Shutdowns, power outages, and conflict: a review of Q1 2026 Internet disruptions
Key Points
- Nationwide shutdowns in Iran and Uganda
- AWS Middle East data centers struck by drones
- Three Cuban grid collapses disrupted connectivity
Summary
Q1 2026 saw a mix of government-directed shutdowns, large-scale power outages, and conflict-driven infrastructure damage that materially disrupted Internet connectivity across multiple regions. Notable events include nationwide shutdowns in Iran and Uganda, repeated national grid collapses in Cuba, regional power-grid impacts in South America and Eastern Europe, and physical strikes on AWS data centers in the Middle East. Cloudflare observed these events via bytes- and request-based traffic signals, BGP/route announcements, and connection-failure metrics.
Key Points
- Government-directed shutdowns: Uganda (Jan 13–26, near-complete mobile shutdown around the election), Iran (two sustained nationwide shutdowns beginning Jan 8 and Feb 28; filtering/whitelists implicated; Iran largely offline through late April), Republic of Congo (near-complete outage during March election, ~60 hours).
- Power outages: Cuba experienced three national electric grid collapses in March; Argentina, Moldova/Ukraine, Paraguay, and the Dominican Republic saw shorter but sharp connectivity drops tied to grid failures.
- Conflict and physical damage: Russian strikes caused region-level Internet drops in Dnipropetrovsk and Kharkiv (~50% drops). Drone strikes damaged AWS me-central-1 and me-south-1 (UAE/Bahrain), causing elevated connection failures and prolonged instability.
- Other disruptions: severe weather in Portugal, subsea/cable damage affecting the Republic of Congo, a Verizon Wireless technical incident in the U.S., and brief unknown provider outages in Guinea and the U.K.
- Observability notes: Cloudflare used both bytes- and requests-based graphs; IPv6 route withdrawals in Iran preceded traffic loss in one case, while aggressive filtering (whitelists/white SIMs) appears to have been used in the later shutdown. The Cloudflare Radar Outage Center contains a broader list of anomalies.
Operational impact
- Shutdowns and grid failures caused full or near-full loss of traffic (country or region level) for hours to weeks; region-level power attacks produced ~50% traffic degradation overnight.
- Cloud and origin availability was affected where physical infrastructure was damaged; customers in affected AWS regions were advised to back up or migrate workloads.
Recommendations for engineers
- Monitor both bytes-based and request-based traffic metrics and set rapid-alerting thresholds for sudden drops in either signal.
- Implement multi-region, multi-cloud redundancy and regular off-region backups for critical services; validate failover playbooks.
- Monitor BGP prefix announcements and IPv6/IPv4 withdrawals; use route-change alerts and heuristics to detect filtering vs. route withdrawal.
- Maintain communications plans for customers during prolonged geopolitical or national-level outages and test DNS/alt-connectivity workarounds.