Back to Flash News

Microsoft sets a record by fixing about 1,400 security vulnerabilities in a single month

Microsoft fixed about 1,400 security vulnerabilities in its monthly security update last month, setting a single-month record; about 170 in the same period a year earlier. The company attributed the surge in vulnerabilities discovered to the assistance of artificial intelligence tools. Microsoft Chief Executive Officer Satya Nadella said on November 18 that AI helps find and fix vulnerabilities faster, but the resulting patching may increase system instability, especially for older infrastructure.

Nadella pointed out that recent outages in services such as Microsoft 365 were caused by potential defects being triggered under specific traffic patterns, and acknowledged that testing should have been completed earlier. He warned that AI’s acceleration of vulnerability discovery may force companies to deploy patches more quickly, thereby introducing new risks on older systems. Several media outlets reported that in 2026 Microsoft had fixed more than 5,000 vulnerabilities in total for the year, far above previous years' levels.

Original source:

  1. Title: How the surge in AI-discovered bugs is straining Microsoft security and threatening tech stability
    Publisher: GeekWire
    Publication date: 2026-11-20
    URL: geekwire.com/2026/how-the-surge-in-ai-discovered-bugs-is-straining-microsoft-security-and-threatening-tech-stability
    Source nature: media report
    Verification status: This read the relevant content, verification date 2026-11-20
    Supporting content: Supports the monthly record of about 1,400 vulnerabilities, the comparison with the previous year, the surge in AI-discovered findings, and Nadella’s comments about patches causing instability.

ABAB AI Insight

The number of vulnerabilities rose from about 1,180 for all of 2025 to more than 1,400 in a single month in 2026, with the change starting after a surge in security researchers’ reports after April 2026. Microsoft publicly attributed this to discovery assisted by large language models. The number of vulnerabilities actually exploited did not increase by the same magnitude; most of the new findings were low- to medium-severity flaws. The larger patch size increased the burden on enterprises for testing and deployment.

Nadella’s comments came after recent Microsoft 365 outages. He described the outages as a perfect storm: potential defects were triggered under specific traffic and capacity conditions, rather than being directly caused by AI code. He also said that AI has sped up the pace of discovery and exploitation, forcing patches to go live immediately and potentially reducing the adequacy of testing. A comparable case is that, in history, large software vendors have seen patch management become an operational bottleneck after vulnerability disclosure accelerated.

On the capital and operational path, Microsoft is integrating AI into security and product processes. Higher patch speed may improve short-term remediation capability, but it increases the risk of cascading failures on older systems that are not sufficiently isolated. Some critical infrastructure, such as firmware or industrial control systems, is described as difficult or impossible to patch. Historical patterns show that advances in discovery tools often precede the maturity of remediation processes, creating a temporary window of instability.

ABAB News · Cognitive Laws

  1. AI-accelerated vulnerability discovery can increase the total number of fixes, but it does not automatically raise the share of vulnerabilities that are actually exploited.

  2. There is a trade-off between faster patching and the stability of older systems; rapid deployment without sufficient testing may introduce new failures.

  3. A single-month vulnerability record must distinguish between improved discovery capability and changes in threat activity; the two do not necessarily move in tandem.