GeoTech Cues
July 27, 2026 • 12:44 pm ET
Ivan Ferrari
In January, a federal judge in the Southern District of New York ordered OpenAI to produce twenty million de-identified ChatGPT conversations to news organizations suing it for copyright infringement. OpenAI could produce the conversations because they sat on OpenAI’s servers. Cloud artificial intelligence (AI) keeps its records with the provider, and control over those records rests with the provider and, when a court decides, with the parties across the table.
The ruling made visible a cost that Microsoft CEO Satya Nadella has since given a name. In a July 12 essay on what he calls the Reverse Information Paradox, Nadella argues that a buyer of a closed model pays twice: once in money and again in the proprietary knowledge it must reveal to make the model useful. Every prompt, correction, and workflow trace teaches the provider’s systems something about the customer’s business. “In consuming intelligence, you are creating intelligence,” he writes, and what you create should belong to you. The court order is one instance of a broader condition: when the model runs on someone else’s infrastructure, both the record and the learning sit outside your walls. Running the model on infrastructure you control brings them back inside.
