Larry Ellison ORCL mentioned a very critical point: models like ChatGPT, Gemini, Grok, and Llama are essentially trained on roughly the same publicly available internet data. When everyone trains on the same information, the models will inevitably become more and more alike. This is also why AI is heading towards commoditization. The real moat has never been just the model itself but the proprietary data behind it. Whoever can train with exclusive data that others can't access will be able to establish an advantage that can't be replicated. When you have data that others don't, you have the opportunity to dominate your market. Trade ORCL and other US stocks using @MSX_CN