morning_wind

Last night, the market was speculating that Google's questioning of MAG7's massive bleeding investments might all be about practicing the Sunflower Manual (a risky, self-destructive strategy);Today, it feels like they're realizing something's wrong while practicing and starting to live within their means, and hardware is going to suffer as a result.

$ Taiwan Semiconductor(TSM.US) Nothing much to say, steadily hitting new all-time highs.

Anthropic: U$62BOpenAI: U$40B-------The recent ARR situation of the two companies in June is still rising.$ XL2CSOPHYNIX(07709.HK)$ NVIDIA(NVDA.US)

The second half of AI semiconductor stock selection

Expressed dissatisfaction with the continuous large outflow of talent related to the model of $ Alphabet(GOOGL.US), sold a portion today. Reallocated into $ Intel(INTC.US) and $ Cerebras(CBRS.US). The reason for buying Intel is also clearly buying into some narratives over the weekend (CEO interview). Still hold quite a bit of Google, but the main expectation is now only on cloud and TPU. The expectations for the model since last year's Q3 are now gone.

Weekly semiconductor highlights By $ Alphabet(GOOGL.US) Big Banana Model Pro.

$ Meta Platforms(META.US)$ Direxion Semicon Bull 3X(SOXL.US)My take on the Meta thing: Market-based measures have optimized the resource allocation efficiency of computing infrastructure, similar to how a planned economy is being adjusted. It doesn't affect the overall supply-demand imbalance issue of the entire silicon-based industry. Locally, it might even increase the tolerance for internal competition among various CSP/CAPEX spenders: computing power is still something to fight for; it's better to scale it up in your own hands; if you really can't scale it up, you won't lose out (you can still resell it).

$ JOYY(JOYY.US), cash equals market cap, a very aggressive dividend plan. Even if it's a cigarette butt, it's made of gold. Bought a position last night. Hope this can set an example for the old guard of Chinese concept stocks and kick things off. The $ PDD(PDD.US) lottery results are out tonight, learn something from it.

$ AMD(AMD.US) AMD and OpenAI are both racing towards trillion-dollar valuations. A few months ago, the two companies signed a bet, and now they are in the final stage of mutual commitment: OpenAI will deploy up to 6GW of AMD Instinct series GPUs in phases, with the first 1GW batch expected to start in the second half of 2026; in return, AMD will grant OpenAI up to 160M equity subscription rights (approximately 10%), linked to delivery/market cap milestones. The market reacted strongly to this deal, bringing a lot of upside potential, but it also comes with significant dilution, execution, and infrastructure risks.

The process of building the investment portfolio in 2026 is quite different from previous years, mainly because I found it difficult to identify new stocks that meet my satisfaction in terms of win rate and odds, and are worth a large position from my circle of competence. In terms of results, the opening portfolio of 2026 doesn't seem to have changed much from the closing positions at the end of 2025.Overall, as briefly mentioned earlier, considering the combined effects of the election year, the change of the Federal Reserve chairman, and the typical strong government cycle, I think the overall capital policy environment for U.S. stocks will be more optimistic than in 2025 (hopefully no misjudgment). On the Hong Kong stock side, new stocks keep draining liquidity. Overall, I allocate 7:3 to U.S. stocks and China concept stocks. From the trend of interest rate changes, it doesn't seem advisable, but in the medium to long term, I think the cost of interest rates is acceptable relative to absolute returns. Specifically for U.S. stocks, gold, Bitcoin, financial innovation, oil and gas, etc., are still not my strengths. I might occasionally speculate on the sidelines, but they won't be included in the portfolio. XBI and LABU have rebounded enough, so I won't allocate them in 2026. Instead, I'll focus on AI.1. $ Alphabet(GOOGL.US) , $ NVIDIA(NVDA.US)  , and $ Taiwan Semiconductor(TSM.US)  are still the absolute main large positions. The market consensus is that Alphabet is leading overall but not cheap, NVIDIA is quite controversial, and TSMC's geopolitical risks are hard to handle. Personally, I think Google's recovery is complete, but there's still about 30% room for improvement. I think NVIDIA is at a low point. In fact, it's now a competition between two major alliances: Google vs. NVIDIA+OpenAI. My view is that they will repeatedly outperform each other and eventually both grow. TSMC is the most stable, with its newly expanded capacity and 2nm mass production ensuring enough growth. As for geopolitical risks, I don't consider them. If the risks materialize, I think other stocks within my circle of competence won't be safe either.2. This year, AI undoubtedly needs to focus on application ROI and Agents. Overall, I favor 2B scenarios > 2C scenarios and predict the most successful AI coding scenario, where Anthropic's lead will shrink. Apart from big players, the certainty of new players in this part is still unclear. The experience of independent applications like Cursor and Manus under the encroachment of LLM models is quite thought-provoking, and I have many ideas to expand on later. In terms of allocation, this is where 2026 differs most from 2025, but the overall position isn't high because it still doesn't feel very clear. Specific names include Apple, PATH, FIGMA, and ServiceNow.

$ Taiwan Semiconductor(TSM.US)$ Alphabet(GOOGL.US), thanks to the 13F, the quality of the two StockPro picks the OP is responsible for has been verified once again.

Under the pressure of interest rate cuts, small-cap stocks like $ SPDR S&P Biotech(XBI.US) are rising much faster than large-cap stocks like $ Alphabet(GOOGL.US)😂