Researchers counted nearly 268,000 systems, but the findings do not establish whether individual deployments are exploitable.
AI-related infrastructure exposed to the public internet increased 63% over nine months, according to Censys data shared with Axios.
Censys identified nearly 268,000 internet-facing systems running 44 AI-related products as of July, up from roughly 164,000 in October 2025. The products include software connecting AI models to databases, corporate applications and other services.
Workflow automation platform n8n accounted for nearly 114,000 deployments. About 77,000 were running versions affected by at least one recently disclosed critical vulnerability, Censys found.
Researchers also identified more than 54,000 exposed control panels for OpenClaw, an open-source agent that can connect to personal accounts and other services.
The findings do not establish whether the individual deployments can actually be exploited, Axios noted.
Censys principal researcher Emily Austin said the tools’ connections to sensitive systems make their exposure a security concern. CEO Zakir Durumeric said employee-built AI applications compound companies’ existing difficulties tracking software.
Austin told Axios that businesses should regularly inventory, update and secure AI infrastructure as they do other software.
