Microsoft outage exposes risks of growing reliance on big tech providers
Major global outages have increasingly far-reaching consequences as companies grow more reliant on a handful of large corporations to provide critical online services, experts say.
A major global outage of multiple Microsoft services that caused heavy disruptions to businesses and governments has exposed a growing vulnerability of an interconnected digital economy, where organizations across the globe rely on a handful of providers to keep essential services running.
The issue, which began at 12:30pm Thursday Eastern Standard Time, restricted access to Outlook, Teams and Copilot. Microsoft initially traced the outage to cloud-based server problems in the Americas, which the company said overloaded remaining infrastructure and spread disruptions to other regions. After initial recovery efforts failed, it implemented a fix and began restoring affected services by Friday morning.
While some applications began to slowly come back online during the disruption, DownDetector, a website that monitors outage reports from users, said in a post on X that outages had “reached 74 countries and territories in about an hour” by Thursday evening, with problems continuing into Friday. The cause of the outage was still unconfirmed on Friday, but Microsoft said it is not a cyber attack.
The episode illustrates what some experts see as a troubling trend. Major global outages have increasingly far-reaching consequences as companies grow more reliant on a handful of large corporations to provide critical online services, raising questions about whether the infrastructure has become too concentrated.
The disruptions can prevent employees from accessing email, businesses from processing transactions and consumers from using everyday online services, becoming all the more significant as those digital experiences increasingly shape everyday existence.
“Recent outages affecting services such as Microsoft 365 and Canvas have reminded us how vulnerable we can be,” Katia Punter, assistant professor of software engineering and data analytics at Northeastern University London, told Northeastern Global News.
In 2024, a faulty software update by cyber security firm CrowdStrike disabled Microsoft Windows computers in offices, supermarkets, airports and sports stadiums in what some called one of the largest IT outages in history. A year ago, global cloud platform Cloudflare suffered an outage that knocked dozens of important websites offline, including OpenAI, Spotify and X.
This year has seen its share of disruptions. In February, another Cloudflare network failure disrupted internet access for customers using its Bring Your Own IP service, causing connection failures for users trying to access a host of services including Uber, Workday, Minecraft, Wikipedia and Microsoft Outlook. A month earlier, Microsoft experienced another disruption affecting Microsoft 365 services.
The outages also underscore the concentrated nature of vital cloud and technology services, and the way disruptions can have cascading effects across the digital economy.
Microsoft sells multiple products bundled as a suite, which can save their customers money. But when a provider like it experiences outages, it tends to take down multiple services at the same time — in this case Outlook, Teams and Copilot — maximizing the disruption. As a gesture, providers sometimes offer their clients outage credits, but these rarely provide adequate compensation, some experts say.
Editor’s Picks
Editor’s Picks
Editor’s Picks
But how can companies realistically protect against costly outages in 2026? A thorough strategic plan could be part of the answer, experts say, and although there’s no way to fully avoid outages, companies can try to limit the damage.
“The bundle is the trade-off — buying email, chat, files and AI from one supplier saves money and makes everything work together, but it also means one failure can hit several tools at once,” said Anees Baqir, assistant professor of data science at Northeastern University London. “Running everything twice across different platforms isn’t the answer either — it’s expensive, adds complexity and for most organizations simply isn’t sustainable.
Baqir said that companies should protect their most essential operations by maintaining backup communication channels through separate providers, ensuring employees can access critical files and contacts offline, and establishing contingency plans for system outages.
“Write a simple plan for what people should do when systems go down, and practice it,” he said. “And check what your contract actually covers — service credits rarely match the cost of a lost day.”
Jake Moore, an advisor at cyber security firm ESET based in London, said a plan B for companies should include alternative ways to communicate when email and Teams go down.
“Relying entirely on one technology provider creates enormous risks and most probably thought this would never happen,” Moore said. “But in a data hungry world this may be something that keeps reoccurring so resilience is vital.”
Causes of internet outages often include outdated or faulty equipment, human error, natural events that damage infrastructure and cyber attacks, according to Moore. Malicious actors may gain unauthorized remote access to a critical control system.
Additionally, Baqir said servers in data centers can fail and overload, which can cause disruptions.
“Outages like this are a reminder that the cloud is still physical infrastructure,” he said. “Microsoft has said it was not a cyber attack, but it’s still a strong case for having backup plans and not putting everything in one basket. Because so many organisations rely on the same provider, one problem can stop work for millions of people at once.”
The restoration of Microsoft’s services was piecemeal. On Friday morning, workers logging on in Australia, Asia and Europe were unable to access emails on Outlook. Government services, schools, hospitals, airlines and police forces are also among those impacted. The wider Microsoft 365 suite of services was also affected.
In a statement, OpenAI, which operates the AI chatbot ChatGPT, said some tools “are not working as expected” as it uses some Microsoft infrastructure.
Punter, who specialises in the teaching of technology at NU London, said she thought something was wrong with her computer on Friday morning when Outlook refused to open before realising the issue also affected other services. She agreed that backups and alternative solutions are essential.
“It is important to recognise how volatile cloud solutions can be,” she said.
