Incident sheet · /infrastructure
- System
- Infrastructure & Cloud
- Filed
- Read
- 4 min · 888 words
- Severity
- Routine
Microsoft outage exposes government vendor lock-in risk
A 12-hour Microsoft Outlook outage hit governments across 74 countries, while federal policy on AI, open-source security and code sharing still leaves commercial cloud email outside key protections.

Event log
- The Nov. 12 Microsoft Outlook outage drew nearly 140,000 reports across 74 countries and lasted beyond 12 hours.
- Nvidia was worth about $4.5 trillion before a scandal that erased roughly $192 billion in one trading session.
- OMB memorandum M-25-22 requires AI buyers to address sourcing, interoperability, licensing and portability.
- The SHARE IT Act, signed Dec. 23, 2024, covers custom government code but excludes commercial off-the-shelf products.
- PHP runs on about 69% of websites with identifiable server-side languages; two October RCE flaws scored 9.2 out of 10.
Nearly 140,000 Microsoft Outlook problem reports were logged across 74 countries by the evening of Nov. 12 as an outage that ran beyond 12 hours exposed government dependence on one cloud provider.
At the Metropolitan Police in London, officers lost the ability to take some crime reports online. The disruption had started around 1 p.m. Eastern time on Thursday and continued into Friday.
What went dark?
- The U.S. House of Representatives, Japan, Australia’s spy agency and federal police were affected.
- Michigan warned residents that email replies would be delayed.
- Some organizations reverted to fax machines.
Microsoft blamed server problems and reported no sign of malicious activity. For public agencies, the incident exposed an availability failure with a governance consequence: one provider’s weakness became their operational weakness.
That concentration is vendor lock-in: reliance on one commercial provider so complete that switching becomes impractical or financially prohibitive. It also exposes users to the provider’s outages, unpatched flaws, service throttling and pricing changes.
Capital markets had concentrated heavily in Nvidia. The concurrent honeypot scandal involved CEO Jensen Huang admitting to an affair with a suspected Chinese spy, and the disclosure erased roughly $192 billion from a company worth about $4.5 trillion in one trading session while the broader market rose. Marcum treated the events as parallel, not causal.
Policy already sees the risk
Federal agencies already treat single-vendor dependence as a risk, but key controls still stop short of the commercial cloud services that failed.
OMB memorandum M-25-22, which Christopher Steven Marcum helped develop, tells agencies buying AI to examine vendor sourcing, data portability and long-term interoperability. It also requires them to scope licensing and intellectual-property rights so contracts avoid costly dependence on one vendor.
The memo directs agencies to write lock-in protections into AI contracts: knowledge transfer, portability of data and models, and rights to code and models produced under the contract. Those controls build on the information-governance requirements of OMB Circular A-130.
In July, the Cybersecurity and Infrastructure Security Agency made the security case for open source. Its guidance says agencies can assess code quality directly rather than rely on vendor assurances. It also recommends that contractors give government the rights to reuse, modify and release custom-built code.
The SHARE IT Act, signed into law Dec. 23, 2024, requires agencies to share custom-developed code, including code written by federal employees and produced under federal contracts, so taxpayers do not pay twice for the same software.
Yet its definition excludes commercial computer software and commercial off-the-shelf products. Email, calendars and collaboration suites usually operate as licensed commercial cloud services, placing the systems hit on Nov. 12 outside the statute.
The open-source comparison
During the outage, Microsoft said it was trying different methods to find the cause and taking a measured approach to repair affected infrastructure; affected agencies could only wait. Marcum argues that closed-source delivery denied them independent teams.
Open source would not have prevented a server-side capacity failure on its own, but it would have widened the pool of investigators and allowed agencies to lend staff to the diagnosis.
What the PHP response showed
PHP runs on about 69% of websites whose server-side language can be identified. Two bugs in PHP itself, CVE-2026-109618 and CVE-2026-109619, allowed unauthenticated remote code execution and scored 9.2 out of 10 when disclosed in October.
Because the source code is public, independent teams found them. Ryan Austin, a researcher at security firm Calif, used an OpenAI model to build a working exploit and reported it Aug. 11. OpenAI reported the same root cause; Wiz reported related attack paths later that month.
The PHP Foundation coordinated with researchers, distributors and security vendors, then released patched versions Oct. 21, a day ahead of schedule after details leaked. OpenAI rallied defenders through its Patch the Planet program, and CISA urged immediate patching.
The response was not a clean win. A core developer first saw signs of the memory bug in 2024 while fixing an issue from an outside audit funded by Germany’s Sovereign Tech Agency, but the fix stalled for months. The foundation called this “a structural failure of its decision process and its resources.”
It published a dated postmortem, credited each researcher and released the corrective code. Microsoft customers will learn only what Microsoft chooses to put in its incident report. The comparison, Marcum says, shows what public investment in open-source systems can buy: transparency and a faster response.
What does this change for government IT?
Government IT has made the same one-basket bet with commercial software, routing email, calendars and meetings agency after agency through one provider. Open source spreads risk across maintainers, hosts and independent reviewers, so one failure does not remove every available path.
The practical answer is to extend M-25-22’s sourcing, portability and interoperability protections, plus SHARE IT’s code-reuse logic, to commercial platforms sold to government. Agencies should favor open standards and open-source options that let them move mail when a provider fails.
Marcum, an open-science advocate and former White House information-policy official, put the case plainly: “No single company should be able to take the world offline when it goes down.”
via fedscoop.com (Source)
More from Olivia Hart
Show full bio
News editor covering money, funding and market shifts around enterprise information technology and the leaders who run at CIO Monitor.
207 articles
The newsroom digest
One email a day with the stories that matter. No noise.
Cross-references
- Microsoft 365 Back After Multi-Day Outage Hit Firms in 70+ Countries
- Microsoft 365 outage reportedly affects 100,000 users worldwide
- Microsoft 365 Suffers Broad Outlook Outage Across All Regions
- Microsoft 365 outage exposes recovery gaps after more than 10 hours
- Microsoft's Global Outage: A $7 Billion Lesson in Cloud Concentration Risk