Amazon continues to keep part of its AWS infrastructure operational in the Gulf, but the situation in Bahrain and the United Arab Emirates region is far from normal. Amazon itself officially acknowledged on March 24 that the AWS region in Bahrain had been “disrupted” due to the ongoing conflict and urged customers with loads in the area to migrate to other regions. The company added that it was working with local authorities, prioritizing the safety of its staff and assisting affected customers in relocating applications to other points within its global network.
This official message makes two things clear. First, that AWS has not abandoned the area nor given up its presence in Bahrain. Second, that recovery is neither quick nor straightforward. In parallel, the AWS Health Dashboard has been showing updates about “ongoing service disruptions” affecting both the Middle East (UAE) – me-central-1 region and Middle East (Bahrain) – me-south-1, confirming that the issue is not limited to a single resolved incident.
The most sensitive part of this story was not shared publicly by Amazon in a statement but was reported by Big Technology based on internal AWS communications. According to that source, Amazon considers one availability zone in Dubai and another in Bahrain to be “hard down”, and both could remain offline for an extended period without a clear timeline for normalization. The internal memo also suggests that AWS is minimizing operational footprint in both regions to facilitate customer migration to other locations. Since this language does not appear in Amazon’s official public statements, it should be treated as journalistic information based on internal material, not as official company communication.
Nevertheless, the overall picture aligns with other public descriptions of the incident. CRN, citing statements from Amazon and updates from the AWS health panel, explained that early March saw two data centers in the Emirates directly hit by drones, causing structural damage, power outages, and in some cases, activation of fire suppression systems that caused water damage. This coverage also indicated that Amazon was already transferring customer workloads to other AWS regions to maintain service continuity.
The key question is whether AWS remains operational in the area. The short answer is yes, partially. If Amazon had fully shut down in the Gulf, it would not make sense for the company to keep urging customers to monitor the health panel or to talk about phased migrations to other regions. Nor would internal communications describe regions as “impaired” alongside zones that are “hard down.” The different signals suggest an intermediate scenario: AWS still has active capacity in the Middle East, but it can no longer guarantee normal levels of redundancy or resilience in Bahrain and the Emirates.
This nuance is crucial because it shifts the understanding of the problem. We are not facing a total and permanent blackout but rather a cloud infrastructure that remains alive but degraded—more fragile and with some of its capabilities severely impaired. In other words, AWS is holding on, albeit with significantly reduced operational capacity in these regions.
Beyond Amazon’s case, what has happened offers a much broader lesson for the tech sector. For years, Gulf cloud regions have been positioned as strategic platforms for banking, government, telecommunications, and increasingly Artificial Intelligence infrastructure. When a regional war directly impacts data centers belonging to a major hyperscaler, the conversation changes: availability, latency, and cost are no longer the only variables; geopolitical risk becomes part of architectural planning.
It also compels a rethink of the traditional idea of regional high availability. Theoretically, a cloud region with multiple availability zones is designed to withstand localized failures. But when the problem involves repeated attacks, physical damage, power outages, and sustained threats to multiple facilities within the same region, the discussion shifts from internal redundancy to the prudence of distributing workloads across multiple sovereign regions or even multiple providers. Amazon’s internal messaging to customers explicitly recommends moving workloads out of Bahrain and the Emirates while this phase of the conflict continues.
In this context, what is happening with AWS in the Gulf could become an uncomfortable reference point for the entire cloud industry. Public cloud has long promoted the idea that its main strength is abstracting physical data center complexity. These attacks remind us of the opposite: ultimately, cloud services also depend on buildings, power, fiber optics, substations, personnel, and territory. When all of these are within the reach of war, “cloud” no longer appears as ethereal as its name suggests.
Frequently Asked Questions
Is AWS still operating in Bahrain and the UAE?
Yes, but in a degraded state. Amazon has officially confirmed disruptions in Bahrain, and the AWS health panel shows ongoing issues in both Bahrain and the Emirates. Internal reports from Big Technology indicate that some zones are “hard down,” while others are deteriorated but still operational.
Has Amazon provided a recovery timeline?
No. In its official statement regarding Bahrain, Amazon mentions ongoing recovery and customer migration efforts but does not specify a timeline for returning to normal. Internal reports from Big Technology suggest there is no clear schedule either.
What is AWS doing for affected customers?
Amazon states it is helping customers migrate their applications to other AWS regions and that many are already operating from elsewhere around the world.
What does this incident tell us about the cloud industry?
It demonstrates that large data centers and cloud regions can become physical targets during conflicts, and that resilience depends not only on architecture but also on geopolitical stability and the physical infrastructure supporting the cloud.
via: About Amazon, LinkedIn and Big Technology