Cisco Expands Nvidia Partnership for Agentic AI in Private Data Centers

Cisco Systems Inc. (CSCO) has expanded its partnership with Nvidia Corp. (NVDA) to bring agentic AI directly to Splunk customers. As part of this collaboration, Cisco and Nvidia introduced the Cisco AI POD for Splunk, a turnkey hardware and software system that lets companies run the Splunk AI Assistant locally within private clouds or air-gapped data centers.

The infrastructure supports self-hosted open and proprietary large language models, including Google's Gemma 4 and OpenAI's GPT-OSS 20B. Justin Boitano, vice president of enterprise AI at NVIDIA, noted that the joint architecture gives businesses a high-performance foundation to run AI agents where their data natively resides.

Cisco also debuted a 'Tokenomics' capability within its Splunk Agent Observability suite, which tracks token consumption across custom AI agents and developer tools. This feature provides predictive analytics to forecast upcoming usage trends, helping IT leaders link AI expenses directly to operational outcomes before billing cycles conclude.