Microsoft chairman and CEO Satya Nadella warned that enterprises adopting AI face a hidden cost beyond subscription fees: the proprietary knowledge they must hand over to make the technology useful. In a post on X, Nadella argued that using AI effectively means "you essentially pay for intelligence twice," first with money, then by revealing institutional know-how. He framed this as the "Reverse Information Paradox," inverting Nobel Prize winner Kenneth Arrow's "Information Paradox." Nadella stated that the better a company wants an AI model to perform, the more proprietary information it must feed it, creating information asymmetry skewed in the seller's favour. He highlighted "exhaust"—accumulated prompts, tool usage, and corrections—which he said becomes "intelligence exhaust" that leaks out "trace by trace, correction by correction, eval by eval." Nadella noted that patents solved Arrow's paradox, but the Reverse Information Paradox needs an equivalent safeguard beyond conventional data protection. He quoted Palantir CEO Alex Karp, saying enterprise customers want to "own the means of production." Nadella outlined a five-part framework built around the five Cs: Control, requiring enterprises to retain ownership of institutional memory, feedback and decisions; Capability, calling for proprietary learning environments within a company's own tenant boundary; and Choice.