Original Insights Roundup

📉 Market Sentiment & Index Setup

• Ugly tape can be a bottoming setup, not a deal-breaker. Awful sentiment, big washouts, retail blowups, and Iran headlines are exactly the kind of stress cluster that usually shows up before bottoms. Bottoms do not print when everyone is bullish and the backdrop feels clean. 1

• $QQQ still has bear-case downside if the current pattern rhymes with “the last time.” Another 6-7% decline from current levels remains on the table unless this week can convincingly hold the key blue support line. Futures bouncing is not enough yet. 2

🧠 AI, Semis & Memory Cycle

• Changxin IPO is the main event for A-shares because it combines scale, policy relevance, and hard fundamentals. It is framed as the largest A-share tech IPO, a domestic semiconductor chain-leader, with 2026 H1 expected revenue of 1100–1200亿元, net profit attributable to parent of 500–570亿元, sharp YoY growth, and roughly No. 4 global DRAM market share. 3

• Changxin opening straight to 3.2万亿 market cap turns it into the new A-share “king,” showing just how aggressively the market is pricing domestic memory leadership. 4

• Nvidia–OpenAI could become the most ambitious financial trade of the U.S. AI boom. The key is not just chips. It is balance-sheet engineering: Nvidia is reportedly discussing roughly $2500亿美元 of backing/guarantees for a 10GW hyperscale data-center project in southern Ohio tied to SoftBank. 5

₿ Crypto Market Structure

• The crypto market is entering a “new normal”: not just exchange closures, but project shutdowns too. The stress is moving from trading venues into protocols themselves, with Radiant (RDNT) announcing DAO orderly exit/maintenance mode on 6月1日, and zkSync-ecosystem L2 Sophon (SOPH) shutting down its own L2 chain on 6月25日. 6

• $BTC oversold RSI divergence is working so far. The setup is still technical and early, but current price action is validating the divergence instead of rejecting it. 7

📌 Single-Stock Positioning & Capital Discipline

• Tesla is a clean reminder that position sizing matters. Over the past five years, going all-in on Tesla would not even have beaten the S&P 500. The latest selloff reinforces the need for strict allocation discipline instead of blind conviction. 8

• Berkshire looks broadly defensive, but Google is the exception that matters. The portfolio reportedly cleared out 16 stocks, held nearly 4000亿 in cash, yet increased Google from 1780万股 to 5780万股 and deployed 100亿美元 in a private placement. In an otherwise cautious book, that exception is the real signal; the implied Google level cited is 350美元. 9