VCF 9.1: From AI Hype to Production Reality

In this episode of the Virtually Speaking Podcast, hosts Pete Fletcher and John Nicholson welcome Chris Wolfe, Global Head of AI at Broadcom, to discuss the real-world evolution of AI deployment with VMware Cloud Foundation (VCF) 9.1. The conversation centers on the growing momentum of private AI, where organizations are moving beyond hype to implement production-grade AI solutions on-premises or at the edge. Chris highlights how customers are leveraging VCF for specific, measurable use cases—like government chatbots and contact center automation—while balancing cloud-based frontier models with local inference for governance, cost efficiency, and data control. A key theme is the shift from theoretical AI adoption to operational reality, with VCF acting as an intelligent orchestrator that enables smart routing of workloads between on-prem and cloud environments. The episode also explores technical advancements in VCF 9.1, including GPU pass-through support, CPU inference via Lama CPP, enhanced observability through detailed metrics, and the new MCP tools registry for governance. Chris emphasizes that production AI is far more complex than just generating tokens, and VCF’s role as a 'production enabler' is gaining traction across enterprises and partner ecosystems. The episode concludes with Chris expressing excitement about VCF’s strategic alignment with the industry’s shift toward private AI inference, the growing partner enthusiasm from model creators, and the increasing recognition that AI deployment requires more than just models—it demands infrastructure, governance, and operational maturity. The hosts underscore that the AI journey is no longer about experimentation but about building scalable, secure, and measurable AI systems, with VCF emerging as a critical foundation for this transition.