OpenAI reportedly lost 1.5 million subscribers in under 48 hours after agreeing to let the US Department of Defense use its AI models.
The Story
This move contrasts sharply with rival Anthropic, which rejected a similar deal and saw its Claude chatbot rise in popularity.
Why It Matters
For your European SMB, this isn't just tech news; it's a stark reminder about AI vendor trust and data sovereignty. When you entrust your firm's data to an AI provider, you're not just buying a tool, you're buying into their values and their data handling policies. This event underscores the critical importance of understanding where your sensitive information resides and who has access to it, especially when geopolitical factors come into play.
What To Do About It
My advice is clear: don't get caught off guard. First, ensure you have a robust data export strategy for any AI tools you use – know *how* to get your data out. Second, start evaluating alternative AI vendors like Anthropic's Claude *now*, not when a crisis hits, focusing on their data policies and ethical stances. This proactive approach protects your firm's continuity and client confidentiality.
Related Signals
Poland's data protection authority fined a door-to-door sales partner of Energa-Obrot for letting reps trade customer contract scans over personal WhatsApp with no oversight.
7 Aug 2026
Germany's BSI set 31 July 2026 as a de facto NIS2 registration deadline, with only about 18,500 of an estimated 29,000 to 40,000 affected companies registered by late May.
28 Jul 2026