For most of Thursday and into Friday, a large chunk of the working world discovered what it actually runs on. Outlook stopped sending mail. Teams went dark. Copilot, Microsoft's AI assistant, fell over alongside the tools it is supposed to assist with. The UK's National Health Service told patients to keep their appointments but expect chaos, police forces asked the public to phone in crimes the old way, and schools from Glasgow to Melbourne lost their lesson platforms. At least 74 countries reported problems, and roughly 430 million people use these tools every day. One company's bad day became everyone's bad day, and that is the story.
What actually happened
Microsoft says it was not a cyberattack, just a technical failure. The company's own explanation is that higher-than-usual stress on its servers caused the collapse, with customers unable to send or receive email, open certain files or see their calendars from late Thursday. Restoration crawled along for nearly 24 hours; by Friday afternoon Microsoft was reporting encouraging progress in Europe and improving queues in North America, while warning that intermittent issues would linger. The painfully vague wording matters, because a failure this big with an explanation this thin is how trust erodes. Source: https://www.bbc.co.uk/news/articles/cwvg238zrnm5o
Why it hurt so much
The outage did not hit equally. It hit institutions that had quietly converted their operations into a Microsoft subscription. NHS England enacted disruption plans and told patients to use 111 or pharmacies instead of messaging their GP. Several large police forces stopped taking online reports and asked people to call. Courts, councils and governments around the world reported knock-on failures. The BBC's framing was blunt: this was a stark reminder of how much of modern life rides on a single vendor's uptime, and the UK government said it was working directly with Microsoft to assess the damage. When email is down at a hospital, that is no longer an IT story. It is a public safety story. Source: https://www.bbc.co.uk/news/articles/c6vg2ljjpj32o
The question nobody at Microsoft wants asked
Every previous outage this size triggered a round of hand-wringing about cloud concentration, followed by nothing. This one should stick, because the math gets worse every year. Schools that moved classes onto Teams, health systems that moved referrals onto Outlook, businesses that moved everything into Microsoft 365 did so one sensible decision at a time, and each individual decision was fine. The failure mode only appears in aggregate, when one stress event in one company's server fleet stalls pharmacy orders and police reports across four continents. Nobody chose fragility. Fragility was the side effect of a thousand reasonable choices, which is exactly why it keeps happening. Source: https://www.bbc.co.uk/news/articles/cxzd9171m51qo
What to watch
Microsoft will publish a post-incident report in a week or two, and the interesting question is whether regulators treat this the way they treat banking outages. The UK's critical national infrastructure rules already cover some of these services in spirit. If hospitals and police forces cannot function for a day when one vendor sneezes, the case for mandatory multi-vendor resilience, or at least honest disclosure of single points of failure, makes itself. Until then, the practical advice is boring and unpopular: keep a way to reach your GP that does not depend on one company's cloud. Thursday showed, at 74-country scale, why that advice is not paranoia.