The immediate consequences were felt not just in corporate boardrooms but in critical public services. In the United Kingdom, the Metropolitan Police noted that some online form submissions were affected, while healthcare providers struggled to share patient records between services. Bloomberg reported that surgical procedures were rescheduled due to the technical failures. In Canada, the impact was similarly severe; Unity Health Toronto, which operates several hospitals including St. Michael’s, declared a “code grey,” an infrastructure failure alert, as staff faced difficulties sending and receiving emails. Although the hospital network stated that direct patient care was not affected, other institutions such as Sunnybrook Health Sciences Centre, the Jewish General Hospital, and CHUM in Montreal reported disruptions to appointment requests and internal activities. The outage also extended to transportation and education, with Via Rail advising customers to check train trackers due to email notification failures, and post-secondary institutions like the University of Toronto and Seneca Polytechnic reporting email problems, though classes proceeded as scheduled.
Microsoft first flagged the outage on Thursday around midday Eastern Time, initially noting issues with Exchange Online mailboxes in the Americas before clarifying that the issue could affect users “in any region.” The company’s technical explanation, provided in updates on X and its status page, pinpointed the root cause not to a surge in customer traffic, but to internal infrastructure strain. Microsoft stated that a “number of servers had unexpectedly become unavailable,” which placed elevated load on the remaining infrastructure across multiple regions. This unexpected unavailability prompted connectivity errors that cascaded through services including Outlook, Teams, and the company’s AI assistant, Copilot.
The ripple effects of the outage extended beyond Microsoft’s own ecosystem, impacting third-party artificial intelligence tools that rely on Microsoft 365 integrations. Both Anthropic and OpenAI issued status updates confirming that their connectors to Microsoft 365 tools were disrupted. Anthropic additionally noted that authentication issues were affecting its Claude Code product, though the company did not explicitly attribute this specific issue to the Microsoft outage. For developers and enterprises relying on these AI integrations, the downtime highlighted the growing dependency of modern AI workflows on stable cloud infrastructure providers.
As the outage entered its second day, Microsoft focused its recovery efforts on stabilizing the network and clearing the backlog of delayed communications. In a 7:09 a.m. Eastern Time update on Friday, the company reported “encouraging recovery progress,” noting that service availability and email delivery performance were improving across a growing portion of affected infrastructure. Specific regions, including Europe and Japan, were becoming more stable, while North American services also showed signs of improvement. However, Microsoft warned that some users might still experience intermittent issues as recovery actions continued.
A key technical measure introduced during the crisis was the extension of the email timeout window to 48 hours. Normally, undeliverable messages might return an error to the sender, but Microsoft kept these messages in the system for up to two days to allow for eventual delivery as the infrastructure recovered. This approach aimed to prevent the permanent loss of business communications during the restoration process. In the NBC 5 Chicago newsroom, the recovery was tangible but imperfect; while emails began hitting inboxes around 6:00 a.m. Central Time and could be sent successfully, producers reported a 30-minute lag in delivery, illustrating the “intermittent issues” Microsoft had cautioned about.
Microsoft explicitly stated it had “no indication” that the disruption was caused by malicious activity, such as a cyberattack, and committed to treating the incident with “urgency and priority until full resolution.” The incident serves as a stark reminder of the fragility of global digital infrastructure, where a failure in server health checks and automated recovery processes can halt operations for governments, hospitals, and businesses worldwide. While the company has not provided a definitive timeline for full recovery, the focus on clearing email backlogs and restoring service consistency remains the primary objective as the second day draws to a close.
