In this episode of the AGNT Podcast, recorded at the New York Stock Exchange, Raphaelle d'Ornano, co-host and enterprise technology analyst, joins theCUBE + NYSE Wired's Gemma Allen to examine who owns the orchestration layer — and the rent — in the agentic AI era. D'Ornano applies a framework built around intent, context and workflow moats to scrutinize Palantir's valuation of 71 times revenue, even after a 35% decline from late 2024 highs. She acknowledges the company's exceptional metrics — 10 consecutive quarters of growth, 70% year-over-year expansion and 138% net revenue retention — but argues the current price encodes a no-doubt scenario: that Palantir is already the orchestrator across all its domains, government and commercial alike, a claim she is not prepared to accept without debate. The conversation also explores the diverging strategies of OpenAI and Anthropic, pushing back on the tendency to treat frontier AI labs as interchangeable. D'Ornano argues Anthropic's enterprise-first approach and Claude Code's wedge into developer workflows represent a credible path to orchestration, while OpenAI's broader ambitions — from consumer advertising to the acquisition of OpenClaw — leave its durable competitive moat less defined. She underscores the fundamental risk embedded in both business models: a potentially catastrophic mismatch between spiraling compute costs and revenue, noting that Anthropic scaled from $1 billion to $14 billion in ARR in under two years. From speculating on IPO timelines for xAI, Anthropic and OpenAI to warning public investors of the single points of failure that no spreadsheet can capture, d'Ornano offers a candid framework for navigating the most consequential — and volatile — chapter in enterprise software history.