Krux
February 12, 2026
Splunk Just Launched Its Own AI Models Inside Your Data
Published: February 12, 2026 at 5:09 AM
Updated: February 12, 2026 at 5:09 AM
100-word summary
Splunk dropped native AI models directly into its Cloud Platform on February 18, 2026. The lineup includes Foundation-sec-1.1-8b-instruct for security tasks (think log analysis and threat detection in plain English) plus GPT OSS models at 20B and 120B parameters for heavy-duty reasoning. Everything runs on Splunk's own GPU infrastructure—no third-party LLM needed. The Model Context Protocol Server went live February 4, letting AI agents tap Splunk data instantly. Plus, the AI Assistant for SPL now writes and explains queries in natural language, with zero data leaving Splunk's walls. This could fast-track "Agentic Ops," where AI agents autonomously query and act on enterprise data without messy integrations—a major shift for security...
What happened
Splunk dropped native AI models directly into its Cloud Platform on February 18, 2026. The lineup includes Foundation-sec-1.1-8b-instruct for security tasks (think log analysis and threat detection in plain English) plus GPT OSS models at 20B and 120B parameters for heavy-duty reasoning. Everything runs on Splunk's own GPU infrastructure—no third-party LLM needed. The Model Context Protocol Server went live February 4, letting AI agents tap Splunk data instantly. Plus, the AI Assistant for SPL now writes and explains queries in natural language, with zero data leaving Splunk's walls.
Why it matters
This could fast-track "Agentic Ops," where AI agents autonomously query and act on enterprise data without messy integrations—a major shift for security and observability teams.