Original Viewpoint Summary

📈 Market Structure & Stock Picking

• “Stock picker’s market” means dispersion is doing the heavy lifting. Broad index beta is less useful; the edge shifts to finding individual names with clean setups and relative strength. 1

🤖 China AI & Agent Platforms

• China AI chip adoption could move offshore from 2028 onward. The key read-through: Chinese companies may start deploying Chinese-made AI accelerators in overseas regional data centers outside China, which would widen the addressable market beyond domestic compute demand. 2

• WeChat Agent looks more important than a standalone AI app because it does not require users to “go somewhere else.” Native placement inside WeChat lowers friction and could make the Agent layer part of daily workflows, not just another chatbot destination. 3

🚀 Private Markets & SpaceX Valuation

• $SPCX is only ~9% away from wiping out Elon Musk’s “trillionaire” paper status. The line is simple: SpaceX at ~$154/share keeps him above $1T, while a drop to around $141/share puts his paper net worth back under $1T. 4

💾 Memory, Korea & Semis

• $EWY / KOSPI weakness is a near-term overhang for $DRAM. With KOSPI selling off overnight, memory names could face pressure in the next session. 5

• If HBM4 ramps slower than expected, memory makers carrying “hockey stick” projections are exposed. The bull case depends on fast HBM4 adoption; any delay risks estimate cuts and multiple compression across $DRAM-linked names. 6

🇭🇰 Hong Kong IPO Risk

• Hong Kong recent IPOs can turn into brutal unlock traps. 药捷安康 was as high as 679元 less than a year ago; after the IPO lock-up ended on 6月22日, it fell nearly 60% today and hit a low of 11.69元. That is the classic post-lockup rug pull in HK new listings. 7