The Microsoft Outage May Be Costing the World $600 Million an Hour

Every hour that Outlook stays dark during a working day, the world's employers may be paying about $600 million, and plausibly anywhere from $300 million to $1 billion, for work that isn't getting done. That is our estimate of the bill for the Microsoft 365 outage that began on Thursday, November 12, and it is still running.

Nobody has published an official figure, and nobody will for weeks. But the arithmetic is simple enough to do on a napkin, and the answer is large enough to matter. Here is how we got there, and how to work out your own share of it.

What broke

Exchange Online, the service behind Outlook email and calendars, began failing on Thursday afternoon US time. Everything in Microsoft 365 that leans on it went down with it. BBC News reports that problem reports for Outlook peaked at nearly 140,000 worldwide, with users affected from the US to Japan and Singapore.

The damage spread well past email. Copilot stopped working properly, and users lost calendar details and access to some files, although chat and online meetings largely held up. Microsoft's incident notice also lists SharePoint Online and its Defender XDR security suite among the affected services. OpenAI and Anthropic both reported problems with their connections to Microsoft 365.

The BBC reports there is no indication of a cyberattack. Microsoft's explanation so far is that a number of servers became unexpectedly unavailable, which overloaded the ones still running. Fixes have repeatedly failed to stick. ABC News reports Microsoft's warning that restored servers can become unresponsive again once they receive traffic, and Reuters reports the company saying its latest efforts were not progressing as expected. At the time of writing, there is no confirmed root cause and no estimated time of recovery.

The math behind the number

The central estimate is about $600 million per working hour, with a plausible range of $300 million to just over $1 billion. It comes from multiplying four things: how many Microsoft 365 users are at work in a given hour, what share of them are affected, how much of their productivity they lose, and what an hour of their time costs.

A few words on each input. Microsoft has roughly 450 million paid commercial seats worldwide (approximate; its last public figure was above 430 million). The sun is never up on all of them at once, so we count only the 30 to 45% who are in working hours at any moment. Not every tenant was hit and recovery came and went, hence the 40 to 60% affected. Email and calendars were down but chat, meetings and documents mostly worked, so the productivity loss is a fraction, not a full stop.

The outage began in the US afternoon, ran through the whole Asia-Pacific Friday and was more than 12 hours old as Europe sat down to work. Adding up the hours region by region gives a running total of roughly $3 billion to $7 billion, and it grows with every hour the fix doesn't hold.

"They'll catch up on Monday" is not a refund

The usual objection to numbers like these is that the work isn't lost, only late. Emails arrive eventually, meetings get rebooked, people catch up. That is partly true for the economy's output. It is not true for the employer's payroll.

A salary buys time. When a third of someone's Friday goes to refreshing an inbox that won't load, the employer has paid for hours it did not receive. One fully lost working day is about 0.4% of an annual salary. Most affected workers will have lost a fraction of that, somewhere between 0.05% and 0.2%, but multiplied across hundreds of millions of people it becomes the billions above.

Catching up has its own price. Clearing a backlog takes paid effort on top of the downtime: either next week's planned work gets pushed out, or people absorb it as unpaid overtime.

So the bill is real even if every delayed email is eventually answered. The cost has only moved to another day, or to someone's evening.

Who pays the most

The average hides a wide spread. A developer with code checked out locally may lose 5% of the day. A recruiter, an account manager or a paralegal whose work lives in the inbox may lose more than half.

• Small businesses. A large bank can absorb a day like this. A five-person accounting firm on thin margins has no fallback mail system and no IT team to improvise one.
• Anyone with a deadline. Electric truck maker Workhorse Group moved its third-quarter results from November 12 to November 16, citing the outage, according to the BBC. Tenders, filings and closings with Friday deadlines face the same problem.
• Security teams. Microsoft's incident notice lists Defender XDR among the affected services, so some security monitoring was degraded during the same hours that staff were improvising workarounds.
• IT support. Helpdesks that take requests by email lost their intake channel just as demand peaked.
• Public services. New York City's emergency management department warned residents of delays in reaching city agencies by email, ABC News reports. Reuters names Australia's spy agency and federal police and the US House of Representatives among the organisations affected.

Life sciences: about 350,000 seats at five companies

Five of the world's largest drugmakers account for about 350,000 Microsoft 365 Copilot seats between them, going only by figures Microsoft has made public. None of them has said it was affected by this outage, so what follows is a measure of exposure, not a report of damage.

Copilot is sold as an add-on to a Microsoft 365 licence, so each seat implies a Microsoft 365 user. Three of the figures are purchase commitments, which can run ahead of active use. Even so, the total number of mailboxes at each company is likely to be higher than its Copilot count.

Applying the same assumptions as the global model, 40 to 60% of those 350,000 users would have been affected: roughly 140,000 to 210,000 people at these five companies alone. At half affected and a fifth of productivity lost, that is about $10 million per eight-hour working day at the global $35 hourly rate, and closer to $17 million at a rate of $60 that better reflects pharma pay.

The salary figure understates what is at stake in this industry, because so much regulated work runs on email and has a clock attached.

• Drug safety. Adverse event reports often arrive by email, and serious cases must reach regulators within fixed windows, typically 15 calendar days and as few as 7 for the most serious clinical trial cases. A lost day of intake eats directly into that margin.
• Clinical trials. Sponsors, research organisations and hospital sites coordinate across companies by email. Safety notifications to investigators and site queries stall when it stops.
• Regulatory affairs. Questions from health authorities come with response deadlines that do not move for a vendor's outage.
• Quality and manufacturing. Deviation reviews and batch-release sign-offs often depend on email notifications and shared documents, even where the system of record is separate.
• Field teams. Sales and medical staff lose calendars and customer correspondence, which Microsoft confirmed were among the features hit.

Work out your own number

The same formula works for a single company, and you will know your inputs far better than we know the world's.

Cost = employees × share affected × productivity lost × hours × hourly cost

Take a 1,000-person firm where 60% of staff were affected, each losing a quarter of their productivity across an 8-hour day, at a loaded cost of $40 an hour. That is 1,000 × 0.6 × 0.25 × 8 × $40, or $48,000 for the day.

Two adjustments make the answer more honest. Use loaded cost, not base salary: benefits and overheads typically add 25 to 40%. And set the productivity figure by role instead of using one number for everyone, because that input moves the result more than any other.

What to take from this

Every large platform has days like this, and the alternatives have had theirs, so switching vendors is not the answer. The more useful finding is that most organisations have never priced what a day without email costs them, so they have never decided what it is worth spending to soften one.

1. Know your hourly number. Run the formula above before the next outage. A figure like $6,000 an hour turns a vague worry into a budget conversation.
2. Keep one channel that doesn't share the failure. A phone tree, a status page on separate hosting, or a second messaging tool. It only has to be good enough to tell people what is happening and what to do.
3. Decide in advance what people do when email is down. The costliest hours are the ones spent waiting to see if it comes back. A short list of offline-capable work per team recovers a surprising share of the day.
4. Don't count on compensation. Service credits are tied to what you pay for the subscription, not to what the downtime cost you. For most firms the credit will be a small fraction of the loss.

All four can be settled in an afternoon, well before the next outage.

How solid is this estimate?

It is a back-of-envelope figure, and it should be read as an order of magnitude. Three things could move it a long way.

• The share affected and the productivity lost are judgment calls. No one has measured them. Halve both and the hourly cost falls to a quarter.
• The seat count and labour cost are approximations. The 450 million seats is a rounded figure from Microsoft's past disclosures, and $35 an hour blends very different wage levels across regions.
• It leaves things out. Lost sales, missed deadlines, IT response costs, security exposure and consumer Outlook users are not in the number. Work that people shifted to other tasks is only roughly accounted for.

No insurer, analyst or major news outlet has published a global cost estimate yet. Firmer figures should follow in the weeks ahead.

About Zelthy Research Desk

Zelthy Research Desk is the research and editorial arm of Zelthy. It covers how technology and AI are changing pharma and life sciences, from R&D and regulatory work to commercial, medical and field operations, for the leaders who have to decide what to adopt and when. The desk is owned by Zelthy and discloses that on its work. Its charter is vendor-neutral, Zelthy included, with no paid placement or sponsored rankings. It aims to publish its sources and methods with every piece and to correct errors openly.

Sources

BBC News: Microsoft outage sees Teams, Outlook and Copilot suffer 'connectivity issues', for scale, affected services and companies hit

ABC News: Microsoft Outlook outage stretches on for hours, for Microsoft's recovery updates and the New York City warning

Reuters: Microsoft outage drags on for hours, disrupting Outlook, Teams for businesses and governments (syndicated copy on AOL), for recovery status and organisations affected

Microsoft 365 service health notice MO1550821, for the list of affected services

Cost-per-hour model and inputs: our own estimate, as set out above

Microsoft fiscal 2026 third-quarter earnings call, April 29, 2026, Microsoft Cloud Blog and Microsoft WorkLab, for Copilot seat figures at life sciences companies