Researchers Find 175,000 Publicly Exposed Ollama AI Servers Across 130 Countries

Yikes — researchers just uncovered over 175,000 Ollama AI servers publicly exposed across 130 countries. This isn't just a minor config slip — many of these servers have tool calling enabled, which means attackers can execute arbitrary code and pull off LLMjacking attacks. Basically, free compute for bad actors.

The scale is wild — 130 countries means this is a global exposure problem, not just a regional fluke. If you're running Ollama instances without proper network controls, you're basically hosting a public AI playground for anyone with an internet connection.

This is the kind of finding that makes security pros sweat — massive attack surface + AI capabilities = a perfect storm for exploitation. If you're in infra or devsecops, time to audit your Ollama deployments ASAP.