Nexus Void Research

Time to Exploit, CISA KEV, N-Day, Zero-Day, Vulnerability Management, Patch Management, EPSS, CVSS, Original Research

Average Time to Exploit a Vulnerability in 2026: KEV Data

The average time to exploit a vulnerability in 2026 is 5 days (median) per our analysis of 297 CISA KEV CVEs. 27% were exploited by disclosure day.

The average time to exploit a vulnerability in 2026 is measured in days, not months. Of the 219 CVEs published in 2026 that CISA has confirmed as exploited, the median went from public disclosure to a Known Exploited Vulnerabilities listing in 5 days. 27% were listed on or before the day the CVE was published in NVD, and 81% within 30 days.

That is our analysis of every CVE added to the CISA Known Exploited Vulnerabilities (KEV) catalog between January 1 and November 13, 2026, matched against publication dates in the National Vulnerability Database (NVD). It updates our 2025 study of the disclosure-to-exploitation window, using the same method.

What is "time to exploit" for a vulnerability?

Time to exploit is the gap between a vulnerability becoming public and attackers using it in real attacks. It is the window defenders have to patch before exploitation starts. Nobody sees the exact moment of first exploitation, so researchers use proxies. We use the date CISA adds a CVE to KEV, which requires reliable evidence of active exploitation. That makes our numbers an upper bound: real exploitation often starts before CISA lists it.

A related term is the n-day vulnerability: a flaw that is already disclosed, and often already patched, but still exploitable on systems that have not applied the fix. Most KEV entries are n-days by the time they are listed. The zero-day cases are the ones listed on or before disclosure.

What is the average time to exploit a vulnerability in 2026?

For CVEs both published and confirmed exploited in 2026, the median is 5 days and the mean is 20.5 days. The mean is pulled up by a long tail of flaws exploited months after disclosure. The middle half of cases fall between 0 and 18 days.

Days from NVD publication to KEV listing

CVEs (published 2026)

Share

On or before publication (0 or less)

60

27.4%

1 to 7 days

70

32.0%

8 to 30 days

47

21.5%

31 to 90 days

28

12.8%

More than 90 days

14

6.4%

Total

219

100%

Six in ten exploited 2026 CVEs were confirmed exploited within a week of disclosure. A monthly patch cycle cannot keep up with that, and an annual test cannot see it at all.

Is time to exploit getting shorter?

Across all KEV additions, yes: the median fell from 26 days in 2025 to 12 days in 2026. But be careful with that headline. Most of the drop comes from CISA adding fewer old vulnerabilities this year (26% of 2026 additions were published before 2026, against 33% in 2025). For newly published CVEs, the speed was already extreme last year.

Metric

2025

2026 (to Nov 13)

CVEs added to KEV

245

297

Median days, all additions

26

12

Listed on or before disclosure, all additions

20.0%

20.2%

Median days, CVEs published that year

4

5

Listed within 7 days, CVEs published that year

58.8%

59.4%

Listed within 30 days, CVEs published that year

75.2%

80.8%

The honest reading is that the window for new vulnerabilities was already about a week in 2025 and has stayed there, while the share caught within 30 days has grown. What is changing is how attackers fill that week. In November 2026, Google reported attackers using AI agents to turn patches into exploits, including a framework that ran discovery and exploitation with no human involved, and Project Zero began cutting its disclosure deadline toward 30 days.

Does a higher CVSS score mean faster exploitation?

Not reliably. Among exploited 2026 CVEs with a CVSS score, critical flaws (9.0 to 10.0) had a median of 5 days to KEV listing, high flaws (7.0 to 8.9) had 3 days, and medium flaws (4.0 to 6.9) had 10 days. 12% of exploited 2026 CVEs scored below 7.0, so a "fix criticals only" rule would have skipped them. Severity measures impact if exploited, not the likelihood of exploitation, which is why CVSS misses so much.

CVSS band (2026 CVEs)

Exploited CVEs

Median days to KEV

Critical (9.0 to 10.0)

108

5

High (7.0 to 8.9)

85

3

Medium (4.0 to 6.9)

25

10

Vulnerabilities CISA marks as used in ransomware campaigns (27 of 2026's additions) had a median of 15 days across all additions, a reminder that ransomware crews often work through n-days after patches ship.

Which vulnerabilities do attackers exploit fastest?

The vendors with the most 2026 KEV additions were Microsoft (41), Cisco (21), Apple (10), Google (10), Linux (9), Fortinet (9), Ivanti (8) and Adobe (7). Among 2026 CVEs, the most common weakness types were OS command injection (CWE-78, 13 CVEs), code injection (CWE-94, 11), improper authentication (CWE-287, 10), path traversal (CWE-22, 9), deserialization of untrusted data (CWE-502, 8) and authentication bypass through an alternate path (CWE-288, 8).

The pattern is internet-facing systems with low-complexity, often unauthenticated flaws: edge devices, VPNs, remote management tools and web applications. These are exactly the bugs that are easiest to reverse-engineer from a patch.

How fast should you patch exploited vulnerabilities?

Within days, and ideally before CISA's own deadline. Each KEV entry carries a due date for US federal agencies, now set under CISA's BOD 26-04. Of 2026's 297 additions, 159 had a deadline 3 days after listing, 91 had 14 days and 44 had 21 days. Add the 5-day median before a CVE is even listed, and a team on a monthly cycle can be exposed for several weeks after exploitation is confirmed. A practical order of work:

  1. Patch anything in KEV that is internet-facing first, within days, regardless of CVSS.

  2. Rank the rest with EPSS, the Exploit Prediction Scoring System, which estimates the probability of exploitation in the next 30 days. Our EPSS guide explains the thresholds.

  3. Watch vendor advisories for your edge stack daily, not monthly. Most fast exploitation hits a small set of vendors.

  4. Verify after patching. A fix that was not applied everywhere, or that left a backdoor behind, does not close the window.

Our vulnerability prioritization guide walks through combining CVSS, EPSS and KEV.

Why does time to exploit matter for penetration testing?

Because a test only sees the vulnerabilities that exist on the day it runs. If most exploited flaws go from disclosure to attack in a week, an annual or quarterly pentest leaves months where new exposure goes untested. Our 2025 analysis found an annual schedule would have missed 67% of actively exploited vulnerabilities. Continuous testing closes that gap by checking your exposure as new vulnerabilities land. Agentic VAPT runs that testing on an ongoing cadence, and Nexus Void offers continuous VAPT for web apps, APIs, cloud and networks.

Methodology

We pulled every CVE with a CISA KEV date-added between January 1 and November 13, 2026 from the NVD API (297 CVEs) and took the NVD publication date for each. Time to exploit is the KEV date-added minus the NVD publication date, in days. "Published in 2026" means an NVD publication date in 2026 (219 CVEs). CVSS is the NVD base score (v3.x or v4.0). Vendor and ransomware fields come from CISA's KEV feed. The 2025 comparison uses the same method on the 245 CVEs added to KEV in 2025. Limits: KEV listing lags real first exploitation, KEV is US-government focused and not exhaustive, and NVD publication can trail a vendor's own advisory by a day or more. The full dataset is available on request.

Frequently asked questions

What is the average time to exploit a vulnerability?

For CVEs published and confirmed exploited in 2026, the median time from public disclosure to CISA KEV listing is 5 days and the mean is 20.5 days. Across all 2026 KEV additions, including older CVEs, the median is 12 days.

What percentage of vulnerabilities are exploited before disclosure?

27.4% of exploited CVEs published in 2026 were added to KEV on or before their NVD publication date, meaning exploitation was confirmed by the time most defenders heard about them. Across all 2026 additions the figure is 20.2%, the same as 2025.

What is an n-day vulnerability?

An n-day vulnerability is a flaw that has been publicly disclosed, and usually patched, but is still exploitable on systems that have not applied the fix. Most exploitation in the KEV catalog targets n-days.

Is time to exploit getting faster in 2026?

The median across all KEV additions fell from 26 days in 2025 to 12 days in 2026, mostly because fewer old CVEs were added. For newly published CVEs it was already about 4 to 5 days in both years.

How quickly should I patch a vulnerability in CISA KEV?

Within days for anything internet-facing. Our data shows a median of 5 days from disclosure to confirmed exploitation for 2026 CVEs, so a monthly patch cycle leaves a long exposure window.

DATA SOURCES

CISA KEV catalog: https://www.cisa.gov/known-exploited-vulnerabilities-catalog ; NVD API 2.0: https://nvd.nist.gov/developers/vulnerabilities ; FIRST EPSS: https://www.first.org/epss/ ; Nexus Void analysis of 297 KEV additions, Jan 1 to Nov 13, 2026

Liked this post? Share it: