How to View the Rebound in AI Applications? @Huatai Computer (0731)

English summary

This morning we reported: short-term negative factors in the AI industry chain have been exhausted, hardware volatility has declined / we see a rebound; this morning AI applications rebounded more strongly than hardware. Why and how sustainable? Reasons for the rise in AI applications: First, capital rebalancing. The core reason is that the short-term logic for computing power has flaws, and it is difficult to directly reverse into a main upward wave (the root is that demand is unclear in the short term; apart from coding, what is the next explosive scenario?). Second, changes in AI application narrative: ① Enterprises are shifting from token-maxxing to token optimizing, starting to move toward cheaper/higher cost-performance models; ② Competition among large models is intensifying, and low-cost open-source models are rapidly approaching top closed-source models. Earlier K3, and later the price impact of the official DeepSeek V4. Today, OAI's 80% price cut for GPT-5.6 Luna (lightweight/entry-level) and 20% cut for Terra (mid-range/main) is a proactive response to this trend. …