InformationalVulnerabilityOngoing

Microsoft analysis: 78% of exploited vulnerabilities are system takeover flaws, and patching lags

Microsoft found that about 0.5% of disclosed CVEs are exploited, 78% of those enable system takeover, and patching on servers and appliances lags far behind

Published
Updated
Exploitation
Yes, Microsoft links 2,056 CVEs to real-world exploitation, but the analysis names no single new vulnerability

What happened

On November 2, 2026, Microsoft Security Research published a data-driven analysis of which vulnerabilities attackers actually exploit Microsoft Security Blog. It is a research study, not a report of a new flaw or breach. Key findings:

  • From 1999 through September 2026, about 383,000 software vulnerabilities were disclosed, yet only about 2,000 CVEs (roughly 0.5%) have been linked to real-world exploitation Microsoft Security Blog. The exploited set is 2,056 CVEs GBHackers.
  • Takeover classes (remote code execution, authentication bypass and local privilege escalation) account for 78% of exploited vulnerabilities Microsoft Security Blog.
  • Server-side RCE rises from 20% of published CVEs to 37% of exploited CVEs. Local privilege escalation rises from 11% to 18%, and client-side RCE from 4% to 13%. Cross-site scripting falls from 15% to 4%, and denial of service from 16% to 7% Microsoft Security Blog.
  • OS and system code is 23% of all CVEs but 41% of exploited ones. Application code is about 77% of all CVEs Microsoft Security Blog.
  • Microsoft Security Exposure Management data from 2026 shows the median vulnerability is patched in about two weeks on Windows workstations and in more than a month on Windows Servers. It takes 10+ weeks on Linux servers and for macOS applications, and more than a year on network appliances Microsoft Security Blog.
  • Exploitation moves faster. According to VulnCheck, 29% of exploits in 2025 were launched within 1 day of CVE publication. According to CISA, the mean time from CVE publication to the KEV catalog fell from 89 days in 2024 to 57 days in 2025. For 2026 through September it is 20 days, with a median of 3 days Microsoft Security Blog.
  • A separate account of VulnCheck’s State of Exploitation 2026 says 884 vulnerabilities had first-time exploitation evidence in 2025, and 28.96% of them were exploited on or before their CVE publication date GBHackers.

Microsoft also cautions that a stated classification does not always match what an attacker can reliably do. Some flaws described as RCE may only cause denial of service in real conditions Microsoft Security Blog. GBHackers adds that publication-to-listing intervals measure reporting and curation processes, not necessarily when attackers first weaponized a flaw GBHackers. Microsoft warns that AI-assisted vulnerability discovery and exploit development could further shorten defenders’ response window Microsoft Security Blog.

No specific CVE IDs, affected versions, fixes or indicators of compromise are published in the analysis. One report notes that the study lists none Cyber Security News.

Am I affected?

This is guidance on prioritization, not a flaw in one product, so there is no affected version list (not confirmed). Your risk depends on exposure:

  • Servers and network appliances are the most exposed, because they are the critical assets and the slowest to patch. The gap between disclosure and remediation is the attacker’s window, and it is widest on critical systems that stay exposed the longest Microsoft Security Blog.
  • Servers often use service accounts, application identities or cloud-managed identities with more permissions than they need. Compromising one can give access to more resources, sensitive data, administrative privileges or trusted network paths Microsoft Security Blog.
  • Legacy or unsupported systems that cannot be patched need compensating controls Microsoft Security Blog.
  • Availability-critical assets and exposed credentials or secrets also deserve attention where the business context makes the impact severe Microsoft Security Blog.

What to do now

  1. Rank open vulnerabilities by demonstrated attacker utility and exposure, not by volume or severity score alone. Treat remote code execution, authentication bypass and local privilege escalation as the first tier GBHackers.
  2. Check what an attacker can actually do with a flaw before you rank it, since a label such as RCE may only produce a crash in practice Microsoft Security Blog.
  3. Shorten patch times on servers, Linux hosts and network appliances, where the medians are a month, 10+ weeks and over a year respectively Microsoft Security Blog.
  4. Inventory your internet-facing assets, including shadow IT and third-party dependencies. Microsoft points to Defender External Attack Surface Management for this Cyber Security News.
  5. Rank weaknesses using exploit signals, internet exposure and asset value. Microsoft suggests Defender Vulnerability Management for this Cyber Security News.
  6. Map attack paths from exposed servers through privileged identities to sensitive data, then cut those routes Cyber Security News.
  7. Where patching is delayed or impossible, remove unnecessary exposure and restrict excessive privileges to limit follow-on compromise GBHackers.
  8. Review service accounts and cloud identities on servers and trim permissions beyond what the workload needs Microsoft Security Blog.

Sources