Raphaelle d'Ornano of Decoding Discontinuity appears on AGNT Podcast Episode 8, broadcasting from the New York Stock Exchange with theCUBE Research. d'Ornano draws on their experience in artificial intelligence, referred to as AI, strategy and examines how intelligent systems reshape enterprise economics. Hosts Gemma Allen of theCUBE Research and theCUBE Research guide a focused conversation on hyperscaler capital expenditure concentration, the four-layer inference stack of chips models harness and applications, agentic pricing models and the evolving roles of AI labs cloud providers and SaaS vendors in the inference economy. Key takeaways include that hyperscalers command disproportionate capital expenditure and influence market structure, and that pricing shifts toward consumption-based agent models, d'Ornano states. They warn that compute scarcity could temporarily make compute costlier than human labor and highlight recent partnerships with Anthropic as signals that hardware and cloud arrangements are critical to scale inference affordably. The episode addresses strategic factors to consider for AI infrastructure pricing model design and deployment of large language models, referred to as LLMs, across enterprise applications.