Too much euphoria in the stock markets about Artificial Intelligence?
Commentary by Steven Smith, Investment Director at Capital Group, on the importance of distinguishing short-term AI euphoria from possible long-term investment opportunities
As the role of artificial intelligence (AI) consolidates on the world stage, it is increasingly clear that it is no longer science fiction, but a technology with a growing impact on our daily lives.
It is always important to distinguish short-term euphoria from possible long-term investment opportunities. “AI could prove more radical than electricity or fire.” Statements like this may make headlines, but they are little indication of how these advances could help a company improve its efficiencies, grow revenues or build lasting competitive advantages.
CONSIDERATIONS FOR INVESTMENT IN AI
From an investment perspective, it is still too early to have a definitive opinion on real-life and industrial applications. What we are trying to do right now is establish the possible positive and negative implications of AI, understand what factors could accelerate or decelerate the pace of implementation, and think about a possible framework for investments in the present and long term. Artificial intelligence is already being seen as the next industrial revolution, with the huge market opportunity that comes with it.
AI is obviously a major structural growth theme for the next decade and beyond, and companies that fail to embrace it risk being left behind. This commercial imperative means that an inflection point for this technology is likely to lie within our investment time horizon, and ChatGPT could be seen as officially kicking off the AI race, albeit in essence the participants were already preparing.
From a broad perspective, this technology is expected to drive efficiency gains, reduce costs, significantly increase the pace of innovation and broaden overall target markets for businesses, which should be good for profit margins and stock valuations, as well as accelerating GDP growth. We also expect to see a proliferation of new companies born thanks to AI that don't exist today.
ALL SECTORS (BEFORE OR AFTER) WILL BE INVOLVED
It's hard to think of sectors that won't be positively affected by AI in some way, but there are areas of the global economy where the short-term effects are clearer, including healthcare, energy, autonomous vehicles and 'agriculture. In the healthcare sector, for example, AI is already starting to spread across fields as diverse as drug discovery, medical diagnostics and scan analysis, impacting personal care through remote monitoring and medical record keeping .
Novo Nordisk, for example, is entering into a strategic partnership agreement with Microsoft to accelerate drug discovery and development by using AI to analyze vast amounts of scientific literature, patents, reports and discussion forums, in order to develop syntheses and analyzes capable of directing researchers towards new discoveries. A current area of interest is training AI models to predict the risk of developing atherosclerosis (cardiovascular disease) by identifying biomarkers that can distinguish possible drug targets.
In other sectors such as energy, AI could have a positive impact on electricity grid optimization and demand forecasting, improving the performance and efficiency of energy storage. In agriculture, it could help pave the way for more precision farming by optimizing irrigation, reducing waste and predicting and improving crop yields.
Overall, we also expect generative AI to be broadly positive for a number of key global sustainability goals, including access to essential information and services such as financial planning, education and healthcare. From an environmental perspective, AI could improve areas such as climate scenario modeling, pollution monitoring and natural disaster forecasting, resulting in better proactive mitigation practices.
Given the diverse nature of the AI ecosystem, there is no single “correct” way to track the range of investable AI-related opportunities. Given this, below is a potentially useful investment framework for determining short- and long-term opportunities.
AI IN COMPUTING
Semiconductors are the brains behind AI, which is computationally intensive in both the training and inference phases. While semiconductors remain growth cyclical, the long-term growth trajectory for this sector is always exponential and the market could nearly double from around US$500 billion in 2022 to over US$1 trillion by the end of decade. A significant portion of this will likely be driven by increased computing needs from artificial intelligence.
AI IN INFRASTRUCTURES
If semiconductors are the fundamental building blocks of AI, the companies that provide the infrastructure are the foundation. These include public cloud hyperscalers (such as Microsoft's Azure), which allow companies to outsource computing tasks to the cloud via massive data centers. The benefit is that customers have on-demand, pay-per-use access to the most advanced and powerful computing services, without having to run them on-premise. Infrastructure also includes companies that provide hardware, such as network components and switchgear, as well as software that makes cloud computing more efficient, considering the high speed and bandwidth required by AI.
AI IN MODELS
Much of the hype surrounding AI at the moment focuses on the companies that “create” the models related to it. These include names like OpenAI, which has attracted a lot of interest following the success of ChatGPT. When considering model developers, we are wary of possible commoditization, given the large and growing open-source AI community promoting the concept of “AI for Humanity.”
This community has been collaborative in sharing research, ideas, coding, and best practices. Data ownership will likely prove to be the most important criterion for identifying the ultimate winners in this space, which naturally favors owners of large, unique, proprietary data sets, such as technology companies. To create a cutting-edge generic fundamental model also requires billions of dollars, as well as talent which is in fact in short supply. While many start-ups are dedicating large sums of capital to these models, we expect that only a small number of such companies will be able to compete sustainably due to scale requirements and high barriers to entry, which is why so we believe that the group of undisputed winners in the AI models segment will be narrow.
AI IN APPLICATIONS
Turning to applications, our analysts believe that software companies focused on the “productivization” of AI could benefit significantly and quickly; such winners will have direct monetization leverage, significantly increasing prices. The opportunity for developers to deliver consumer- or enterprise-level software that integrates AI capabilities is clear. Let's think about how a company like Microsoft can integrate AI into its 365 suite, which includes Outlook, Word, Excel and PowerPoint, and charge a large recurring premium.
In our opinion, this segment of the value chain will evolve dramatically over the next decade, based on previous paradigms. In the early years of smartphones, for example, few could have predicted the birth of applications like Uber or Airbnb, which would become services for everyday life.
LARGE AI COMPANIES VS STARTUPS
As for AI, what we can currently imagine about its possible applications is based on our limited understanding of this new technology. Of course, there are hundreds, if not thousands, of startups trying to build the next potentially revolutionary product, but it's hard to predict which of these might succeed. Our current focus is instead on leading companies that are already successfully integrating AI capabilities into their existing suite of applications, as they are likely to be the ones to capture the most value in the near term.
Finally, at the bottom of this investment framework are the real-life and industrial beneficiaries of AI, which could ultimately be limitless, spanning multiple generations. Again, however, it is important to remember that AI is still in an early stage of development: what the technology might look like in 10 years' time, how long it might take to gain consumer trust, and how integrated AI applications could be into our daily lives.
We continue to remain focused on the opportunities that can arise from AI and believe that an in-depth research approach will become even more instrumental in identifying potential winners and avoiding losers. A layered framework, such as the one described, begins to examine and reveal some of the key opportunities that will become available and allows us to develop a broader perspective in terms of successful long-term investing.
This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/economia/borse-intelligenza-artificiale/ on Sat, 02 Mar 2024 06:09:10 +0000.