In this inaugural episode of the AGNT Podcast, Raphaëlle d'Ornano, co-host and agentic AI analyst, joins theCUBE + NYSE Wired's Gemma Allen to discuss how Fortune 500 companies can build their readiness for the emerging agentic AI era. D'Ornano argues that the defining competitive advantage in the new paradigm is not data alone — it's context. She explains why agents require access to deep, defensible and non-portable data, structured through the right memory systems, to make grounded decisions, and why the model context protocol (MCP) has been a pivotal development in enabling agents to access that context securely. The conversation also explores the MiniMax IPO — one of the first LLM companies to go public — which listed on the Hong Kong Stock Exchange alongside Zhipu and doubled its valuation on day one. D'Ornano offers a measured assessment of MiniMax's M2.1 model, noting its multimodal architecture and coding-agent strategy as genuine strengths, while flagging DeepSeek's anticipated training breakthrough as a meaningful competitive risk. She also addresses the broader investment thesis for LLM companies, explaining why these plays are best framed as optionality bets rather than traditional discounted cash flow analyses. From the geopolitical dimensions of Chinese open source AI to a UC Berkeley study showing closed models dominating enterprise agentic deployments, d'Ornano provides a clear-eyed framework for how investors and enterprises should navigate the fast-moving world of agentic AI.