Why Sovereign AI Is Worth Building, and Where It Breaks

What happened
On May 13, 2025, a federal judge in Manhattan ordered OpenAI to keep user chat logs, including those users had previously deleted, as part of a legal case with The New York Times. This order applied to ChatGPT Free, Plus, Pro, and Team accounts, as well as API users without specific data retention agreements. The order lasted until late September, but data flagged by the Times remains under legal hold. Enterprise customers with certain agreements were excluded from the order, but many business users were still affected.
Key details
- The court order was signed on May 13, 2025, in Manhattan.
- It required OpenAI to preserve chat logs from ChatGPT and API users, even if users had deleted them.
- Enterprise customers with zero-data-retention agreements were not included in the order.
- The order ended in late September 2025, but some data remains on legal hold.
- Surveys show 71% of executives and officials consider sovereign AI a strategic imperative, and 90% say open source is essential for it.
Why it matters
Organizations using external AI platforms may lose control over their data if legal actions force providers to retain information, even after deletion requests. This affects businesses, governments, and any group handling sensitive data through third-party AI services, increasing risks around privacy and compliance. The situation is driving strong interest in sovereign AI solutions, where organizations can control their own infrastructure and data, often using open-source tools to reduce dependency on outside vendors.
