I saw someone replied to you before, and it was actually quite fair. Xiaomi is downstream in the token ecosystem, Alibaba is upstream, and their work scopes are different. If you want Xiaomi to do everything, it's just a case of the ideal being very full, but reality is very skinny. Tencent's conference call also mentioned they are considering renting computing power. Only the super giants have the capability for real AI infrastructure.

My view hasn't changed. Those in the know already went all-in last week. Potential new narratives that may emerge next include: accelerated overseas expansion of new energy under the US-Iran situation, new developments in AI large models, human-vehicle-home ecosystem interconnection, miclaw, and chips, etc. Looking forward to tonight's launch event!

I think Xiaomi will replace Alibaba in the second half of the game. After all, when it comes to AI models, their capabilities will eventually converge. Models like Qianwen, Yuanbao, and Doubao will likely become similar in performance, as talent poaching can somewhat bridge the gap.Currently, the focus of AI competition is on large models and computing power, but in the second half, the key will shift to end-user applications. From Doubao phone, we can see ByteDance is eager to integrate models with devices. In contrast, Xiaomi, as the player with the most devices, is in a strong position to win effortlessly.From a business development perspective, Xiaomi has a closed-loop ecosystem covering people, cars, and smart homes, with growing international capabilities and exponential performance growth. On the other hand, Alibaba is losing ground in e-commerce to PDD and in food delivery to Meituan, relying solely on Alibaba Cloud to sustain performance. This is too fragile, and they still have to subsidize loss-making businesses with profits from this segment, making sustainable development hard to guarantee.