AWS teams handling Middle East strikes

Amazon Web Services says its engineers are working “around the clock” after drone strikes damaged company data centers in Bahrain and the United Arab Emirates, leaving dozens of cloud services in the region still unavailable as of Tuesday, April 7. Amazon Web Services chief executive Matt Garman described the disruption on CNBC and said the company was still focused on keeping Middle East services running. (cnbc.com) That sentence lands differently if you remember what a data center is. A cloud service sounds weightless, but it runs inside real buildings filled with servers, power systems, cooling equipment, and fiber links, so a drone strike can break “the cloud” the same way it can break an airport or a refinery. (cnbc.com, datacenterknowledge.com) Amazon says the damage was not limited to one site. CBS News reported that drones struck two Amazon facilities in the United Arab Emirates directly and also damaged a data center in Bahrain, turning a regional war story into an infrastructure outage for companies that rent computing power from Amazon Web Services. (cbsnews.com, cnbc.com) The affected footprint matters because Amazon Web Services organizes its cloud by region. If a business stores applications, databases, or backups in the Middle East region and has not built copies elsewhere, damage to a few buildings can ripple outward into login failures, storage problems, and application downtime across many customers at once. (datacenterknowledge.com, crn.com) This was not a one-day hiccup. CNBC reported that the strikes happened last month and that dozens of services relying on Amazon Web Services data centers in the region remained unavailable on April 7, which suggests the hard part is no longer just extinguishing fires or restoring power, but repairing physical infrastructure and shifting customer workloads. (cnbc.com, crn.com) The Bahrain angle had already been flashing red before this latest Amazon update. On April 6, 2026, the U.S. Embassy in Manama issued a security alert telling Americans to find a secure location in a safe building, keep food, water, medications, and essential items on hand, and stay away from debris in case of an attack. (bh.usembassy.gov) That April 6 alert came after a string of earlier warnings. The U.S. Embassy in Bahrain said on March 2 the State Department had ordered non-emergency U.S. government employees and family members to leave Bahrain because of the threat of armed conflict, and a March 3 travel advisory told Americans to reconsider travel to Bahrain due to terrorism and armed conflict. (bh.usembassy.gov, bh.usembassy.gov) The military backdrop also widened beyond Bahrain. CBS News reported in March that Iran had named major U.S. technology companies as potential targets in the Middle East, explicitly tying commercial technology infrastructure to the expanding conflict and making cloud platforms part of the battlefield map. (cbsnews.com, cbsnews.com) By early April, the war reporting had started to mix escalation with ceasefire talk. CBS News reported on April 1 that President Donald Trump said he expected the U.S. war with Iran to end within several weeks, while later live coverage said Iran and the United States had agreed to a two-week ceasefire contingent on reopening the Strait of Hormuz. (cbsnews.com, cbsnews.com) For companies using Amazon Web Services in the region, the lesson is not abstract. “Cloud resilience” now includes missile defense, fuel supplies, physical security, and evacuation plans, because the same regional conflict that can close roads or empty embassies can also knock out storage systems, databases, and application servers inside a commercial data center. (cnbc.com, bh.usembassy.gov, cbsnews.com) That is why this story reaches beyond Amazon. A cloud outage used to mean a software bug, a bad update, or a power failure inside one facility; in March and April 2026, it also meant that a war in the Persian Gulf could physically damage the buildings behind the internet and leave global customers waiting while engineers work around the clock to keep services alive. (cnbc.com, datacenterknowledge.com)