admin Cloudflare’s review of disruptions from January to March showed that the most damaging incidents were not dominated by software faults alone. Political orders, military strikes, fragile electricity grids, storms and physical cable damage repeatedly interrupted internet traffic, leaving businesses with fewer practical safeguards than many continuity plans assume.
Government-directed shutdowns formed one of the clearest patterns. Uganda ordered a nationwide internet shutdown ahead of its January 15 presidential election, with public access suspended from the evening of January 13. Traffic remained close to zero until partial restoration after President Yoweri Museveni was declared winner of a seventh term. Full restoration was announced later in January, but the interruption triggered legal challenges and criticism from digital-rights groups.
Iran presented the most severe case. A nationwide shutdown began on January 8 during anti-government protests, pushing traffic close to zero for nearly two weeks before partial restoration. A second shutdown began on February 28 as military strikes escalated, with traffic falling to well under 1 per cent of earlier levels. By late April, access remained heavily restricted, leaving many users dependent on whitelisted services, approved connections or costly workarounds.
For companies, Iran’s blackout illustrated a risk that extends beyond consumer access. Payments, freelance work, logistics, cloud access, customer support and software operations can be disrupted even when physical offices remain intact. Limited business-access programmes may reduce some losses, but tiered connectivity creates uneven access, compliance uncertainty and operational exposure for companies with suppliers, contractors or customers inside affected markets.
Military action also moved closer to the core of enterprise infrastructure. Drone strikes in the Middle East damaged Amazon Web Services facilities in the United Arab Emirates and Bahrain in early March, causing elevated connection failures in affected cloud regions. Amazon acknowledged structural damage, power disruption and water damage linked to fire-suppression activity, and urged customers with workloads in affected areas to back up data or migrate to other regions.
That episode has sharpened attention on cloud-region concentration. Enterprises often treat major hyperscale regions as highly resilient by default, but physical damage to data-centre campuses can affect origin servers, availability zones, latency, failover arrangements and customer-facing applications. The risk is particularly acute for firms that run regional workloads without tested multi-region recovery, independent DNS resilience, replicated data stores and clear authority to shift traffic during a crisis.
Ukraine’s connectivity was hit by attacks on energy infrastructure. Traffic in the Dnipropetrovsk region fell by nearly half after strikes on January 7 and 8, while Kharkiv recorded a similar drop after drone and missile attacks on January 26. These incidents showed how power systems and digital systems now fail together, especially where mobile towers, fibre networks and data centres depend on unstable grids and limited backup capacity.
Power failures produced disruption across several other markets. A heatwave-related outage in Buenos Aires affected multiple providers in January. Moldova and parts of Ukraine experienced traffic drops after an emergency grid event on January 31. Paraguay saw traffic fall by as much as 72 per cent on February 18 after key transmission lines went out of service. The Dominican Republic faced a sharp nationwide drop after a failure in its electricity system on February 23.
Cuba suffered three separate collapses of its national electrical system in March, each affecting internet connectivity. One failure cut power across large parts of the island, including Havana, while another lasted more than 30 hours. Traffic fell by about 65 per cent during one collapse and as much as 77 per cent during another, underscoring how deteriorating power infrastructure can become a direct threat to digital continuity.
Severe weather added another layer of vulnerability. Storm Kristin struck Portugal on January 28, damaging infrastructure and leaving hundreds of thousands of electricity customers without power. Internet traffic dropped sharply in Leiria, Coimbra and Santarém, with some areas recovering slowly over several weeks. The pattern reinforced the need for enterprises to examine not only cloud redundancy but also last-mile connectivity, field operations and workforce access.
Physical cable damage remained a persistent hazard. The Republic of Congo suffered disruption after an incident on the West Africa Cable System in early January, with traffic falling 82 per cent below expected levels before recovering after repairs and backup measures. Such events carry particular risk for markets with limited international routes, where a single subsea cable problem can impair banking, government services, media platforms and corporate communications.
Technical and unexplained failures added to the quarter’s disruption map. Verizon Wireless users in the United States were affected by a software issue in January. Flow Grenada experienced an island-wide outage in February, with routing data pointing to a possible routing-related event. Orange Guinée and TalkTalk in the United Kingdom also faced service disruption, with root causes either described only broadly or not publicly disclosed.
