Microsoft's global Microsoft 365 outage this week may have cost the world economy as much as $7 billion in lost productivity, according to an estimate cited by CNN — a figure that turns a 20-hour service failure into one of the most expensive cloud outages on record.

The estimate comes from Anderson Economic Group, a Michigan-based consultancy known for cautious assessments of disruption costs. Its president, Patrick Anderson, said as many as 26 million users saw their work disrupted on Thursday and 51 million on Friday, and warned the final bill could climb higher still as the full impact is counted.

The outage knocked out Outlook, Teams and Copilot for users in 74 countries from midday Thursday on the US East Coast, and stretched into Friday before services were restored. At its peak, more than 130,000 people reported problems with Outlook alone on Downdetector.

The $7 billion productivity bill

The scale of the disruption explains the size of the estimate. More than 430 million people worldwide rely on the affected Microsoft products for email, meetings and documents, and the outage struck during two full working days for many of them. Anderson's firm put the lost output at up to $7 billion, a number that has not been independently audited but carries weight because the consultancy's estimates are typically conservative.

Microsoft itself has not published a cost figure. The company apologised to customers and partners and said it would publish a full post-incident report once its investigation is complete. For businesses that lost a day of email and meetings, the wait for that report will be a long one.

What went wrong inside Microsoft's network

Microsoft traced the failure to a fault introduced during an earlier network capacity upgrade, which surfaced after a recent network change. Servers began failing health checks and a process that sets up secure connections was overwhelmed, leaving users unable to sign in or reach their mailboxes.

The company stressed the incident was a technical issue, not a cyberattack. The fault was logged under incident IDs EX1550543 and MO1550821, starting at 1:26 p.m. ET on Thursday, and it took engineers the best part of a day to work through the backlog of stalled connections.

Recovery nearly complete, but questions remain

By Friday morning in the US, Microsoft said mail backlogs had shrunk to near-normal levels, and services were restored in stages, starting in Asia-Pacific before moving to North America and Europe. Email timeouts were extended to 48 hours so delayed messages would still arrive rather than bounce.

Some users still reported intermittent problems into Saturday, and Microsoft has given no date for full recovery. The outage is the latest reminder of how much of the working world runs on a handful of cloud platforms — and how quickly a single faulty upgrade can bring it to a halt.