Smart AI investment is not about chasing the big one

(Bloomberg opinion) – How AI will reform our lives is blurred with uncertainty. Researchers are debating what it means to achieve artificial general intelligence, or AGI, or the neural network systems that drive chatbots such as chatgpt, can lead us to the great price. But smart investors should avoid trying to predict future breakthroughs of a technology still at an evolutionary stage. Rather than participating in OpenAI Inc. expensive sales of secondary stocks, they can get exposure to the AI ​​theme using a time-testing approach. The supply chain is long, complex and prone to bottlenecks. At each chokepoint there are some economic goats-to use the phrase of Warren Buffett-which delivers much faster and more than ambitious, great language model developers. The best -known restriction occurs in designing the chip. The question for Nvidia Corp. products are so strong that Openai had to find alternative sources, with the conclusion of a mega agreement with the competitive designer Advanced Micro Devices Inc. This week for inferiority functions. Distractions, or using pre -trained models to generate user content, may have to do with less advanced chips. Diggle deeper and there are more unclear, but equally interesting opportunities. We see a boom in the data center, especially in the US. In July, spending of construction reached $ 41 billion on an annual basis, or 47% of the total on offices. But developers are afraid of different bottlenecks, from acquiring slides to the security of power generation. Try to get hold of gas turbines, burning natural gas and turning a generator to produce electricity. Earlier, it took two years to get a new one, and it is now taking up to five or more. This is partly due to an oligopoline market structure, dominated by GE Vernova Inc. In the US, Germany’s Siemens Energy AG and Japan’s Mitsubishi Heavy Industries Ltd. After a gas turbine boom in the early 2000s, the three producers became cautious and expanded the capacity, fearing that the demand of the data could not be. Meanwhile, startups are also not trying to lead, because the technological obstacle for access is high. In some cases, turbine designs have taken decades to test and adjust. The high-banded memory market, or HBM, which is essential to support AI systems such as those of Nvidia, is another good example. South Korea’s Sk Hynix Inc. is the industry leader, followed by Samsung Electronics Co. and Micron Technology Inc. The technology is so complicated that Huawei Technologies Co. Korean products in some of its leading AI processors had to use, even if the Chinese conglomerate preferred to source domestic. It is not surprising that Openai has also embarked on strategic partnerships with Samsung and Hynix in an effort to ensure his supply chain and scaling up in front of everyone. All three shares are on fire this year. To be sure, identifying AI winners is not as simple as studying technical access barriers and market structures. The Netherlands’ Asml Holding NV has an almost monopoly on the specialized lithographic equipment used to make high-performance chips, but the share price missed the AI ​​rally this year. One important reason is a concentrated customer base. ASML is not too much on Foundry Taiwan Semiconductor Manufacturing Corp, which does not sell too much to the Chinese, but it is too much relied on Foundry Taiwan Semiconductor Manufacturing Corp. So what are the lessons? First, for investors without a computer scientific background, identifying issues in supply chain is perhaps easier than asking if AI research can lead to training models that display human intelligence. After all, some money managers have been through a similar exercise and looked at suppliers around the world during the smartphone era. Secondly, the AI-fired stock boom is not limited to the US or China because the supply chain is so complicated that no country can claim that it can do everything on its own. Third, there were many concerns about an AI bubble – investors are understandably cautious as Mega Caps from Oracle Corp. until AMD can melt in a single trading day by more than 20%. But the diverse stock performance of, for example, gas turbine manufacturers and ASML indicate that there is still rationality in this rally. Not all AI names have yet to shine. Most importantly, most of the profits in this AI race are probably earned upstream, where the biggest issues are. By 2029, when Openai finally began to break, boring turbine engineering firms may have already earned a truck money. More from Bloomberg opinion: This column reflects the author’s personal views and does not necessarily reflect the opinion of the editorial or Bloomberg MP and his owners. Shuli Ren is a columnist for Bloomberg covering Asian markets. She was a former investment banker and was a reporter for the Barron market. She is a CFA charter holder. More stories like these are available on Bloomberg.com/opinion © 2025 Bloomberg LP