A bigger checklist of detected visitors anomalies is accessible within the Cloudflare Radar Outage Heart. Notice that each bytes-based and request-based visitors graphs are used inside this put up for example the impression of the noticed disruptions, with the selection of metric typically made based mostly on which higher illustrates the impression of the disruption.

Upfront of the January 15 presidential election, Ugandan authorities ordered a nationwide Web shutdown. The Uganda Communications Fee (UCC) instructed cellular community operators to droop public Web entry, efficient 18:00 native time (15:00 UTC) on January 13. The UCC reportedly defended the shutdown as essential to “curb misinformation, disinformation, electoral fraud and related risks.” Home visitors on the Uganda Web Alternate Level (UIXP) dropped from roughly 72 Gbps to 1 Gbps because of the motion taken.

Equally, Cloudflare information exhibits a near-complete lack of visitors from Uganda coincident with the beginning of the shutdown, with visitors remaining successfully at zero via 23:00 native time (20:00 UTC) on January 17, when Web connectivity was partially restored after incumbent President Yoweri Museveni was declared winner of his seventh time period. 

Full Web restoration was introduced by the UCC on January 26, with cellular community operators MTN Uganda and Airtel Uganda each confirming on social media that restrictions had been lifted. The shutdown prompted lawsuits in opposition to UCC and the telecoms corporations and drew criticism from digital rights organizations together with CIPESA.

Uganda additionally blocked Web entry throughout its 2021 election. Authorities had repeatedly promised this time could be completely different, stating as lately as January 5 that “claims suggesting otherwise are false, misleading.”

Iranian residents spent a big a part of Q1 2026 offline, or with severely restricted connectivity, because of two nationwide Web shutdowns. The primary started round 20:00 native time (16:30 UTC) on January 8, and we explored the impression seen over the primary few days in our What we find out about Iran’s Web shutdown weblog put up. Visitors from Iran remained close to zero till January 21, when a small quantity of visitors returned, solely to vanish a bit of over 24 hours later. The same temporary restoration additionally occurred on January 25, earlier than visitors recovered extra aggressively beginning on January 27.

A near-complete lack of introduced IPv6 handle house began a number of hours earlier than the drop in visitors befell on January 8. Asiatech (AS43754) was by far the only largest contributor, dropping 4.46 million /48-equivalents, accounting for ~9.4% of Iran’s total IPv6 house loss by itself. RASANA (AS31549) was the second-largest, dropping 4.19 million /48-equivalents (~8.8% of the nation whole). As could be anticipated, this resulted within the share of IPv6 visitors in Iran going to zero. Given the hole in timing between this variation and the lack of visitors throughout the nation, this may occasionally have been a number one indicator of what was about to occur, however probably not a direct reason behind it. Some nominal shifts in introduced IPv4 handle house are seen throughout the shutdown, however ranges remained pretty constant throughout the shutdown interval. These observations counsel that the shutdown was applied by different means, resembling filtering.

Cloudflare Radar social media posts (X, Bluesky, Mastodon) all through January and into early February documented our observations in regards to the state of connectivity in Iran over the course of that month.

On February 28, as navy strikes on Iran escalated, a second nationwide Web shutdown started. Cloudflare Radar noticed a pointy drop in visitors from Iran starting round 10:30 native time (07:00 UTC). Visitors ranges fell to nicely beneath 1% of earlier ranges, with solely small quantities of Internet and DNS visitors egressing the nation.

No vital shifts in introduced IP handle house had been noticed across the onset of this shutdown. IPv4 house remained pretty constant, and IPv6 house remained persistently unstable, suggesting that route withdrawals weren’t the reason for this second shutdown.

The continued announcement of IP handle house, and the presence of visitors from the nation, even when only a small quantity, helps studies that the shutdown was successfully achieved via aggressive filtering, with so-called “whitelists” and “white SIM cards” proscribing entry to solely accepted Web websites by chosen customers.

Iran remained successfully offline via the top of the quarter. As of late April, this shutdown stays largely in place, making it one of many longest sustained Web disruptions noticed lately.

On March 15, because the Republic of Congo held a presidential election anticipated to increase President Denis Sassou Nguesso’s 42-year rule, a near-complete shutdown of Web connectivity was noticed within the nation. Visitors from the nation dropped precipitously round 06:30 native time (05:30 UTC), falling to close zero for about 60 hours via the election interval and its fast aftermath. Visitors started recovering round March 17 at 18:20 native time (17:20 UTC), quickly returning to pre-shutdown ranges. Whereas Congolese authorities supplied no official clarification for the drop in visitors, related shutdowns had been put into place throughout the 2021 and 2016 elections.

On January 7-8, Russian assaults on vitality infrastructure in Ukraine triggered energy outages that disrupted Web connectivity in Dnipropetrovsk and surrounding areas. Cloudflare Radar noticed a major drop in visitors from the area, reaching practically 50% under the prior week’s ranges, beginning round 22:45 native time (20:45 UTC) on January 7. Restoration started roughly 06:00 native time (04:00 UTC) on January 8.

On January 26, Russia launched a drone and missile assault concentrating on vitality infrastructure in Kharkiv. Cloudflare Radar noticed an roughly 50% drop in visitors from the area starting round 19:15 native time (17:15 UTC). Restoration progressed via January 27 as energy was step by step restored.

One of the uncommon disruptions of the quarter was the bodily injury inflicted on Amazon Internet Companies information facilities within the Center East by drone strikes tied to the continued regional battle. On the morning of March 1 (UTC), Amazon reported a fireplace began after objects hit a UAE information heart. The next day, the corporate confirmed that two of its services within the United Arab Emirates (me-central-1 area) had been “directly struck” by drones and {that a} facility in Bahrain (me-south-1 area) was additionally taken offline after being broken by a close-by strike.

Cloudflare’s Cloud Observatory information confirmed elevated connection failure charges for the me-central-1 and me-south-1 areas starting March 1-2 and remaining increased for a number of days. Connection failures happen when Cloudflare fails to efficiently hook up with an origin server when trying to retrieve uncacheable content material, or content material not in/expired from cache. These graphs illustrate the elevated fee of failures skilled when trying to connect with servers in these impacted areas.

In a standing put up on the AWS Well being Dashboard, Amazon acknowledged: “These strikes have caused structural damage, disrupted power delivery to our infrastructure, and in some cases required fire suppression activities that resulted in additional water damage.” The corporate warned that instability was more likely to proceed within the Center East, making operations “unpredictable,” and urged clients with workloads within the affected areas to again up their information or migrate to different AWS areas.

The AWS me-south-1 area in Bahrain suffered an extra disruption on March 23, following additional drone exercise.

On January 15, a energy outage struck Buenos Aires throughout a summer time warmth wave. The outage triggered nominal disruptions in Web connectivity for purchasers of a number of suppliers within the Buenos Aires space, together with Telecom Argentina (AS7303), Telecentro (AS27747), and IPLAN (AS16814), with visitors from these networks dropping between 17:30 and 19:30 native time (20:30 – 22:30 UTC). Visitors returned to anticipated ranges roughly two hours after the outage started.

An emergency energy minimize on Ukraine’s electrical energy grid on January 31 triggered widespread energy outages affecting Moldova and a number of other Ukrainian areas together with Kyiv and Kharkiv. Moldova was reportedly hit by widespread energy cuts amid the Ukrainian grid issues, and the Ukrainian Vitality Minister defined the cross-border impression, noting “Today at 10:42 a.m. (08:42 GMT), a technical malfunction occurred, causing a simultaneous shutdown of the 400 kilovolt line between the power grids of Romania and Moldova and the 750 kilovolt line between western and central Ukraine.” Visitors from Moldova, Kyiv, and Kharkiv started falling round 10:42 native time (08:42 UTC), reaching as a lot as 46% under the prior week, with restoration occurring round 14:00 native time (12:00 UTC).

On February 18, widespread energy outages struck Paraguay after key transmission traces went out of service. The Nationwide Electrical energy Administration (ANDE) posted a sequence of updates on X documenting the incident and efforts to revive energy. Web visitors from Paraguay dropped as a lot as 72% in comparison with the prior week starting round 15:15 native time (18:15 UTC), and the disruption lasted practically three hours, with restoration occurring by roughly 18:30 native time (21:30 UTC).

A main failure within the Interconnected Nationwide Electrical System (SENI) of the Dominican Republic triggered a widespread energy outage on February 23. The state-owned electrical firm Empresa de Transmisión Eléctrica Dominicana (ETED) posted updates on X documenting the failure and the restoration effort. Web visitors from the nation dropped sharply starting round 10:50 native time (14:50 UTC), and recovered round midnight native time (04:00 UTC) on February 24, in keeping with a affirmation posted by ETED that “The authorities of the electric sector reported that the Interconnected National Electric System (SENI) was fully restored to 100% at 11:53 p.m. on this Monday…”.

Cuba skilled three separate collapses of its Nationwide Electrical System (SEN) throughout March, every inflicting widespread Web disruption, reflecting the extreme deterioration of the nation’s electrical infrastructure. (Energy outages additionally disrupted Web connectivity in Cuba throughout September and March 2025, and October 2024.) 

The primary collapse occurred on March 4, when a disconnection of Cuba’s Nationwide Electroenergy System cascaded from Camagüey to Pinar del Río, reducing energy to the western half of the island, together with Havana. OSDE/UNE (Cuba’s Electrical Union) confirmed the failure on social media. Cloudflare Radar information confirmed visitors from the island dropping by practically half starting round 12:15 native time (17:15 UTC), with visitors recovering by roughly 05:01 native time (10:01 UTC) on March 5.

The second collapse occurred on March 16, when Cuba’s total Nationwide Electrical Energy System was disconnected. EnergíaMinas Cuba posted updates on the scenario on X. Cloudflare Radar information once more exhibits a major lack of visitors from Cuba starting round 13:35 native time (17:35 UTC) on March 16, dropping roughly 65%. Visitors returned to anticipated ranges by roughly 20:00 native time on March 17 (00:00 UTC on March 18), with the disruption lasting over 30 hours.

The third collapse (the second in only a week) occurred simply days later, on March 21-22. EnergíaMinas Cuba and OSDE/UNE once more supplied scenario updates through X. Cloudflare Radar information exhibits one other vital lack of visitors from Cuba starting round 18:30 native time (22:30 UTC) on March 21, falling as a lot as 77% in comparison with the earlier week. Visitors recovered round 21:39 native time on March 22 (01:39 UTC on March 23).

In line with a Fb put up from the Virgin Islands Water and Energy Authority (WAPA) on March 24, a lack of era on the Richmond Energy Plant mixed with injury to an underground cable triggered an influence outage affecting St. Croix and St. Thomas within the U.S. Virgin Islands. Cloudflare Radar information exhibits visitors from native supplier VI Powernet (AS14434), the first ISP for the U.S. Virgin Islands, dropping to close zero starting round 12:15 native time (16:15 UTC), with restoration occurring by roughly 14:45 native time (18:45 UTC). Though VI Powernet skilled a near-complete outage, visitors from St. Thomas solely fell by round 60%, and roughly 40% from St. Croix as a result of presence of different suppliers.

Storm Kristin made landfall in Portugal on January 28, inflicting widespread injury and energy outages throughout the nation. Roughly 1,500 incidents had been registered by Civil Safety between midnight and 08:00 native time (00:00 – 08:00 UTC), with the hardest-hit areas being the districts of Leiria and Coimbra. Vital infrastructure injury was reported, and by 07:00 native time (07:00 UTC), over 850,000 E-Redes clients had been with out electrical energy.

The related energy outages disrupted Web connectivity throughout Portugal, which Cloudflare Radar noticed primarily within the areas of Leiria, Santarém, and Coimbra starting round 04:10 native time (04:10 UTC) on January 28. Web visitors dropped as a lot as 70% in Leiria, and 52% in Coimbra.

Restoration was sluggish: over 290,000 clients remained with out energy as late as January 30, and Cloudflare continued monitoring gradual restoration of regional visitors over the next weeks. (Coimbra returned to anticipated ranges inside the first a number of days after the storm.) Greater than three weeks after the storm, over 6,000 clients in Leiria reportedly remained with out electrical energy.

Simply after the New 12 months, Web connectivity within the Republic of Congo was disrupted by an incident on the WACS (West Africa Cable System) submarine cable. Congo Telecom (AS37451) posted on X saying “an international incident on the WACS cable” was inflicting Web disruptions, and stating that backup options had been activated. Cloudflare Radar noticed a major drop in visitors from Congo starting round 00:00 native time on January 2 (23:00 UTC on January 1), falling to 82% under anticipated ranges. A follow-up put up from Congo Telecom confirmed that repairs had been ongoing, with customers doubtlessly experiencing slowdowns throughout peak hours. Visitors returned to anticipated ranges by roughly 15:00 native time (14:00 UTC) on January 4.

On January 14, a software program problem impacted voice and information providers for purchasers of Verizon Wi-fi (AS6167) throughout the United States. Verizon printed an official assertion acknowledging that the outage started January 14 and that by 22:15 ET (03:15 UTC on January 15) the difficulty had been resolved. A number of updates on X from @VerizonNews stored subscribers knowledgeable all through the night. Cloudflare Radar information exhibits a minor drop in visitors starting round 12:30 ET (17:30 UTC) on January 14, in line with the reported onset of the outage.

On February 9-10, clients of Move Grenada (AS46650) – the first Web supplier serving Grenada – skilled an island-wide service disruption lasting roughly 12 hours. The supplier posted on Fb acknowledging a service disruption, although no particulars in regards to the root trigger had been supplied. Cloudflare Radar information exhibits visitors from the community initially dropping round 11:30 native time (15:30) UTC on February 9, disappearing fully round 20:00 native time (midnight UTC on February 10), and recovering by roughly 23:30 native time (03:30 UTC on February 10). Routing information exhibits a whole lack of introduced IPv4 house on the identical time visitors dropped to zero. Main spikes in BGP bulletins across the time the disruption initially began, and bookending the entire outage, counsel that the entire occasion could have been routing-related.

Prospects of Orange Guinée (AS37461) in Guinea had been unable to make cellphone calls or entry the Web beginning round 10:45 native time (10:45 UTC) on January 6. Orange Guinée subsequently confirmed an “exceptional breakdown” affecting cell phone and Web providers because of a technical incident, with groups mobilized to revive service. Service was restored by roughly 14:00 native time (14:00 UTC) that very same day. No additional particulars on the basis reason behind the incident had been publicly disclosed.

On March 25, clients of UK broadband supplier TalkTalk (AS13285) reported widespread service disruptions. TalkTalk acknowledged the problems on X however didn’t publicly disclose a root trigger. Cloudflare Radar noticed visitors from the supplier drop practically 50% as in comparison with the earlier week starting round 07:00 native time (07:00 UTC), with service restored by roughly 08:15 native time (08:15 UTC).

The primary quarter of 2026 was marked by an unusually excessive variety of extreme and extended Web disruptions. The key government-directed shutdowns, notably the prolonged blackouts in Uganda and Iran, underscore how Web entry continues to be weaponized as a device of political management. Cuba’s three separate nationwide grid collapses in a single month paint a troubling image of infrastructure fragility with direct penalties for connectivity. And the drone strikes on AWS information facilities within the Center East signify an unprecedented escalation as lively navy battle straight and bodily broken main cloud infrastructure, with disastrous penalties for the web sites and functions hosted there.

The Cloudflare Radar workforce is continually monitoring for Web disruptions, sharing our observations on the Cloudflare Radar Outage Heart, through social media, and in posts on weblog.cloudflare.com. Comply with us on social media at @CloudflareRadar(X), noc.social/@cloudflareradar (Mastodon), and radar.cloudflare.com (Bluesky), or contact us through e mail.