Shares of semiconductor giant NVIDIA, the leading maker of AI chips, rose 150% in the first half of the year, giving NVIDIA a market value of $3 trillion and briefly making it the world’s most valuable company. Tech giants Microsoft, Amazon and Broadcom, which are among the most prominent participants in the AI boom, also rallied.
Generative AI has the potential to increase productivity in all sectors of the economy. That is why it is an important factor for all of us to consider not only from the perspective of how we organize our lives and economic activities, but also in terms of investments.
As much as I believe in the long-term potential of AI, I’ve also learned to be skeptical when stocks rise so much so quickly. I worked as an analyst for telecommunications companies during the technology and telecommunications boom in the late 1990s. I remember the enormous excitement about the potential of the Internet to fundamentally transform the economy. Unfortunately, there was a gap between the enthusiasm of investors, which was there immediately, and the real positive impact on the economy, which only came to fruition years later.
That was an important lesson for me. The Internet has had a huge impact on the economy ever since, but the returns weren’t immediate – and investors became impatient.
When making AI-related investments in my portfolios, I consider the following four risks.
1. Investors often overestimate the short-term impact of technology
We tend to overestimate the immediate impact of technological innovations and underestimate the longer-term impact. This can be explained by the J-curve used to depict productivity development. When new technologies are introduced, companies and investors are often excited by their transformative potential and therefore invest heavily in the development of the corresponding infrastructure.
Adapting to new productivity-enhancing tools takes time. New technologies can initially reduce productivity as companies and individuals must continue to work with old processes while learning and integrating new ones. As a result, it can take several years before tangible economic benefits are realized.
I believe that in 10 years, AI will have transformed the way we do business. But that doesn’t mean that we can expect immediate returns or that AI will only lead to better development for companies. We are dealing with a cycle that is subject to the same psychology and the same economic rules that we see in other areas of innovation.
2. the pace of capital investment depends on the results
Tech giants Microsoft, Meta, Alphabet and Amazon are investing tens of billions of dollars in AI infrastructure, much of which has been spent on semiconductors and other components needed to build out data centers. In order for these so-called hyperscalers (leading internet and cloud platform providers) to continue to spend so much, I believe they need a significant return on investment in the form of revenue and ultimately profit growth.
Will AI deliver returns in the next year or two? I expect it will for certain companies, but not for many others. The companies that will likely suffer setbacks are those where the share price already reflects future growth expectations related to AI.
I have seen such a scenario before. The market and companies are extremely euphoric because they see great growth opportunities and therefore invest significantly in this opportunity. In the current situation, this is about high spending on AI infrastructure. But then, when the market sentiment changes and a reluctance to spend sets in, companies start to reduce their investments one by one. At the moment, we do not see anything like this in AI-related investments. However, when the market situation changes, the result will be a change in mindset across the industry and spending will decrease.
In fact, in 2022, the market punished some companies with the best growth figures and high investments, simply because they did not generate a return on them. A prominent example: Meta, the parent company of Facebook, responded by declaring 2023 the year of efficiency and significantly reducing its expenses.
I think that this focus on efficiency is still there among Big Tech companies. I expect that these companies will be more disciplined in terms of investments and profit-seeking than in the past, and that discipline will ultimately translate into AI investments. If that happens, that could lead to a decline in capital spending over the next few years, but also surprises in profits and margins.
3. Resource scarcity could further slow down AI rollout
Expanding AI infrastructure requires a wide range of resources, not least specially qualified specialists. Not only do you need people who can create the basic models on which generative AI is based, but you also need people who know how these models can be used in practice in companies. AI also requires a lot of electricity to run data centers.
As a result, demand for energy is increasing, pushing the electricity grid to its limits. To meet their enormous energy needs, hyperscalers have already turned to utilities that rely on nuclear energy. In March, Amazon acquired a data center from Talen Energy to gain access to nuclear power from a nearby Talen nuclear power plant in Pennsylvania. Last June, Microsoft entered into an agreement with Constellation Energy to supply one of its data centers with nuclear power.
Due to potential capacity constraints, AI data centers may not grow as quickly as some expect in the next few years.
4. Blisters can hurt like hell
Are we already in an AI bubble? I don’t know, but we are moving more and more in that direction. As I mentioned, as a telecom analyst in the late 1990s, I had a front-row seat to the expansion and bursting of the technology and telecom bubble, so to speak, so I know the catastrophic impact that the downturns in such market cycles can have. I also know that the ups and downs tend to occur even when the long-term promise is ultimately largely fulfilled. That was the case with the Internet in the 1990s and 2000s, and it will be the case with AI today.
Let there be no misunderstanding: The euphoria in AI stocks is clearly different from the bubble of the 1990s. First of all, NVIDIA and other tech giants have posted strong earnings growth in recent quarters and are overall better valued than the leading stocks were in 2000. But I still expect that at some point in the next 12 to 24 months we will reach a period of disillusionment where growth stagnates. Despite a strong, sustained trend, there is a possibility that some leading AI stocks will experience sharp price drops. In addition, at the start of a major innovation cycle, it is not always clear which companies will emerge as long-term winners, so it is important for investors to be aware of the risks.
Even if AI meets the most optimistic expectations about its potential, investors can suffer significant losses on the way to realizing that potential. When the technology and telecommunications bubble burst in 2000, many companies disappeared from the market, others saw a sharp decline in their value. And for some, recovery took several years. In the late 1990s, Cisco Systems, a manufacturer of networking hardware and telecommunications equipment for the Internet, saw its stock soar to become the most valuable company in the S&P 500. After the bubble burst, shares collapsed by almost 80% as telecommunications companies cut spending. Cisco has not yet regained the highs of that era.
Conclusion
There is currently a lot of euphoria about AI – and I think that is fundamentally justified. I believe that AI will bring us spectacular things and will be very important in the future. But I also believe that investors should choose their investments very carefully in the current phase of the cycle and carefully weigh up the risks.
By Chris Buchbinder, Equity Portfolio Manager at Capital Group
Past performance results do not allow any conclusions to be drawn about the future development of an investment fund or security. The value and return of an investment in funds or securities can rise or fall. Investors may only receive less than the capital invested. Currency fluctuations can also affect the investment. Please note the regulations for advertising and offering shares in InvFG 2011 §128 ff. The information on www.e-fundresearch.com does not represent any recommendations for the purchase, sale or holding of securities, funds or other assets. The information on the e-fundresearch.com AG website has been carefully prepared. Nevertheless, unintentional errors may occur. No liability or guarantee can therefore be accepted for the timeliness, accuracy and completeness of the information provided. The same applies to all other websites to which reference is made via hyperlink. e-fundresearch.com AG rejects any liability for direct, specific or other damages that arise in connection with the information offered or otherwise available. The NewsCenter is a special form of advertising from e-fundresearch.com AG for asset management companies, which is subject to a fee. Copyright and sole responsibility for the content lies with the asset management company as the user of the NewsCenter special form of advertising. All NewsCenter messages represent press information or marketing communications.
