The latest episode of theCUBE Pod finds John Furrier and Dave Vellante looking past the GPU race to the bigger shift: AI infrastructure becoming a full intelligence production system. Furrier and Vellante connect sovereignty, data movement, networking, storage and compute as enterprises rethink where AI runs, who controls it and why today.The conversation moves from Dell’s infrastructure momentum and CoreWeave’s expansion to agent security, software disruption and the growing battle for control. Furrier and Vellante also weigh the risks of autonomous agents, rising infrastructure costs and the new cloud landscape taking shape as enterprise AI demand keeps accelerating across the industry.

The interface is fading, and AI agents are ready to run the show. On the latest episode of theCUBE Pod, theCUBE Research’s John Furrier and Dave Vellante break down Salesforce’s Agentforce gamble, the rise of headless experiences and why Data Cloud could become the context engine behind enterprise agents everywhere.Furrier points to survey data showing that fewer screens could mean more usage, more activity and a larger role for agents. Vellante follows the money to chips, data centers, CoreWeave and NeoCloud infrastructure, asking who ultimately owns the AI production stack as governance, pricing and a possible “SaaSpocalypse” loom ahead.

Nvidia’s latest quarter raises a bigger question: How far can its AI infrastructure ambitions reach? On theCUBE Pod, theCUBE Research’s John Furrier and Dave Vellante break down the company’s shift beyond GPUs toward full-stack AI factories, tracing monetization gains across Hopper, Blackwell and Vera Rubin architectures and enterprise adoption patterns.Furrier sees more potential per gigawatt as Nvidia expands its platform, while Vellante points to networking, memory and inference economics as constraints. They connect Cisco’s Nvidia collaboration, neoclouds and rack-scale edge systems to changing deployment models, while Salesforce’s results suggest AI can expand SaaS adoption rather than displace it.

AI’s biggest balancing act is no longer theoretical. In the latest episode of theCUBE Pod, theCUBE Research’s John Furrier and Dave Vellante break down the economics, sovereignty concerns and infrastructure choices shaping enterprise strategies, connecting soaring token and GPU costs with governance, architecture and the push toward on-premises AI factories.Furrier and Vellante turn those pressures into practical guidance, from modeling token economics and ROI to comparing cloud and on-premises deployments for steady workloads. They also discuss Canva’s cost reduction, AI virtualization and why separating the control plane, data governance and model governance helps organizations control spending, compliance and risk.

Is the AI infrastructure boom built to last, or are the warning signs getting harder to ignore? On theCUBE Pod, theCUBE Research’s John Furrier and Dave Vellante debate soaring investment, NeoCloud financing and enterprise demand while tracking CoreWeave, Nebius and the supply constraints shaping the market’s next phase for enterprises.Vellante sees familiar bubble signals in rising valuations, memory pricing and financing risk, while Furrier argues real demand and Nvidia-backed capital could keep expansion moving. They also assess utilization, contract structures, residual values and Cisco’s latest results, giving technology leaders a sharper view of what could matter through 2030 and beyond.

AI is moving beyond the model era, and John Furrier and Dave Vellante dig into what comes next on the latest episode of theCUBE Pod. Furrier and Vellante examine the rise of AI systems spanning infrastructure, cloud architecture, security and economics, while also reflecting on the lasting contributions of David Floyer.Drawing on theCUBE Research and reporting from Black Hat and AMD Advancing AI, Furrier and Vellante examine the pressures shaping AI deployment. They focus on high bandwidth memory, power constraints, neo-clouds, AI factories, agentic workflows and security as enterprises confront new bottlenecks and rethink how large-scale AI systems operate.

In this week's theCUBE Pod , John Furrier gives his take on two major enterprise stories: AMD’s shift from chips to integrated AI systems and IBM’s emerging role as an operating system for enterprise AI. He argues that Wall Street is overlooking the bigger transformation underway across infrastructure, software and business value.Furrier then joins Dave Vellante for a broader discussion recorded at AMD Advancing AI. They debate AMD’s transformation into a systems company, its role as an alternative to Nvidia and the economics shaping enterprise AI, including inference growth, token costs, Databricks and the tradeoffs between open ecosystems and tightly integrated platforms.

IBM’s earnings shock, NVIDIA’s commanding position and Databricks’ rising valuation set the agenda in the latest episode of theCUBE Pod. John Furrier and Dave Vellante break down what these signals reveal about enterprise AI adoption, infrastructure spending and the widening gap between legacy technology and systems built for intelligence today.Furrier and Vellante also examine how data pipelines, agentic workloads, token economics and AI FinOps are reshaping enterprise strategy. They explain why companies must connect proprietary data with systems of intelligence, improve cost discipline and turn AI investments into measurable business value as competition across the infrastructure market intensifies today.