Cisco and Splunk launch on-premises AI infrastructure

Cisco and Splunk have announced the availability of the Cisco AI POD for Splunk, a turnkey on-premises infrastructure designed to host agentic AI workloads within enterprise data centers. This development allows organizations to deploy AI capabilities locally, providing an alternative to public cloud providers for companies with strict data sovereignty or regulatory requirements.

The Cisco AI POD integrates Cisco UCS compute, NVIDIA GPU acceleration, Red Hat OpenShift, and Splunk’s AI runtime software. It is designed to support the Splunk AI Assistant and the upcoming Agent Launchpad. This infrastructure enables enterprises to host various models locally, including Google Gemma 4 and OpenAI GPT-OSS 20B, and is compatible with air-gapped environments.

The move addresses growing concerns regarding data governance and the difficulty of deploying AI in regulated industries, particularly as frameworks like the EU AI Act become enforceable. By bringing AI to the data rather than moving data to the cloud, Cisco aims to resolve challenges related to data security and management complexity.