Near AI Says Security Breach Affected 100 Users and 18 Legacy Agents
Edited by Steinar Vigdel Kolnes, Editor, Oslo
Near AI said a Nov. 8 intrusion exploited an open-source inference engine flaw and may have exposed confidential AI requests.
Near AI has disclosed a security incident affecting account information tied to roughly 100 users, as well as data associated with 18 legacy AI agents.
The intrusion occurred on Nov. 8, 2026, after attackers exploited a vulnerability in an open-source inference engine used within Near AI's infrastructure.
The company said the vulnerability may have allowed confidential AI requests to be exposed. Details on the specific account information involved were not included in the initial disclosure.
Near AI linked the breach to a security weakness in SGLang, highlighting the risks that vulnerabilities in third-party open-source components can pose to AI services.
Source: Bitcoin.com News
