AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: AI And CVE Growth: What We Know About Bug Hunters And Fixes on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

A source analysis of vulnerability records found 13,934 new CVE records published from Oct. 1 to Oct. 25, 2026, more than three times the total for October 2025. The records show a sharp rise in disclosures but do not establish how much AI contributed to finding bugs; the analysis also reports that only 516 of 6,157 vulnerabilities disclosed by Anthropic had been patched to its knowledge.

A total of 13,934 new CVE records were published from Oct. 1 through Oct. 25, 2026, according to an analysis of US National Vulnerability Database and CVE.org data. The figure is more than three times the 4,389 records published in all of October 2025, but the records do not show how much artificial intelligence contributed to finding the flaws or how quickly affected software is being fixed.

The analysis, published by Thorsten Meyer on Oct. 26, counted vulnerability records by their publication date and cross-checked October entries against CVE.org. It reported 88,772 records so far in 2026, 78% more than the full-year 2025 total. The October count includes records later rejected: the analysis said 13,766 remained records and 202 had been rejected at the time of checking.

The rise was concentrated among a small number of organizations. Microsoft accounted for 1,455 October records, compared with 180 in October 2025; Oracle rose from 78 to 1,078, and VulnCheck from 163 to 2,085. The analysis linked much of Microsoft’s total to its October security update, which SecurityWeek reported fixed 1,413 CVEs. These are publication counts, not a direct measure of how many flaws were newly introduced or exploited.

The data offers limited evidence about who discovered the bugs. Of October’s 13,934 records, 6,346 named a finder, while 200 explicitly credited an AI lab or AI security company, according to the analysis. That 1.4% is a count of explicit credits, not a reliable estimate of AI’s overall contribution: many records lack a finder, and the published attribution may not capture tools used during research.

At a glance
reportWhen: Data retrieved October 26, 2026; Octobe…
The developmentAn October 26 analysis of public vulnerability data documents a sharp rise in 2026 CVE publications and argues that the pace of fixes, rather than discovery alone, is a central concern.
Thorsten Meyer AI
NVD + CVE.org · data to 25 Oct 2026

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.

13,934new CVE records, 1–25 Oct 2026
3.2×all of October 2025 (4,389)
88,772in 2026 so far vs. 49,972 in all of 2025
+11%daily rate, Oct vs. Sep: growth is cooling

New CVE records per month

January 2024 – 25 October 2026 · phones show the last 13 months

0
5k
10k
15k
3.6k
4.4k
15.1k
13.9k*
Jan ’24Jul ’24Jan ’25Jul ’25O’25N D J’26F M A M J J A S O’26
2024–20252026September 2026 (record)* October 2026: 1–25 Oct only
+18%Jun vs. May
+20%Jul vs. Jun
+28%Aug vs. Jul
+22%Sep vs. Aug
+11%Oct vs. Sep

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

  • MicrosoftOct 2025 → Oct 2026180 → 1,455
  • OracleOct 2025 → Oct 202678 → 1,078
  • VulnCheckOct 2025 → Oct 2026163 → 2,085
  • GitHub (open-source maintainers)Oct 2025 → Oct 2026276 → 1,159
  • Linux kernelOct 2025 → Oct 2026660 → 925

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

Why the patch queue matters

A larger volume of disclosures can help software users understand and address known weaknesses, but publication alone does not protect systems. The analysis says Anthropic reported 6,157 vulnerabilities to maintainers and knew of patches for 516 of them, roughly one in twelve. That figure is limited to Anthropic’s reported findings and its knowledge of patch status; it is not a measure of all vulnerabilities or all fixes.

The gap matters because organizations need time to assess disclosures, develop corrections and deploy them. A rapid rise in records can strain maintainers and security teams, especially when fixes depend on volunteer developers or software vendors. The figures do not establish that the increase has caused more breaches, but they highlight why discovery rates should be read alongside remediation and exploitation data.

The analysis also points to a distinction between finding a vulnerability and attackers using it. It says CISA’s Known Exploited Vulnerabilities catalog grew by 245 entries in 2025, far less than the reported rise in CVE publications. That comparison is not a complete measure of real-world attacks: the catalog covers vulnerabilities CISA identifies as exploited, while a CVE record can describe a flaw without evidence of exploitation.

Products Worth Considering

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDR True Black 400

LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDR True Black 400

As an affiliate, we earn on qualifying purchases.

How the October count was built

Meyer queried the NVD CVE API 2.0 for records published by month from January 2022 onward, using all statuses, and retrieved the data at 07:16 UTC on Oct. 26. The analysis says it checked October IDs against CVE.org. It found that 98% of October records carried a 2026 ID and that the median time between reserving an ID and publication was 15 days, which it used to argue the increase was not mainly a delayed release of older records.

The pace was high but not accelerating at the same rate each month. The analysis reported daily publication rates rising 20% from June to July, 28% from July to August and 22% from August to September. For October through the 25th, it estimated growth of about 11%. Its projection that October could finish near 17,000 records is an estimate, not a final count.

A CVE is a coordinated public identifier and disclosure for a vulnerability, commonly published through a CVE Numbering Authority. The analysis notes that a CVE entry records the disclosure process, not necessarily the identity of the original discoverer. Its AI attribution counts therefore reflect what entries explicitly state, rather than all AI-assisted security work.

“To its knowledge, 516 are patched, about one in twelve.”

— Thorsten Meyer, describing the reported Anthropic disclosures

What the records cannot show

The available records do not establish what share of vulnerabilities were discovered with AI, whether AI materially sped up each discovery, or whether more disclosures reflect better detection, changes in reporting practices, or both. Only 46% of October entries named a finder, and explicit credits may not capture indirect or unreported use of AI tools.

The patch figure also has limits: Anthropic’s count of 516 patched vulnerabilities is based on what the company knew, and the supplied material does not give a comparable patch-status measure across all October CVEs. Nor does the source provide enough detail here to quantify the full October change in confirmed exploitation. The number of published records alone cannot show whether attackers are exploiting more flaws or whether organizations are becoming less secure.

Track fixes and exploitation

The October publication count remains incomplete because it covers only the first 25 days of the month. A final monthly total could show whether the pace continued, slowed or changed after Oct. 25; the estimate of roughly 17,000 should not be treated as an observed result.

For a clearer picture of the security impact, readers will need comparable follow-up data on patched vulnerabilities, time to remediation and confirmed exploitation. Further details from vendors, maintainers and AI security programs could also clarify how many findings were AI-assisted and how those findings translated into public advisories or CVE records.

Key Questions

How many CVE records were published in October 2026?

The analysis counted 13,934 records from Oct. 1 through Oct. 25. It projected a total near 17,000 for the full month if that pace held, but that was an estimate, not a final count.

Does the surge prove AI is finding most vulnerabilities?

No. The analysis found explicit AI-related credits on 200 October records, or 1.4%, but more than half of the records named no finder. The data cannot establish AI’s full contribution.

What does the Anthropic patch figure mean?

Anthropic said it had reported 6,157 vulnerabilities to maintainers and knew of patches for 516. The figure is specific to those reports and reflects Anthropic’s knowledge of patch status.

Do more CVE records mean more attacks?

Not by themselves. CVE counts track published vulnerability records, while exploitation is a separate question. The supplied analysis cites CISA’s Known Exploited Vulnerabilities catalog but does not provide enough evidence to conclude that attacks rose in proportion to CVE publications.

Source: ThorstenMeyerAI.com

You May Also Like

Show HN: I Wrote An Email App That Accepts Only One Sender Per Inbox

Clova.Email says its new email service can create separate addresses that accept mail only from the first sender to claim each inbox.

Inside PATEO’s AI Box Built On NVIDIA DRIVE AGX Thor

A Yahoo Finance headline links PATEO’s in-vehicle AI Box with NVIDIA DRIVE AGX Thor, Nemotron and ByteDance Doubao, but gives few product details.

SenseTime Open-Sources SenseNova-Loop-9B-Preview, A Looped Model That Runs Its Middle Layers Twice – Pandaily

SenseTime says its open-source 9B preview runs its middle layers twice, but available details do not establish its license or performance.

3 Best AI Content Creation Tools In 2027: Pick A Tool For Your Needs

A comparison of three books on AI-assisted content creation, from Canva business visuals to social-media captions and beginner workflows.