Because the LLM field is now crowded and so it’s not about being good enough to attract a customer who hasn’t used A.I. It’s about being good enough to get someone to switch from their current option. Which means you need proprietary data. The first batch got distribution from novelty, the second wave has to get it from specialization. Apple lacks that data. Even Perplexity and OpenAI lack that which is why they are trying to do things like buy browsers and other data heavy startups. It’s why Google with search + browser + YouTube + email, will overshoot long term (example Nano-Banana and Veo being better than other image and video contenders by 10x)