Overview

I've reorganized the statistics. The Longbridge account has a relatively good return rate, so it has a Top 2% badge. Started investing in Hong Kong and US stocks from 2025-01-01:

• Total Return: +136%
• Annualized Return: +78%
• Total Return for H1 2026: +34%

It's a bit awkward. Almost all returns this year come from memory (Micron and SK Hynix). Other stocks are roughly breakeven. The price increase in memory completely exceeded my expectations. The allocation was about 5% and it roughly increased 8 times. The top three holdings have basically remained unchanged. I'm bullish on their long-term value.

Current Holdings

1、$NVIDIA(NVDA.US)  Approx. 39.4% allocation, approx. +9.5% P&L

2、$Eli Lilly(LLY.US)  Approx. 23.6% allocation, approx. +15.9% P&L

3、$Palantir Tech(PLTR.US)  Approx. 18.1% allocation, approx. -24.6% P&L

4、$Applied Optoelectronics(AAOI.US)  Approx. 8.8% allocation, approx. -21.7% P&L

5、$Roundhill Memory ETF(DRAM.US)  Approx. 6.5% allocation, approx. -2.5% P&L

6、$Bitwise Hyperliquid ETF(BHYP.US)  Approx. 3.6% allocation, approx. +7.0% P&L

Stock Commentary

$NVIDIA(NVDA.US) ： Ten years ago, Jensen Huang said: GPUs are not just for processing images; GPUs can rebuild this world. NVIDIA is focused on creating the future. Today, large models have emerged, and AI agents are accelerating. Huang's prophecy seems about to come true, with credibility points MAX (unlike Musk who talks nonsense). Today, Huang says the next decade is about AI infrastructure, aiming to upgrade NVIDIA from a GPU company to a "smart production infrastructure company." I believe in the future he describes, so I went all-in.

• Valuation: Even if growth slows by 2027, I give it a forward PE of 30%, PEG≈1, undervalued.
• Recent Company Actions: Investing in upstream supply chain, increasing buyback plans—all positive long-term behaviors. Remain bullish.
• Next Step: Hold heavy position. Might buy some long calls if it drops significantly.

$Eli Lilly(LLY.US) ： *WALL-E* depicts future humans as obese with atrophied limbs. Human nature is to be lazy and love food. The weight-loss drug (or metabolic drug) market is a sustained growth market. The better life gets, the higher the demand for weight-loss drugs.

• Capability: Lilly basically dominates its rival NVO in all aspects. It has been continuously acquiring new drugs in the past six months to solidify its moat. Its current pipeline products also remain strong.
• Recent Company Actions: The oral weight-loss drug Orforglipron underperformed Wegovy in both efficacy and sales—a negative. Needs continued observation.
• Valuation: Lilly's current valuation is reasonable. It needs a performance breakthrough to drive the stock price higher.
• Next Step: Hold position.

$Palantir Tech(PLTR.US) ： Wall Street is rife with "SaaS crash theory." Palantir was directly labeled an "old man stock." Its stock price trend is on par with the Hang Seng Tech Index, becoming my second-largest losing stock. My view remains the same: Coding has long ceased to be a moat for any company. AI Coding merely reduces the cost of coding, not necessarily creating more value. What enables AI to create more value is precisely the specialized knowledge accumulated by companies like Palantir through years of deep vertical expertise, along with long-accumulated channels, data, reputation, and customer trust.

• Valuation: It fell for half a year because it was too expensive. The market just found reasons to cut valuations. The current price is relatively reasonable, but sentiment hasn't reversed; it might still fall.
• Recent Company Actions: Recent expansion in Europe hasn't gone smoothly—slightly negative.
• Next Step: Hold position.

Memory： Over the past six months, "old man" stocks were abandoned by the market, and memory became the "new favorite." My understanding of memory is the same as most people's: remain bullish. Demand far outstrips supply. However, starting with Apple's price hikes, upstream price increases have already been passed down to end-users. The pressure for memory prices to rise further will increase.

• Stock Selection: DRAM to participate in the next phase, seeking stability.
• Next Step: Will continue to add on dips.

$Applied Optoelectronics(AAOI.US) ： Missed the optics rally this year. Bought AAOI to amplify portfolio elasticity. Remain bullish. No additions or reductions.

$HYPE： Fell at the beginning of the year following the crypto market and the unlock narrative. Actual buybacks were larger than the unlock amount. Narrative reversed.

• Valuation: Crypto can't be valued. Benchmarking against BNB, short-term target $100.
• Recent Company Actions: Looking forward to the HIP-4 prediction platform.
• Next Step: Add on dips. Already took profits on PURR. Now switching to BHYP and continuing to hold.

Summary

My long-term thesis hasn't failed, but the market is currently rewarding another, stronger supply-demand consensus. The old consensus hasn't disappeared; it's just priced in slower. The new consensus is accelerating, and I only captured a small part of it—unexpected but reasonable. Finally, ending with a quote from Charlie Munger:

You don't need a lot of action; you need a lot of patience. You have to stick to your principles, wait for the opportunity to come, and then pounce on it with vigor.