Summary of Key Points
ASML, the global leader in advanced chip equipment, released financial results that exceeded expectations (both revenue and gross margin were better than the market anticipated), yet its stock price plummeted, causing a decline in the entire U.S. semiconductor sector. This does not indicate the end of the AI cycle; rather, it marks the beginning of a new phase in AI investment: in the past, any company related to AI would see its stock price rise, but now the market is more discerning. Investors are no longer solely focused on performance but also on whether future growth can support high valuations, the sustainability of orders, and whether the current trend is just a temporary peak. The investment logic has shifted from a general upward trend to one of differentiation, with a need to distinguish between assets with certainty (long-term leaders), assets with volatility (which may offer higher returns), and ancillary assets (related to AI but not necessarily semiconductors).
Detailed Analysis
1. Why Did ASML’s Strong Financial Results Lead to a Stock Sell-Off?
There are four main reasons for this paradox:
- Positive News Was Already Overpriced: The AI hardware sector (such as NVIDIA GPUs, TSMC’s advanced manufacturing processes, and HBM memory) had already seen significant gains in the past year, so ASML’s good financial results were more of a confirmation of existing trends rather than new news. Those who bought shares early took the opportunity to sell for profit.
- High Valuations and Increased Market Rigor: Semiconductor stocks, especially those related to AI, are already at high prices. While good performance used to drive price increases, investors now require clearer evidence of sustained growth—questions like whether growth will continue into 2027, whether new equipment (High-NA EUV) will sell well, and whether customer investment will remain stable remain unanswered, leading to lower valuations.
- Concerns About the Future of AI Investment: Cloud companies (such as Microsoft and Amazon) have invested heavily in AI data centers, driving chip demand. However, there are concerns about whether this trend has peaked and whether there might be an oversupply after 2026, which is deterring further investment.
- Crowded Markets and Fear of Selling: AI is a highly competitive field, and even minor disruptions (such as ASML’s financial report) can cause panic among investors, leading some to sell their shares before the market stabilizes.
2. Has the AI Cycle Really Ended? Don’t Panic—It’s Just Changing
The cycle has not ended; it has just entered a new phase:
- Phase One: Focusing on Demand: Stocks of companies that demonstrated a need for AI computing power (such as NVIDIA) rose sharply due to strong demand.
- Phase Two: Focusing on Supply, Profitability, and Differentiation: Producing AI chips requires advanced equipment, HBM memory, packaging technology, etc. Each segment presents different opportunities: some companies may have their valuations overstated, while others still have room for growth; some rely on hype, while others on actual profits.
- Key Reminder: Avoid extremes—don’t conclude that the AI era is over based on short-term declines, nor blindly chase high valuations just because of long-term trends. Even the best sectors can experience corrections, and even the best companies’ stock prices can fluctuate.
3. What Positive Signals Are Hidden in the Financial Results?
Despite the stock price drop, the financial report contains important industry indicators:
- Increase in Equipment Service Revenue: ASML’s revenue from equipment maintenance and upgrades accounts for nearly 30%, indicating that customers (chip manufacturers) are making the most of their existing equipment due to delays in new shipments. This confirms ongoing demand for AI chips.
- 75% Increase in Storage Demand: There is a surge in demand for HBM memory, particularly for AI applications, which will drive demand for related equipment (etching, testing).
- Progress with New Equipment: ASML’s High-NA EUV technology is already being used by Intel, indicating that chip manufacturers are pursuing even more advanced manufacturing processes (2nm, 1nm). This long-term trend will not be halted by a single stock price drop.
4. How to Invest Now?
Semiconductor and related assets can be categorized into three types based on risk and return:
- Certified Assets: Hold for the long term and buy when prices adjust. These are industry leaders with high barriers to entry, such as ASML (in lithography equipment), TSMC (core in AI chip manufacturing), and NVIDIA (driving demand for AI). Their long-term prospects remain strong, but current valuations are high; wait for prices to drop before buying.
- Volatile Assets: Invest in companies with large price fluctuations that could offer higher returns, such as Micron (HBM memory), Ram Research (storage equipment), and Corley (chip testing). These companies’ performance is closely tied to the AI cycle, so monitor orders and profit realization.
- Ancillary Assets: Avoid high-valued semiconductors and focus on related infrastructure components. Examples include Vertiv (power supply for data centers), Arista (high-speed networking), and Eaton (power management). Although not semiconductors themselves, these companies are essential for AI implementation and have lower valuations.
5. What Signals Should You Watch Next?
To predict the future direction of the semiconductor sector, focus on these four indicators:
- Cloud Companies’ Capital Spending: Will Microsoft, Amazon, and others continue to invest in AI data centers? If spending continues, demand will remain strong; if it slows, be cautious.
- NVIDIA and ASIC Orders: NVIDIA’s GPU orders reflect direct AI demand, as do Broadcom and Marvell’s custom chip orders. A strong performance in these areas indicates that AI demand is not monopolized; however, any divergence could indicate structural changes or overall demand declines.
- HBM Prices and Supply: Will there be a shortage of HBM from companies like Micron and Samsung? If prices remain high, these stocks may perform well; if supply increases, prices may fluctuate.
- Valuations: Have semiconductor valuations dropped to a reasonable level? Only when valuations are more balanced can positive news drive price increases.
Conclusion
The AI cycle has not ended, but semiconductor investment requires a more rational approach. Evaluate companies based on their actual performance and valuations, and avoid crowded markets. The current phase of differentiation presents both risks and opportunities; choosing the right assets will help you profit in the new phase of the AI era.
(End of Article)