Analysis
Microsoft's Outage Enters Day Two. What It Means for Stocks
Microsoft closed up 0.36% at $523.00 on November 12, 2026, the day a worldwide Microsoft 365 outage began. The outage was still open on Friday morning. History says the share price barely notices a short outage; the questions that could change that are how long this one lasts, what caused it and whether it touches Azure.
Outlook, Teams and Copilot stopped working for customers around the world on Thursday, November 12, and Microsoft shares finished the day higher. The stock closed at $523.00, up 0.36%, on a session when the Nasdaq Composite fell 0.39% and the S&P 500 lost 0.35%. By Friday morning in Europe the outage was still not fixed, which makes it one of the longest Microsoft 365 failures in years. The gap between a very public failure and a calm share price is the place to start, because it says a lot about how the market values this company and about what would have to happen for that calm to end.
The facts first. Microsoft's own advisory, numbered MO1550821, puts the start of the wider incident at 1:26 PM Eastern time on Thursday, BleepingComputer reported, after the company had spent the morning looking at connection problems in Exchange Online, the service behind Outlook mailboxes. The advisory lists Exchange Online, Teams, SharePoint Online, Microsoft 365 Copilot and the security product Defender XDR. Fox Business added Intune, Purview, Power BI and Windows 365 to the list of services with problems. Downdetector had counted more than 130,000 reports within hours, and Microsoft said on X that users in any region could be affected.
The cause is still not confirmed. Microsoft first pointed to elevated load on its infrastructure across several regions, then said that a number of servers had become unexpectedly unavailable and left the rest overloaded. According to SecurityWeek, a process that sets up secure connections was being overwhelmed, affected servers were failing their automated health checks, and the company began rolling back a recent network configuration change that coincided with the start of the trouble. Microsoft has not said that this change was the trigger. Charles Lamanna, an executive vice president, wrote that there was no indication of malicious activity or a cyberattack.
Recovery has gone back and forth. Microsoft reported limited recovery shortly before 7 PM Eastern on Thursday, then said that some servers stopped responding again once traffic returned. Restarts produced improvement around 9 PM, an update after 10 PM said the work was not producing the expected results, and a partial recovery in some regions before midnight did not hold. The Belgian outlet ITdaily, following the incident through the European morning, quoted Microsoft at 8:50 AM Central European time on Friday as still working to restore service, with no time given for a full recovery. Anyone reading this later should check Microsoft's service health page for the current status.
Now the share price. Thursday's small gain needs context, because the week had already been poor for the stock. Microsoft closed at $545.18 on Monday, November 9, and has lost 4.07% since then. It set a record close of $555.80 on November 4 and now sits 5.9% below it. Thursday was also a bad day for technology shares in general: Nvidia's chief executive resigned and the 10-year Treasury yield moved back above 5%. A stock that rises on a day like that, while its best-known products are offline, is telling you that investors did not treat the outage as a reason to sell. At least not yet.
There is a simple reason for that. Microsoft does not sell Outlook and Teams by the hour. Companies pay per user, per month, usually on contracts that run for a year or longer, and the money keeps arriving whether the service works on a given Thursday or not. An outage does not cancel those contracts. It does not reduce the number of seats a customer has bought, and it does not change the revenue Microsoft will report for the quarter in any way that an analyst could see. In its last quarter the company reported revenue of $92.28 billion, up 19%. One bad day, or two, does not move a number of that size.
The clearest precedent is July 19, 2024, when a faulty update from the security company CrowdStrike crashed millions of Windows computers and grounded flights around the world. Microsoft's name was in every headline. Its shares fell 0.74% that day, from $440.37 to $437.11, and were higher the next session. CrowdStrike, whose software was at fault, fell 11.1% on the day and was 36.5% below its pre-outage close two weeks later. The market made a clear distinction between the company that caused the failure, and whose customers could plausibly leave or sue, and the platform that everybody would still be using a month later.
That precedent cuts both ways this time. In 2024 Microsoft was the platform and somebody else's code was the cause. On Thursday the servers, the network and the configuration change under investigation were all Microsoft's own. There is no supplier to blame. This is also not the first long failure of the year: ITdaily noted that Microsoft 365 was down in North America for almost ten hours in January. One long outage is an accident. Two in the same calendar year, on a product sold to businesses as something they can depend on, starts to become a pattern that customers bring up when contracts are renewed.
The direct cost is the easiest part to estimate, and it is small. Microsoft's service agreement for Microsoft 365 promises 99.9% uptime in each month. A 30-day month has 720 hours, so that promise allows about 43 minutes of downtime. Below 99.9%, a customer can claim a credit of 25% of that month's fee for the affected service. Below 99%, which means more than 7.2 hours down, the credit is 50%. Below 95%, or more than 36 hours, it is 100%. An outage that began at 1:26 PM on Thursday and was still running on Friday morning had already passed the second threshold for the tenants that lost service throughout.
Even so, the arithmetic is modest. A 50% credit on one month is one twenty-fourth of what that customer pays in a year, a little over 4%. A full month's credit is one twelfth. The credit applies to the affected services and to the customers who were actually down, and customers have to file a claim to receive it, which many never do. Credits are also deducted from future bills, so they appear as slightly lower revenue spread over later months. For a company that produced $18.6 billion of free cash flow in a single quarter, this is a cost that will be hard to find in the accounts.
The costs that matter are the ones that do not appear on an invoice, and the first of them is the story Microsoft is telling investors about artificial intelligence. The company's valuation rests on the idea that businesses will pay more per user for Copilot and will run AI agents that work through their email, documents and meetings. Microsoft even reorganised its reporting around that idea: its main segment is now called Agents and Infra and brought in $76.8 billion last quarter, up 25%. Copilot was on the list of services that failed on Thursday. An agent that does a company's work is only as reliable as the platform underneath it.
The second is security. Defender XDR, the product that many companies use to detect attacks across their networks, was among the affected services, and so were Intune, which manages employee devices, and Purview, which handles compliance. A security team that cannot see its own alerts for a day has a real problem, and it is the kind of problem that gets written up for the board. Microsoft has spent two years trying to repair its reputation on security after several serious breaches. Rival security vendors, CrowdStrike among them, will not need much encouragement to mention this week in their next sales meetings.
The third is regulation. Since January 2025 the European Union's Digital Operational Resilience Act has required banks, insurers and investment firms to manage the risk of depending on a small number of technology suppliers, and it gives supervisors direct oversight of the providers judged critical. Regulators in Britain and the United States have raised the same concern about concentration in cloud services. An outage that reaches users in every region, with a cause still unconfirmed after a day, is the scenario those rules were written for. Nothing here threatens Microsoft's earnings this year, but it adds weight to the argument that large customers should keep a second supplier.
The biggest question for the share price is Azure. Microsoft disclosed Azure's revenue in dollars for the first time in October: $33.17 billion in the quarter, up 48%, and it forecast growth of 49% to 50% for the current quarter. That growth is the main reason the stock trades where it does. Microsoft's official list of affected services is a Microsoft 365 list. Some outlets reported user complaints about Azure on Downdetector, but the advisory does not name it. If the post-incident report shows that the failure was limited to the systems that serve Microsoft 365, investors will probably move on. If it shows a weakness in networking shared with Azure, the conversation changes.
What about competitors? If investors believed customers would leave Microsoft over this, the obvious beneficiaries would be Alphabet, which sells Google Workspace, and Zoom. Alphabet rose 1.79% on Thursday to $298.70, but it had been climbing all week for its own reasons. Zoom slipped 0.4% to $97.70 after jumping 7.9% the day before the outage. There was no visible rush into the alternatives. That fits how these decisions are made: moving tens of thousands of employees from one email and document system to another takes a year or more, and no company starts that project because of one bad Thursday.
The outage also showed how far Microsoft's systems reach into other companies' products. ITdaily reported that Anthropic and OpenAI said their Microsoft 365 connectors for Claude and ChatGPT had stopped working. SecurityWeek noted that some Google Workspace customers could not reach their Google-hosted mail through the Outlook apps. Portland Fire & Rescue in Oregon warned the public that replies to emails would be delayed. None of these is a financial event. Together they show that a large part of the economy's daily communication runs through one company's servers, which is the source of Microsoft's pricing power and, on days like this, of its exposure.
For the shares of Microsoft's customers, the effect is likely to be close to nothing. Lost email and missed meetings are a real cost, but it is spread thinly across millions of organisations and does not show up in anyone's quarterly figures. The 2024 CrowdStrike failure was different because it stopped physical operations: Delta Air Lines cancelled thousands of flights and later said the episode cost it about $500 million. An office that cannot send email for a day catches up the following week. Unless reports emerge of trading desks, hospitals or airlines that could not operate, there is no reason to expect a mark on other companies' results.
The wider market is exposed mainly through Microsoft's size. It is one of the largest companies in both the S&P 500 and the Nasdaq-100, so a large move in its shares moves the indexes and every fund that tracks them. That did not happen on Thursday. The indexes fell because of Nvidia and bond yields, and Microsoft was a small support, not a drag. For the outage to become an index story, Microsoft itself would have to fall several percent, and that would need news that changes estimates of future revenue. Service credits and a day of embarrassment are not that kind of news.
Traders who follow price levels have a few to note. Microsoft's 50-day moving average was $518.72 at Thursday's close, about $4 below the share price, and the stock has stayed above that average through this week's decline. Below it, the November 2 close of $507.12 is the low of the past two weeks. Above, the record close of $555.80 is the reference. The 200-day average is far lower, at $440.97. The stock has also been unusually active: its 30-day realised volatility is about 32% on an annualised basis, so daily moves of 2% in either direction have been common lately, with or without an outage.
From here, three outcomes seem possible. In the first, and the most likely on past form, service returns on Friday, Microsoft publishes a report that traces the failure to a specific configuration change, and the subject fades by the middle of next week. The shares then go back to trading on interest rates, on the mood around AI spending after Nvidia's news, and on the next data about Azure demand. Thursday's close suggests this is what most investors expect. Analysts tracked by TipRanks have 35 buy ratings and four holds on the stock, with an average price target of $632.48, and one outage is unlikely to change those.
In the second outcome, the outage runs through Friday's US session or into the weekend, or service returns and then fails again. Duration matters more than anything else here. Past 36 hours, affected customers move into the tier where a full month's fee can be credited. More important, a company that cannot restore its own main product in a day and a half invites harder questions about how its systems are built. This is the case where the shares could come under real pressure and the 50-day average would be tested, because the story stops being about one error and becomes about the design.
The third outcome is the least likely and would be the most serious: that the cause turns out to be a security incident. Microsoft has said plainly that it sees no sign of one, and nothing in the public reporting contradicts that. The scenario is worth naming only because it is the one that would produce a reaction like CrowdStrike's in 2024, with customers reviewing contracts and regulators opening inquiries. A related, milder version would be a finding that the outage was linked to this week's security patches. SecurityWeek reported that Microsoft has not connected the two.
The dates to watch are straightforward. The first is the moment Microsoft declares the incident resolved, and whether it stays resolved. The second is the post-incident report the company has promised to publish once service is restored, which should say what failed and what will change. After that come the research notes from analysts, who will ask about service credits and customer reaction, and then Microsoft's next quarterly results, due in late January. The company has guided to revenue of $96 billion to $97.1 billion for the December quarter. If management keeps that range, the financial question is settled.
For anyone holding or trading the stock, the lesson of previous outages is that the headline and the price often point in different directions. Selling a subscription business because its service is down has usually been the wrong reaction, since the revenue does not stop. Assuming that every outage is harmless would be a mistake too. The evidence so far points to an operational failure that is embarrassing, expensive in engineering time and limited in direct cost. What would change that reading is a second day without a fix, a cause that reaches Azure, or a security finding. None of the three had appeared by Friday morning.
Microsoft ended Thursday worth slightly more than it was on Wednesday, with its most widely used products offline. That is the market's first verdict, and the 2024 precedent supports it. It is a verdict on a short outage, though, and this one has already lasted longer than most. Friday's session will show whether investors still see it that way.