331,341 vulnerabilities.
Here's what actually matters.

Software vulnerabilities are flaws that attackers can exploit to break into systems. Thousands are discovered every year. Most of the data about them is incomprehensible. This page changes that.

Scroll to explore the data
0
New vulnerabilities discovered in 2025 — a record-setting year
0.44%
Of all vulnerabilities are in CISA's Known Exploited list
0 days
Median time between discovery and CVE publication
0
Known exploited vulnerabilities in CISA's KEV catalog

01 — The Flood

Vulnerabilities are growing exponentially

Every year, researchers and companies discover more software flaws. A "CVE" is the ID number assigned to each one. The trend line tells a clear story: we're finding more flaws, faster.

38,951
CVEs published in 2025 — the highest single-year total ever recorded, up from 27,463 in 2023
42%
Increase from 2023 to 2025. The flood is accelerating, not slowing down.
331,341
Total CVEs since 1999. The first 100,000 took 17 years. The last 100,000 took just 3.

02 — Who's Filing

MITRE used to assign every CVE. Not anymore.

As the volume exploded, MITRE authorized companies (called CNAs — CVE Numbering Authorities) to assign their own CVE IDs. Today, companies assign 88% of all CVEs.

MITRE is the organization that originally managed CVE assignment. As the volume exploded, they authorized companies (called CNAs — CVE Numbering Authorities) to assign their own. CNAs include Microsoft, Google, Red Hat, and hundreds of others.

88%
Of CVEs in 2025 were assigned by CNAs (companies), not MITRE. The system has fundamentally decentralized.
2010
The year when non-MITRE CNAs surpassed MITRE in CVE assignments. The crossover has only accelerated since.

03 — Severity Spectrum

Not all vulnerabilities are created equal

Each vulnerability gets a severity score from 0 to 10. The score measures how much damage it could cause if exploited. Click a category below to learn what each level means.

Critical: 5%
High: 52%
Medium: 33%
Low: 10%
Critical
~1,950
5% of scored CVEs
High
~20,250
52% of scored CVEs
Medium
~12,850
33% of scored CVEs
Low
~3,900
10% of scored CVEs

Critical (Score 9.0 – 10.0)

These are the most dangerous flaws. An attacker could take full control of your system remotely, often without needing a password or any interaction from you. Think: someone can access your entire computer from across the internet. These get emergency patches.

High (Score 7.0 – 8.9)

Serious vulnerabilities that could let attackers steal data, crash systems, or gain significant access. They usually require some preconditions (like being on the same network or tricking you into clicking a link), but the damage potential is real.

Medium (Score 4.0 – 6.9)

Moderate-risk flaws. These typically need multiple conditions to exploit and cause limited damage. By themselves they're not emergencies, but attackers sometimes chain several medium-severity bugs together to build a full attack.

Low (Score 0.1 – 3.9)

Minor issues with minimal direct impact. An attacker would gain very little from exploiting these alone. They're worth fixing eventually, but they don't keep security teams up at night.


04 — Score Reality

What the score distributions actually look like

CVSS scores aren't evenly distributed. Exploited vulnerabilities cluster at higher scores — but not exclusively. Some medium-severity bugs are heavily targeted too.

Score Distribution

When you toggle to "Known Exploited," the distribution shifts dramatically toward higher scores — but not exclusively. Some medium-severity vulnerabilities are heavily exploited too.

Average Score Trends

6.71
Average CVSS score for all scored CVEs. Scores have stabilized around 6.7 since 2018.
8.33
Average CVSS score for CISA KEV vulnerabilities — consistently higher than the overall average.

05 — The Scoring Gap

Thousands of vulnerabilities have no score at all

The National Vulnerability Database (NVD) fell behind processing the flood of CVEs. You can't prioritize what you can't measure.

3,991
Unscored CVEs in 2025 — down dramatically from near-100% unscored before 2017
7.7%
Of 2024 CVEs had no severity score — a massive improvement from 92.5% unscored in 2017

06 — The Government's List

CISA tracks what's actually being exploited

CISA (the US Cybersecurity and Infrastructure Security Agency) maintains a list of vulnerabilities that are KNOWN to be actively exploited. Federal agencies are required to patch these.

The high numbers in 2021–2022 are because the catalog launched in November 2021 with a large backfill of historically exploited vulnerabilities. Ongoing additions are lower but steady.

1,463
Total KEVs in CISA's catalog — the definitive list of "known bad" vulnerabilities
209
KEVs in 2021 — the peak year, when CISA launched the catalog with a large backfill
3
New KEVs added per week on average since the catalog launched in 2021
07 — Critical vs Exploited

"Critical" doesn't mean "exploited"

A severity label measures potential damage, not likelihood of attack. Most critical CVEs are never exploited. And many exploited CVEs aren't rated Critical.

100 Critical CVEs — how many are actually exploited?

100 Exploited CVEs — what severity were they?

3.5%
Of critical CVEs end up on the KEV list. The other 96.5% are never confirmed as actively exploited.
53.4%
Of KEV entries are High severity, not Critical. Attackers choose practical targets, not just dangerous ones.
08 — Active Exploitation

Exploitation is rising — but PoC code rises faster

Proof-of-concept exploit code exists for far more vulnerabilities than are actually exploited in the wild. Having a recipe doesn't mean anyone will cook the dish.

209
Peak year for actively exploited CVEs was 2021, when CISA launched the KEV catalog.
7,595
CVEs with exploit references in 2025 — up dramatically from 276 in 2021. 47x more than actively exploited.

09 — The Race

The publication delay problem

When a vulnerability is discovered, it takes time before it gets an official CVE publication. That delay leaves a gap where the flaw may be known but untracked.

35
days median
Half of all CVEs are published within 35 days of discovery. But the other half can take much longer.
126
days mean
The average is pulled up to 126 days by a long tail of delayed publications. 10% take over 296 days.
Day 0 – 35
Median — 50% published
Half of all CVEs are published within 35 days of discovery. This is the fastest half of the pipeline.
Day 36 – 126
Mean — average point
The average publication point, skewed upward by outliers. Most CVEs in this range are still being processed.
Day 127 – 296
90th percentile
10% of CVEs still haven't been published by this point. These are the stragglers in the system.
Day 296+
Long tail
Some CVEs take years to be officially published. The longest delays stretch to over 6,500 days.

Publication Delays

How long between a vulnerability being discovered and its CVE being published?

50% of CVEs are published within 35 days, but 10% take over 296 days. The mean delay of 126 days is dragged up by these outliers.


10 — The Scoring System

CVSS has evolved — and scores aren't comparable across versions

Comparing scores across CVSS versions is like comparing Fahrenheit to Celsius — the numbers mean different things.

CVSS v2 (2007–2019)

The original scoring system. Simpler formula, fewer inputs. Tended to cluster scores around certain values. Officially retired.

CVSS v3 (2015–present)

Added "Scope" and refined attack complexity. Better at distinguishing severity levels. Still the dominant version in use today.

CVSS v4 (2023–present)

The newest version. Adds granularity with supplemental metrics and better reflects modern attack patterns. Adoption is growing but still early.


11 — Scoring Quirks

The formula has "favorite numbers"

CVSS scores aren't evenly distributed. The formula's math creates spikes at certain scores — 7.5 and 9.8 appear far more often than 7.4 or 9.7.

The CVSS formula uses discrete inputs (like "Low/High" for complexity) that collapse into certain numeric outputs. This means some scores are mathematically impossible to reach, creating gaps and spikes in the distribution.


12 — Predicting Attacks

EPSS predicts what will actually be exploited

EPSS (Exploit Prediction Scoring System) tries to predict which vulnerabilities will actually be exploited in the next 30 days. Unlike CVSS which measures theoretical damage, EPSS measures real-world likelihood.

70%
Of all vulnerabilities have less than a 1% chance of being exploited in the next month. Most CVEs are noise.
3%
Of vulnerabilities have a 50%+ chance of exploitation. EPSS helps focus on the small percentage that actually matter.

13 — Severity vs Prediction

High severity does not mean high exploitation

The upper-left quadrant (high severity, low exploitation probability) is full. Many "critical" vulnerabilities are never exploited. The lower-right deserves more attention than it gets.

Low EPSS
Med EPSS
High EPSS
High CVSS
8,000
High severity, rarely exploited
1,200
Watch list
500
Top priority
Med CVSS
15,000
Low priority
3,500
Monitor
200
Underrated threats
Low CVSS
12,000
Ignore safely
800
Unusual
50
Surprising exploits

The red cells are where security teams should focus. The large gray cell (8,000 high-severity, low-exploitation) represents wasted urgency — "critical" CVEs that nobody actually attacks.


14 — Attack DNA

Attackers choose easy targets

Attackers overwhelmingly target vulnerabilities that are network-accessible, low complexity, no authentication required, no user interaction needed.


15 — Who's Watching

No single tracker catches everything

Different organizations track exploited vulnerabilities. They don't all agree on what to include.

Tracker Comparison

VulnCheck tracks 3x more exploited vulnerabilities than the US government's official list. All 1,463 CISA KEVs are included in VulnCheck's broader set of 4,395.

Top Vendors by Exploited Vulnerabilities

Microsoft products account for the most exploited vulnerabilities — not because they're less secure, but because they're the most widely deployed. Attackers go where the users are.


16 — Experts Disagree

NVD and CISA score the same bugs differently

When both NVD and CISA ADP/CNA score the same vulnerability, they agree exactly 84% of the time. But for the remaining 16%, the disagreements can be significant.

84%
Exact agreement — when both NVD and CISA ADP/CNA score a CVE, they match 84% of the time
5%
Differ by more than 1 point. Most disagreements are small.
93.1%
NVD coverage — NVD has scores for 93.1% of all CVEs (299,632 of 321,690)

17 — Global Coverage

Different countries, different databases

Different countries maintain their own vulnerability databases. None of them cover everything.

The European EUVD covers 91% of all CVEs. Japan's JVN covers 82%. Russia's BDU covers just 21%. ENISA's EU KEV is tiny (only 19 entries) but 84% of those overlap with CISA's list.

91.2%
EUVD (EU) overlap with CVE. Europe's database covers 302,174 of 331,341 CVEs.
84%
ENISA EU KEV overlap with CISA KEV. But ENISA only tracks 19 entries vs CISA's 1,463.

18 — Quality Control

The CVE system is straining under the load

As CVE volume has exploded, so have quality problems. Over 17,000 CVEs have been rejected — duplicates, errors, and non-vulnerabilities that slipped through the system.

1,660
Rejected CVEs in 2023 — the peak year. These are duplicates, errors, and non-vulnerabilities.
127
Disputed CVEs in 2024 — cases where vendors disagree it's a real vulnerability
5.2%
Overall rejection rate across all CVEs. 17,205 of 331,341 CVEs have been rejected.

19 — The Ransomware Connection

1 in 5 exploited vulnerabilities lead to ransomware

Ransomware groups specifically target known vulnerabilities to break in. Here's how the numbers break down.

304
Known exploited vulnerabilities used by ransomware groups — 20.5% of all KEVs
24
New vulnerabilities added to the KEV list in 2025 were linked to ransomware attacks
245
Total new entries added to CISA's KEV catalog in 2025, a 20% growth year-over-year
94
Older vulnerabilities (from 2024 and earlier) retroactively added to KEV in 2025

20 — What It All Means

Six things everyone should know

01

Volume is not risk

38,000+ new vulnerabilities sounds terrifying, but less than 0.5% are ever exploited. The real skill is knowing which ones matter.

02

Speed is everything

The window between disclosure and exploitation has collapsed from a month to days. Fast patching is a survival skill.

03

Old bugs don't die

94 vulnerabilities added to the government's "known exploited" list in 2025 were from previous years. If it's unpatched, it's useful.

04

Severity scores lie

CVSS measures theoretical damage, not real-world likelihood. Only 3.5% of "critical" CVEs are ever actually exploited.

05

No one sees the whole picture

CISA, VulnCheck, NVD, and international databases all disagree. No single source tracks every exploited vulnerability.

06

The system is straining

Over 17,000 CVEs have been rejected. Scoring coverage only recently improved. The infrastructure for tracking vulnerabilities is catching up but still strained.