November 18 was supposed to be a routine workday for Rwanda-based entrepreneur Derrick Ikenga. But it turned out to be a nightmare he wasn’t prepared for.
Cloudflare, a U.S.-based cloud services and cybersecurity firm, faced a massive global outage that day, taking down large parts of the internet, including operations of giants like X (formerly Twitter) and ChatGPT.
Ikenga’s company, which builds sales automation tools for clients, also uses Cloudflare. The outage brought the tools to a grinding halt, sending his clients into panic mode.
For many entrepreneurs and tech workers, this was a familiar experience in 2025. Today, an estimated 94% of enterprise services worldwide rely on at least one major cloud provider, such as Amazon Web Services, Microsoft Azure, or Google Cloud, for hosting, data storage, or delivery. Together, the three largest providers control more than 62% of the global cloud market.
Between August 2024 and August 2025, AWS, Azure, and Google Cloud together experienced more than 100 service outages. The duration of these outages varied between a few minutes and several hours. AWS saw one of the biggest and most widespread cloud outages of 2025 when its services went down for 15 hours in October. The outage affected over 4 million users and more than 1,000 companies.
When an outage occurs, cloud service providers are responsible for restoring services as soon as possible. In some cases, providers compensate affected customers with free cloud credit.
Each time such disruptions occurred, everyday moments began to fail in unexpected ways. Online payments failed. Doctors were unable to access patients’ health records. CCTV cameras went offline. Doorbells stopped working. Smart mattresses lost their sleep-tracking features.
The outage offered a glimpse into just how deeply cloud infrastructure is woven into modern life, powering everything from bank transfers to flight bookings and movie streams.
Rest of World brings together first-person accounts from entrepreneurs and technology workers around the world, tracing what happened when the cloud failed in 2025.
The quotes have been edited for length and brevity.
Helpless because of a cloud outage
Francisco Osorio, a Mexico-based senior go-to-market marketing engineer at Simetrik, an operational control and management automation company
I couldn’t do anything other than feel helpless.
I am responsible for the smooth functioning of my company’s website, among other things. During the November AWS outage, the website stopped functioning, but I was unable to take any action because I lost access to all my work-related platforms, like HubSpot, Slack, and Salesforce. [Some users of AWS experienced issues on November 5, according to outage tracking website Downdetector.com. An AWS spokesperson denied the claims in comments to Reuters.]
Even as we were sorting through internal issues, our clients also couldn’t log into the platform, which was a more serious concern because we need to comply with service level agreements under which we assure quality and quantity of service.
I couldn’t do anything other than feel helpless. All I could tell everyone around was, “I’m deeply sorry, I cannot help you today. Our platforms are not running.”
Standard operating procedures for cloud outage
Sundeep Narwani, India-based co-founder of Narrative Research Labs, a firm focusing on AI in media
Being an AI company, we’re quite heavily dependent on the cloud. The outages have wasted a lot of time for our clients and us, rendering us helpless in most cases.
The outages have wasted a lot of our time.
The first time it happened, we didn’t know what to do or how to deal with it, and we entered panic mode. My biggest concern was on two fronts: downtime disrupting our processes, and our clients being obviously very angry because we deliver a lot of daily reports for them.
When it happened the second time, we created a standard operating procedure to deal with it because someone like me, who is involved in business development apart from being the CEO, cannot be available all the time to handle these issues.
We delegate a project manager to handle issues of cloud outages and ensure our customers are taken care of during such times while finding a quick solution to the downtime.
We have also purchased a local server to replicate what AWS does, so that if there is any kind of outage again, we have one local server to tide us over. Obviously, it doesn’t have the same kind of performance that AWS gives us, but at least the team can function. But these are just temporary options.
On-premise cloud solution
David Nandwa, founder of Honeycoin, a Kenyan cross-border payment company
We have faced cloud outage issues in the past and were also hit with the recent AWS outage in November.
The approach we have applied to dealing with this is to have a “distributed stack” or multiple fallback options. We rely on multiple clouds, including AWS, Google Cloud, and Heroku. During the recent AWS outage, we temporarily switched to Google Cloud to maintain service availability for most of our customers.
If all cloud providers are down, we then fall back to using an on-premise server that we’ve created. It doesn’t perform as efficiently as our providers, but it comes in handy in moments where all the major providers are down. It’s locally hosted, so it’s an option to process requests when all else fails.
Subscribing to multiple cloud providers definitely helps us not only to hedge against future risk but also to decrease reliance on a single one, and risk losing our data completely.
However, multiple providers mean multiple bills at the end of the month.
The one-off cost of a single on-premise server installation is something we’re seriously considering to decrease reliance on the large players and also have more digital sovereignty. It’s pretty simple to set up, but it requires a lot of work to move our stack from any of the globally reputable providers to our own direct stack.
Navigating through a cloud outage
Olumide Egbigbola, Nigeria-based product manager at Leatherback, a cross-border payment startup
Companies like ours can’t do much in such cases.
Cloud infrastructure is the backbone of most of the things that we do, as most of our critical services are hosted in the cloud. So any downtime from our cloud provider is immediately felt by our individuals and business users. We face complaints of transaction delays, and the app slowing down or being unresponsive.
This usually causes a lot of anxiety amongst users who can’t access their money.
Companies like ours can’t do much in such cases other than inform our users about the downtime.
We try to spread our cloud services and route traffic through regions that are less affected or unaffected, so that users can at least perform minimal activities while we wait for the major downtime to be over.
