AI investment 2nd round, from GPU to power·industrial goods·space

AI investment funds are shifting from GPU-focused semiconductor stocks to broader infrastructure sectors including power, industrial goods, and space, as profit-taking hits leading AI chipmakers like Broadcom. The U.S. market is rotating capital into power equipment, cooling systems, optical communication, and healthcare stocks, reflecting a view that AI bottlenecks are expanding beyond GPUs. This trend was reinforced by stronger-than-expected U.S. employment data, which boosted Treasury yields and the dollar, signaling continued economic growth that supports wider AI infrastructure spending.

🚀 AI investment 2nd round, from GPU to power·industrial goods·space 🌐 Auto-translated from Korean The market has been showing an interesting scene recently. On one hand, profit-taking is occurring in leading AI semiconductor stocks. Broadcom's earnings were not bad, but there was no additional upward revision to AI guidance to the extent the market had expected. As a result, profit-taking occurred in some parts of the AI value chain. However, there is a more important point. It's not that AI-related funds are leaving the market; rather, they are moving elsewhere. Indeed, in the recent U.S. market, funds are seen shifting towards industrial goods, power, healthcare, and AI infrastructure-related stocks. In China, the AI hardware and robotics supply chain continues to show strength. The core point, as I see it, is simple. AI is not over; rather, AI bottlenecks are expanding beyond GPUs. This week's market and U.S. employment figures were another example demonstrating this trend. 1. Stronger-than-Expected U.S. Employment This U.S. employment report came out much stronger than expected. Indicator | Actual | Expected | |---|---|---| | Non-Farm Payrolls NFP | 172K | 85K | | Private Payrolls | 120K | 85K | | Unemployment Rate | 4.3% | 4.3% | | Average Hourly Earnings YoY | 3.4% | 3.4% | Looking at the headlines alone, it's a very strong employment report. However, a closer look at the details reveals a slightly different picture. Key Growth Sectors | Increase Size | |---|---| | Leisure & Hospitality | +70K | | Government | +52K | | Education & Healthcare | +40K | Conversely, - Finance -22K - Information Services -2K - Wholesale & Retail -3.7K were relatively sluggish. In other words, this employment growth was characterized more by increases in government, healthcare, and service sectors rather than an explosive improvement in manufacturing or IT. 2. Market Interpretation Reflected in Interest Rates and Exchange Rates Following the employment announcement, U.S. Treasury yields rose rapidly. Maturity | Yield | |---|---| | U.S. 2-year | 4.12% | | U.S. 10-year | 4.54% | | U.S. 30-year | 5.02% | From the market's perspective, “The U.S. economy is still robust, more than expected” This was interpreted as a signal. The KRW/USD exchange rate also rose to ₩1,550. While many interpret the won's weakness solely as a Korean issue, the strength of the dollar played a significant role this time. Market participants are once again - U.S. growth - U.S. interest rates - U.S. assets moving in a direction that favors. 3. But Why Did AI Take a Break? This week, AI-related stocks showed a somewhat subdued performance. The direct trigger was Broadcom's earnings. AI revenue remained strong, but there was no additional upward revision to the extent the market had expected, leading to profit-taking, especially in AI semiconductor stocks that had risen significantly. However, the important point here is that funds did not leave the market. What the market chose were other AI beneficiaries. 4. From GPUs to the Entire System When discussing AI infrastructure, many investors still first think of GPUs. Of course, GPUs are still important. However, what the market is showing recently is a slightly different direction. AI semiconductor demand doesn't end with just GPUs. As AI data centers increase, what's also needed is - Power - Cooling - Optical Communication - Packaging - Servers - Network - Industrial Infrastructure Indeed, looking at the stocks that have shown strength recently, this trend is visible. Theme | Key Area | Reason for Focus | |---|---|---| | Power | Transformers & Power Equipment | AI Data Center Power Demand | | Industrials | Cooling & Equipment | Expanding Infrastructure Construction | | Optical Communication | Optical Fiber & Optical Modules | Increasing AI Traffic | | Space | Satellites, RF, & Communication | Expansion of AI and Data Infrastructure | | Healthcare | Expanding AI Utilization | Defensive Nature and Growth Potential | If AI investment Round 1 was GPUs, AI investment Round 2 can be seen as the entire system. 5. China is Moving in Another Direction The Chinese market is also interesting. Currently, AI hardware in China is stronger than the overall index. The key is - CXMT - YMTC - Unitree the anticipation of large IPOs from companies like. As AI localization and the expansion of the robotics industry intertwine, - Semiconductors - Servers - Optical Communication - Robotics supply and demand are spreading across the entire value chain. The Chinese market currently appears to be driven by AI self-sufficiency and supply chain construction as a more important investment thesis than economic recovery. 6. Space Can Also Be Part of the AI Supply Chain One of the areas I've been watching with interest recently is space. Many people think of rocket companies when they hear 'space industry,' but the actual supply chain is much broader. For example, $APH /symbol/APH Amphenol $TEL /symbol/TEL TE Connectivity $MCHP /symbol/MCHP Microchip $ADI /symbol/ADI Analog Devices $KEYS /symbol/KEYS Keysight $QRVO /symbol/QRVO Qorvo $VECO /symbol/VECO Veeco are deeply connected to the supply chains for satellites, RF, test equipment, and space-grade solar cells. As the data generated by AI increases, eventually - Communication - Network - Satellite Infrastructure demand will also increase. This trend is less of a short-term theme and more akin to long-term infrastructure expansion. Summary from an Investment Perspective Ultimately, what the market showed this week is one thing. AI is not over. Rather, AI is expanding beyond semiconductors. After Broadcom, some semiconductor stocks took a break, but market funds are still seeking AI-related infrastructure. Theme | Key Stock Groups | Reason for Focus | |---|---|---| | AI Semiconductors | | $K000660 /symbol/000660 SK Hynix, $K005930 /symbol/005930 Samsung Electronics, $MU /symbol/MU $MCHP /symbol/MCHP , $ADI /symbol/ADI , $QRVO /symbol/QRVO , $KEYS /symbol/KEYS , $VECO /symbol/VECO Conclusion This U.S. employment report significantly alleviated recession concerns. However, a closer look at the details shows that service and government employment accounted for a large proportion, rather than explosive growth centered on manufacturing or IT. Nevertheless, the market is still not giving up on AI. It's just that the scope of AI investment is broadening from GPUs to power, cooling, optical communication, industrials, and space infrastructure. Therefore, going forward, rather than simply following leading AI stocks, tracking where the essential bottlenecks move as AI grows may become more important. The market seems to be indicating that these bottlenecks are already expanding beyond semiconductors. This post reflects the author’s own opinion and is not investment advice or a solicitation from bullbear.ninja.