Global chip stocks plummeted Morgan Stanley: The artificial intelligence investment cycle has changed
[Global chip stocks plummeted, Morgan Stanley: Artificial Intelligence investment cycle has changed] On July 6, local time, investment bank Morgan Stanley released a report saying that the recent weakness in U.S. semiconductor stocks indicates that the market has accelerated profit-taking, and the direction of stock market investment is shifting. Investors may turn to "hyper-scale companies" that have invested in large amounts of data centers for artificial intelligence. These companies include Google, Amazon, Meta, etc.
On July 6, local time, investment bank Morgan Stanley issued a report saying that the recent weakness in U.S. semiconductor stocks indicates that the market is accelerating profit-taking and that the direction of stock market investment is shifting. Investors may turn to "hyper-scale companies" that invest in large data centers for artificial intelligence. These companies include Google, Amazon, Meta, etc. After the Asia-Pacific market opened on July 7, Samsung's stock price fell by more than 6.8%, and SK Hynix also fell by nearly 6%, triggering a circuit breaker in the Korean stock market. Driven by the artificial intelligence craze, SK Hynix officially launched its IPO plan in the United States on July 6, aiming to raise 43 trillion won (approximately US$28 billion). However, the recent sharp decline in chip stocks in the U.S. market has also triggered market discussions on how long the artificial intelligence boom will last. In the past two weeks, the U.S. chip stock index has fallen more than 11%. Morgan Stanley believes that as the artificial intelligence cycle shifts, hyperscale cloud service providers such as Google and Amazon may benefit from the rotation of semiconductor stocks. In the past few months, memory chip stocks represented by Samsung, SK Hynix and Micron Technology have experienced a super cycle. An investor who holds SK Hynix told China Business News that he believes that SK Hynix’s IPO in the United States is positive for the company itself, because it will expand the accessibility of global investors and help compete in the global market. "Although the market has faced recent fluctuations, the overall trend of chip stocks is still improving, and it is expected that demand for SK Hynix in U.S. stocks will remain strong." He said. SK Hynix’s listing in the United States will be second only to SpaceX’s record-breaking US$85.7 billion IPO. It is also expected to join the Philadelphia Semiconductor Index after listing. Last month, HSBC raised SK Hynix’s valuation by 20%. In Samsung's latest performance forecast, the company said that operating profit in the second quarter will increase 19 times year-on-year, reaching a scale of nearly 90 trillion won (approximately US$58.4 billion), exceeding the total revenue of the past three years. However, before the release of Samsung's performance expectations, the market had already digested most of the positive news, which also led to a sharp drop in the stock price after the company officially released its results on July 7. Some cautious analysts believe the biggest risk to the memory chip boom will be a slowdown in investment in artificial intelligence infrastructure. Memory chips have faced significant recent price increases that could undermine spending on artificial intelligence infrastructure as well as mobile phones and other consumer electronics. "We believe the memory chip cycle has moved beyond its early stages," Standard Chartered Bank said in a recent report. In addition, delays in U.S. data center construction due to labor shortages, power constraints and other factors may ultimately weaken demand across the entire AI hardware supply chain. Analysts also said that in the long term, the trend of continued growth in artificial intelligence infrastructure spending remains unchanged. In a recent report, Goldman Sachs predicted that total capital expenditures in artificial intelligence-related fields by the four largest hyperscale companies in the world - Meta, Microsoft, Amazon and Google - will reach US$5.3 trillion between 2025 and 2030. These large tech companies are the largest investors in infrastructure such as AI data centers. Sameer Wasson, CEO of MIPS, a chip IP manufacturer under GlobalFoundries, recently said in an interview with China Business News: "On the application side, we are now also involved in the construction of AI data centers. These products have now achieved mass production and can be deployed or have met the conditions for later deployment, and the volume of data center projects is very large." Watson said that whether it is automobiles, industry, consumer electronics or data centers, the demand for physical AI is huge. "We are also concerned that the entire industry is facing certain bottlenecks in terms of production capacity recently." He told China Business News. He also said that the CPU and AI chip design fields where MIPS is located are highly capital-intensive projects, which not only involve scale, but also involve customization requirements, which are related to whether the products can meet end-to-end needs. (