Censys data, shared first with Axios, indicates that the amount of AI-related infrastructure visible on the public internet rose by 63% over nine months. According to the report, nearly 268,000 internet-facing systems were running AI-related products by July 2026, up from roughly 164,000 in October 2025.
Why this matters for companies
The AI boom is producing not only new models and chatbots but also a broad layer of software that companies must inventory, patch and secure. Many of these tools connect AI applications to sensitive corporate data and internal services, creating potential entry points for attackers.
"It's a little bit beyond just like the tool itself, but these [tools] talk to some really sensitive systems," Emily Austin, a principal researcher at Censys and co-author of the report, told Axios.
What products and numbers were found
- Censys identified 44 different AI-related products deployed in internet-facing configurations.
- The most widely exposed product was n8n, a workflow automation platform that can connect AI agents to other applications and services. Censys found nearly 114,000 internet-facing n8n deployments, including roughly 77,000 running software versions affected by at least one recently disclosed critical vulnerability.
- Researchers also discovered more than 54,000 exposed control panels for OpenClaw, an open-source AI agent that users can connect to personal accounts and other services.
Risks and limitations
The findings do not prove that every identified deployment is exploitable, but they underscore the risk of exposing software that can access corporate systems — particularly when such software may connect to emails, financial records or internal databases.
Censys CEO Zakir Durumeric told Axios that companies were already having difficulty tracking their existing software before employees began building their own AI applications, and the rise of these new apps complicates asset visibility and security.
Recommended actions for organizations
Censys researchers and report authors recommend treating AI infrastructure like any other deployed software. That includes regularly reviewing deployments, applying updates and patches, and implementing security controls to reduce the exposure of systems that can access sensitive corporate resources.
Bottom line
The rapid increase in publicly exposed AI components adds another layer to corporate cyber risk. The scale of exposure reported by Censys — with concrete examples such as n8n and OpenClaw deployments — highlights the need for companies to incorporate AI-related software into their normal vulnerability management and asset-inventory practices.
