🔍 Read the full analysis: The Patch Gap Behind The CVE Climb: How AI Enters The Picture on ThorstenMeyerAI.com
TL;DR
A report published Oct. 26 counts 13,934 CVE records published from Oct. 1–25, 2026, more than three times the total for all of October 2025. The data does not establish how much AI contributed to the increase, while Anthropic says it knows of patches for 516 of 6,157 vulnerabilities it reported to maintainers.
A tally of vulnerability disclosures found 13,934 new CVE records published from Oct. 1–25, 2026, more than three times the 4,389 records for all of October 2025. The Oct. 26 analysis says the figures show a steep rise in disclosures, but do not establish how much artificial intelligence contributed; it identifies the gap between finding bugs and getting them patched as a central risk.
The analysis author says the tally came from the U.S. National Vulnerability Database (NVD) and was cross-checked against CVE.org records. The sources agreed on 13,766 published records for October as of the review, with 202 additional records later rejected. The separate count of 13,934 covers the period through Oct. 25 and is the figure used for the comparison with October 2025. The report says 98% of October records carried a 2026 identifier, and the median time between reserving an identifier and publication was 15 days.
The increase is concentrated among a small number of organizations, according to the report. Microsoft’s October count rose from 180 records in October 2025 to 1,455; Oracle’s rose from 78 to 1,078, most published on Oct. 20. GitHub rose from 276 to 1,159, while VulnCheck rose from 163 to 2,085. The Linux kernel count increased from 660 to 925. The author attributes Microsoft’s figure in part to its October Patch Tuesday, which SecurityWeek reported addressed 1,413 CVEs.
The records offer limited evidence about who discovered the vulnerabilities. Of the 13,934 October records reviewed, 6,346 named a finder, while 200—1.4%—explicitly credited an AI lab or AI security firm. The report cautions that this is a floor, not a full measure of AI’s contribution: many records do not name a finder, and disclosure records are not designed to track every tool used in discovery.
Infographic 01 · How big the CVE wave really is
The CVE chart is real. Steep, not parabolic.
New vulnerability records have tripled year over year. A handful of big vendors and clearinghouses drive the jump, and the growth rate cooled in October.
New CVE records per month
January 2024 – 25 October 2026 · phones show the last 13 months
Growth of new CVEs per day versus the previous month. Steep, but the acceleration slowed in October: not „parabolic“.
Who publishes the jump
records by CVE Numbering Authority, October vs. October
Microsoft’s October Patch Tuesday fixed a record 1,413 CVEs (SecurityWeek (14 Oct 2026)). The Linux kernel grew far less.
Not a backlog dump
checks on the October 2026 records
- 98% carry a 2026 ID
- 15 days median from reserving an ID to publishing it
- 217 numbering authorities published, vs. 163 in Oct 2025
- 202 records were later rejected (counted, as NVD does)
- ~17,000 if October’s pace holds: a projection, not a count
Sources: NVD CVE API 2.0 (counts by publication date, all statuses, queried 26 Oct 2026); CVE.org records (credits field, fetched 26 Oct 2026); SecurityWeek (14 Oct 2026). Own analysis; raw files and queries in the kit.
Powered by Thorsten Meyer AI
The Fixing Gap Shapes Risk
A growing disclosure count means more known flaws are being documented, but it does not by itself show that attackers are exploiting them or that software has become less secure. The practical question is whether maintainers and vendors can ship and deploy fixes before attackers take advantage of vulnerable systems.
Anthropic’s disclosure dashboard, as summarized in the report, listed 6,157 vulnerabilities reported to maintainers and 516 known to be patched as of the company’s reporting. That is about one in twelve, based on Anthropic’s knowledge; it is not an independently verified measure of all fixes. Still, the gap illustrates why discovery totals alone are an incomplete measure of security outcomes. A vulnerability may be reported, fixed, and then remain unpatched on systems that have not received the update.
The report also cites a rise in attackers’ use of inexpensive AI agents against older bugs, attributing that account to NorForge and Reuters. That points to a concern distinct from the number of new CVEs: AI may help defenders find weaknesses, while also making it easier to search for or exploit known flaws. The available figures do not quantify the balance between those effects.
Popular right now

Dyson V9 Motorbar Cordless Vacuum

Bose QuietComfort Ultra Headphones Gen 1, Lunar Blue

De'Longhi Classic Signature Espresso Machine with Adjustable Milk Frother
As an affiliate, we earn on qualifying purchases.
How the October Count Was Built
The report queried the NVD CVE API 2.0 for records by publication date, covering monthly totals from January 2022 onward. It says the data was retrieved on Oct. 26 at 07:16 UTC, included all record statuses, and was cross-checked against CVE.org for October. These counts refer to CVE records published during the period, not necessarily to vulnerabilities first discovered during that same period.
A CVE is a public identifier and disclosure record for a vulnerability. The report explains that records are generally published through CVE Numbering Authorities, often a vendor, after a fix or a disclosure deadline. A criminal who finds a vulnerability does not file a CVE; a record may instead reflect a defender’s later disclosure or response. That distinction means a rise in CVEs is not a direct count of attacker discoveries or attacks.
The monthly pace in the report rose from 8,001 records in June to 9,919 in July, 12,719 in August and 15,074 in September. The author calculates monthly growth rates of 20%, 28% and 22% through September, then about 11% for October so far. A projection that October could end near 17,000 records is the author’s estimate if the pace continues, not a final count.
„29,439 candidate vulnerabilities found by Claude models.“
— Anthropic, as described in its disclosure dashboard
What the Records Cannot Show
The October data cannot establish how much AI contributed to the overall CVE rise. Only 1.4% of records explicitly credited an AI lab or security firm, but more than half named no finder. The report says AI attribution in CVE records is therefore incomplete; its 200 credited records should not be treated as a full count of AI-assisted discoveries.
The patch figure also has limits. Anthropic says 516 of the 6,157 vulnerabilities it reported were patched to its knowledge, but the analysis does not independently verify that status or establish how many of the remaining reports are still unpatched, under review, or fixed without notice. The supplied source material also ends before giving the full comparison for CISA’s Known Exploited Vulnerabilities catalog. It says the catalog grew less than the CVE count, but does not provide the relevant later total or enough detail to assess the comparison.
It remains unclear how many of the newly recorded vulnerabilities are exploitable in real-world conditions, how many have been fixed across deployed systems, and whether AI use by attackers has changed the rate of confirmed exploitation. The source’s estimate of about 17,000 October records is a projection, not a completed monthly count.
Patch Status and October Totals
The October figure can be checked against the completed month once the remaining records are published and the databases are updated. The report’s projected total of roughly 17,000 CVEs should be treated as an estimate until then. Further disclosure by vendors and maintainers may also change the recorded totals or patch-status counts.
For readers tracking exposure, the next useful developments are confirmed patch releases, details from maintainers about outstanding reports, and updated entries in public exploitation catalogs. Those measures can help clarify whether the sharp rise in disclosures is accompanied by more active exploitation. The current report establishes a surge in published records; it does not settle the scale of AI’s role or the resulting risk to systems.
Key Questions
How many CVE records were published in October 2026?
The report counts 13,934 records from Oct. 1–25. Its cross-check found 13,766 October records in the NVD and CVE.org data, plus 202 later-rejected records; the author’s date range and record-status handling explain why the figures are presented separately.
Does the increase prove AI caused the CVE surge?
No. The report says 200 records, or 1.4%, explicitly credited an AI lab or AI security firm, while 54% named no finder. Those records cannot measure every instance of AI-assisted discovery.
What does the patch figure mean?
Anthropic’s dashboard listed 516 of 6,157 reported vulnerabilities as patched to the company’s knowledge. The report does not independently verify the status or show that every other report remains unfixed.
Is the October total of 17,000 confirmed?
No. It is the author’s projection if the pace through Oct. 25 continues. The final monthly count was not available in the report.
Source: ThorstenMeyerAI.com