AI Threat Alert
AI Vulnerabilities by Vendor
Browse AI/ML vulnerabilities by the vendor that ships them. AI Threat Alert tracks 1,538 CVEs across 165 vendors — foundation-model providers, framework maintainers, and AI tooling companies including OpenAI, Anthropic, Google, Meta, Microsoft, and Hugging Face — with 220 critical issues, so you can see which vendors carry the most AI security exposure.
- 165Vendors
- 1,538Total CVEs
- 220Critical
Frequently asked questions
Which AI vendors does AI Threat Alert track?
AI Threat Alert tracks 165 vendors that publish AI/ML models, frameworks, or tooling — including OpenAI, Anthropic, Google, Meta, Microsoft, and Hugging Face — ranked by the number of CVEs affecting their products.
How many AI vendor vulnerabilities are tracked?
1,538 AI/ML CVEs across 165 vendors, of which 220 are rated critical.
How are vendors ranked?
Vendors are ranked by total CVE count, with critical-severity issues highlighted, so the vendors with the largest AI security exposure surface first.
What counts as an AI vendor here?
Any organization whose AI/ML models, frameworks, libraries, or tools have tracked vulnerabilities — from foundation-model providers to open-source framework maintainers.
Where does the vendor CVE data come from?
Vendor vulnerability data is sourced from NVD and GitHub Advisory, then mapped to the specific AI/ML packages each vendor maintains.