拓市

$Micron Tech(MU.US) hasn't hit the bottom yet; the range of 700-800 is the heavy position interval.

“CRCL represents the stablecoin and digital dollar infrastructure direction. Its investment logic is not simply based on Bitcoin's rise and fall, but rather on the future growth in the use of stablecoins in payments, cross-border transactions, and the financial system. Regulatory recognition can boost market expectations for its long-term value, but its valuation will also be affected by market sentiment.”

It comes from improvements in advertising efficiency, optimization of recommendation systems, and the large model ecosystem. Unlike NVDA which sells computing power, it uses AI to enhance its own business efficiency while continuously investing in AI infrastructure. Its advantage is strong cash flow, but the scale of AI investment is huge."

"Google has advantages in AI models, search, cloud computing, and its self-developed TPUs, making it one of the few companies with a complete layout from foundational models to applications. The challenge lies in whether AI will change the traditional search business model and whether the cloud business can continue to improve."

"MRVL primarily benefits from demand in AI data center connectivity, networking, and custom chips. Unlike NVDA, which sells GPUs directly, it provides key components in AI infrastructure, making it a niche beneficiary within the AI industry chain with higher elasticity but lower certainty compared to the leading players."

“TSM is the core of the underlying manufacturing in the AI industry chain. Advanced chips like NVDA and AMD all need it for production. Its biggest advantages are its advanced manufacturing processes and advanced packaging capabilities. The more complex the AI chips are, the higher the demands on manufacturing capabilities. However, the company also faces geopolitical risks and capital expenditure pressure.”

“AVGO benefits from AI infrastructure upgrades, with one direction being AI networking equipment and the other being large-scale custom ASIC chips. In the future, AI computing power may not entirely rely on GPUs, as cloud providers will increase their self-developed chips, bringing new growth opportunities to Broadcom.”

MU is part of the AI memory cycle logic. The significant increase in demand for HBM and DRAM from AI servers has raised the value of memory. Its characteristic is high elasticity because the memory industry has obvious cycles. Profits surge during good times, but the market also trades ahead of the cycle, so volatility is usually greater than NVDA.”

AMD's biggest highlight is the opportunity to be the second supplier of AI chips. It may not necessarily challenge NVIDIA's absolute dominance, but if it can consistently secure data center and AI accelerator card orders, the market will give it a higher valuation. The advantage is the large growth potential, while the risk is intense competition, requiring proof of actual revenue growth.

NVDA remains the core of AI computing power. The company's greatest advantages are its GPU ecosystem and CUDA barrier. Global large tech companies cannot build AI data centers without it. The market has already priced in high expectations, so future stock price increases depend more on whether AI capital expenditures can be sustained and whether profit growth can exceed market expectations."

$Sandisk(SNDK.US) SanDisk is currently in the "high prosperity + high volatility" phase of the memory chip cycle, with its pricing still essentially driven by the AI-induced supply-demand mismatch.

From a fundamental perspective, the company's latest financial report continues its strong performance, with revenue and profits showing significant year-on-year growth and a notable increase in gross margin. The core reasons are tight NAND supply coupled with explosive growth in demand from AI data centers, pushing the industry's price cycle into an upward phase. The market is also beginning to reassess its profit stability through "multi-year long-term contracts + customer price-lock models," which is weakening its cyclical nature.

From a stock price structure perspective, SanDisk has moved from the early revaluation phase into a high-level game range: on one hand, it continues to be driven by capital flows in the AI computing chain, and the trend remains; on the other hand, after the substantial gains earlier, marginal trading now relies more on earnings exceeding expectations and continued upward movement in industry prices. Once guidance is not strong enough, volatility will significantly amplify.

Overall, SanDisk at this stage is no longer about "valuation repair logic" but rather "high-elasticity assets within an upward cycle." The trend is not over, but the difficulty of trading has increased. Key observation points are the trend of NAND prices, the sustainability of data center demand, and the potential for upward revisions in subsequent financial report guidance.

$NVIDIA(NVDA.US) NVIDIA's recent price action continues to oscillate at high levels around the AI theme, with the overall trend intact but entering a phase of shifting momentum.

From a fundamental perspective, data centers and the Blackwell architecture remain the core growth drivers, supported by sustained capital expenditures from cloud providers, which underpin the long-term logic. However, the market is currently more focused on marginal changes in growth rates and the stability of gross margins during the transition to the new product generation.

From a trading perspective, it is currently a high-consensus stock. At elevated levels, capital is more inclined towards rebalancing rather than one-sided accumulation. Any catalyst vacuum or macro disturbance could significantly amplify volatility.

Overall, NVDA remains the core anchor for AI pricing, but in the short term, it is in a "verification phase," with its performance more dependent on subsequent earnings guidance and the pace of AI demand realization.

My view is that GOOGL may still experience short-term volatility, but the current position seems more like digesting negative news rather than a trend reversal. For long-term investors, the discussion now might not be whether Google can survive, but whether the market is underestimating its ability to monetize AI in the coming years.$Alphabet - C(GOOG.US)

$Alphabet - C(GOOG.US) Regarding GOOGL's recent trend, I believe the market is pricing short-term risks with the most pessimistic sentiment. The departure of two top AI scientists has raised market concerns about Google's AI competitiveness and talent drain, leading to a significant pullback in the stock price. In the short term, these concerns are real and are the main source of recent selling pressure.

However, from a fundamental perspective, Google Cloud continues to maintain strong growth, AI business commercialization is ongoing, and the core competitiveness of search, cloud computing, and the Gemini ecosystem has not fundamentally changed due to individual talent movements. The current market worry is about excessive AI investment and a long payback period, not problems with the business itself.

Historical experience tells us that pullbacks in high-quality tech companies often come from expectation adjustments, not value disappearance. When the market focuses on talent drain and capital expenditure, it's easy to overlook that Alphabet still possesses one of the world's most powerful traffic portals, advertising platforms, and AI infrastructures.

$Micron Tech(MU.US) Regarding MU's earnings report, my view is that the market has already priced in most of the positive news and released some risks ahead of time. The recent stock price pullback, volatility in the memory sector this round, and adjustments to production expansion expectations by South Korean memory manufacturers have already led capital to start reassessing the future growth pace. The market is not currently worried about whether this earnings report is profitable, but rather whether it can continue to exceed expectations.

From an industry fundamental perspective, HBM demand remains strong, with AI servers driving tight supply and demand for DRAM and HBM. Micron's management previously stated that HBM capacity has been largely pre-booked. The issue is that these positives have already been fully anticipated by the market. The current situation is more like a game of "expectation gap" rather than a game of fundamentals.

Therefore, even if the earnings exceed expectations, I believe it's highly likely to be a moderate rise rather than a sharp surge. This is because what determines the stock price is no longer the performance itself, but whether the performance far exceeds the market's highest expectations. If the guidance is not revised upward again, or if the management's wording becomes slightly conservative, a "sell the news" scenario could occur.

After a relatively thorough round of adjustments recently, SPCX's valuation has gradually returned to a rational range. From an industrial logic perspective, whether it's commercial aerospace, satellite internet, or global AI infrastructure construction, the demand for computing power, communication, and aerospace capabilities continues to rise. These long-term trends have not changed due to short-term market fluctuations. The current market is more about trading sentiment and liquidity, rather than a deterioration in fundamentals.

From a risk-reward ratio perspective, compared to previous highs, the current price has already released a lot of short-term risks, with relatively limited downside potential. In the future, as industry catalysts are released, institutional capital is allocated, and market risk appetite recovers, the upward elasticity is actually more worthy of attention. Often, truly valuable assets are not bought when everyone is optimistic, but are gradually positioned when there is market divergence.

For me, now is more like a stage of patiently waiting for value realization, rather than a stage of worrying about short-term fluctuations. Prices will fluctuate, but the long-term growth logic remains clear.

$SpaceX(SPCX.US) The market is in the process of redefining SpaceX from an aerospace company to an AI and space infrastructure platform. The stock price reflects the growth expectations for the next decade, not today's profit level.

$SpaceX(SPCX.US) SpaceX's greatest value currently lies not in its rockets, but in building the infrastructure ecosystem that connects Earth, space, and the AI computing power network. Behind the valuation fluctuations is essentially the market pricing of the future productivity platform.

$Sandisk(SNDK.US) AI has elevated the importance of computing power and simultaneously raised the strategic position of storage. SanDisk is undergoing a shift from a cyclical logic to a growth logic.

$Aerospace and Defense(IN00299.US) The core logic of the aerospace and defense sector is no longer short-term event-driven catalysts, but rather a long-term reassessment of demand against the backdrop of global security restructuring and space economy expansion.

$Micron Tech(MU.US) The current investment thesis has shifted from a traditional memory cycle stock to a core beneficiary of AI infrastructure. With HBM supply and demand remaining tight and data center capital expenditures staying high, the company's performance and profit guidance are being continuously revised upward. The core of market trading is no longer the DRAM price cycle, but the revaluation of memory value in the AI era.

$Nokia Oyj(NOK.US) Nokia is transforming from a traditional communications equipment cyclical stock into an AI data center network infrastructure company. The core of current market trading is "the repricing of AI network dividends, not the traditional operator cycle."

$S&P 500(.SPX.US)The current nature of the S&P 500 is not a broad bull market, but a highly concentrated structural trend driven by AI. The index strength masks the divergence in internal breadth.

$Semiconductors(IN00297.US) The core of this semiconductor cycle is not "demand recovery," but the acceleration of structural changes. AI is turning computing power from a "cost item" into "productive infrastructure," and industry pricing power is shifting from end-users to upstream players. The real beneficiaries are not all manufacturers, but the few players with advanced processes, packaging capabilities, and ecosystem binding power.