Satya Nadella, chairman and chief executive of Microsoft, argued on July 12, 2026, that companies using artificial intelligence models face a structural problem where getting good results requires disclosing proprietary knowledge. Nadella published this argument in a post on X titled "The Reverse Information Paradox," which had accumulated 6.2 million views. The post inverts Kenneth Arrow's 1962 paper "Economic Welfare and the Allocation of Resources for Invention," which described a paradox where a seller of information cannot prove its value without giving it away. Nadella wrote, "AI creates the reverse problem. In the AI age, the buyer risks giving away knowledge, just in order to use what they bought." He argues that a company paying for a model subscription or API access must reveal its own operational knowledge to get useful output, stating, "You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful." The mechanism Nadella describes is not a single data-sharing event but an accumulation over time.