Value Investor's Thoughts on AI

Allow me to start off by making it very clear that I am not a technological expert and my knowledge of AI is rather superficial. The essay that follows is entirely my own personal view of Artificial Intelligence, based on my understanding of human nature, historical financial market events of huge significance, and my position as a contrarian value investor.

I believe people's attitudes towards AI and its advancements in recent years vary but can be put in one of the following categories. Some, who might have watched a bit too many Terminator movies, fear that the more sophisticated it becomes, the more likely it is to trigger a fight-or-flight response, which will then enable it to turn against us, take over our Internet-connected systems, and start the process of enslaving humanity. Others believe it might not go as far, but it will eventually take our jobs and leave us on the street, where we will starve to death. The third group of people think slightly more on the positive side and focus on all the efficiency it will bring to our everyday lives and the massive economic growth it will spark. And the last group of people think AI is just a hype bubble waiting to burst and burn everybody who is heavily invested in all those companies, whose stock prices skyrocketed as soon as they mentioned "AI" in their quarterly investor updates.

I, as an investor, try to focus on its impact on company profits and how AI is going to change the corporate world. Those stories taken out of the "Terminator" movie lore I find to be closer to fiction, and thinking about them brings no true value to my life and work. It might sound like I am burying my head in the sand here, but rather I chose to focus on concepts and ideas that are more likely than not to happen. Concepts I can use to my advantage while thinking about my current investments as well as where to allocate my capital.

This leaves us with the following three categories for me to delve into: AI is going to take our jobs; AI is going to improve our lives and make us richer; AI is just another "tulip" bubble, and it is not going to end well.

AI is going to take our jobs and leave us on the street

Let's go back to 1804 for a moment and think about how people were traveling back then. Riding a horse was the best way to get to your destination. It was neither convenient nor comfortable, but it was faster than walking. Then came the invention of the train. It was far more convenient for people since all they had to do was simply take a seat, either read a book or chat with their fellow travelers for the duration of their journey, and just wait until the train arrived at their stop. Also, train journeys were far smoother, warmer, drier, and faster compared to the alternative of riding a horse. Moreover, steam-powered railroads completely revolutionized the transportation of heavy goods and raw materials over long distances. You can easily understand why trains became the preferred mode of transportation for citizens of the more developed countries back then and decimated the demand for horses. Naturally, the expense of this transportation revolution fell on saddle makers, horse breeders, blacksmiths, and all other occupations that had a direct connection to horses. They were forced to choose between changing crafts or facing starvation after losing their jobs. This whole story sounds a bit familiar now, doesn't it?

In order to bring this analogy up to scratch and match it to the current AI concerns, we need to answer the following simple question: What is AI good at? Here are the first things that come to mind: digesting massive amounts of data and summarizing it within seconds; image recognition; automation; solving complex problems based on data it has been fed; writing software; translating text and speech in a large number of languages in near real-time; generating visual as well as audio in the form of music, images, videos, and even movies. So who are the "saddle makers" that are going to be affected the most by this shift in information technology? To name a few: online customer service representatives, technical customer support personnel, software engineers, movie directors, actors, music artists, translators, and people with administrative jobs. In essence, individuals from diverse fields of labor but mostly entry-level jobs. At least this is how it seems for the time being.

AI is going to improve our lives and make us richer

Now that we have covered the most likely worst-case scenario, let me put on my "positive thinking" hat and explain how I see AI as a tool that can help push humanity for a greater good. The human brain is perhaps one of the most remarkable products that have come out of human evolution. It is utterly remarkable how far it has enabled us to go as a species with its ability to compute mathematical problems, envision physical as well as fictional matter in three dimensions, recall previously consumed information, and form mental pictures or ideas. But lets be real, it comes with its own flaws, such as bugs in the form of feelings and emotions (subconscious feelings), it works at a low frequency compared to modern computer hardware, and it can sometimes get sick and develop more complex and unfixable bugs. This is where computers enter the room. They sure do not have what we humans call "sixth sense" and do not possess a micron of social skills, but they leave humans in the dust when it comes to computation speeds. So, people did what we have been doing throughout history and utilized their creation to our advantage. We developed the Internet so we can communicate with one another over vast distances; we learned to speak their language and program software applications that will enable millions of engineers around the world to create programs that make our lives more efficient as well as fun. We soon discovered that the bottleneck of the evolution of the computer was the human brain. Software code could only be written as fast as the developer's brain could function. Source codes were full of bugs and inconsistencies, which made applications misbehave. So, humans did what we do best and came up with a solution to this problem in the form of Artificial Intelligence. We created the AI technology, which in turn enabled computers to do the thinking for us, with the sole purpose of removing that data throughput bottleneck.

Here is a simple example of how AI can do something you and I can't. Imagine someone giving you a task to summarize a book you have never read in a matter of 2 seconds. Impossible, right? Well, not for AI. Obviously this is not a very useful use case, so let me give you another example. Let's say you are a software engineer on your first day at work. You are assigned a ticket to solve a problem you know nothing about. So you start doing what every self-respected engineer would do: read the company WiKi to familiarize yourself with the application you are troubleshooting. It might take you hours, days, or weeks to find the solution, which is purely because our brains are very slow when it comes to ingesting and analyzing data. Now imagine the same situation, but you have an AI bot that your company has developed and integrated into their software development stack. Within a matter of seconds, you will have a solution to your complex problem. Moreover, the most time-consuming part of that second scenario is you trying to figure out how to ask your question and typing it into the AI console. This is just one use-case of how powerful and efficient AI can be.

Humans have always been able to build or develop tools in order to solve previously unsolvable problems. AI is no different. The capitalistic world we live in truly loves efficiency, and AI is designed to maximize exactly that. It will enable companies to either increase their sales or expand their margins, or both.

AI is just another "tulip" bubble, and it is not going to end well

"A financial bubble is an economic cycle characterized by rapidly increasing prices of an asset to a point that is unsustainable, causing the asset to burst or contract in value."

Financial bubbles have been part of human history for centuries. Our fear of missing out (aka FOMO) has dictated the financial decisions made by a large portion of the population of first- and second-world countries. To use our transport revolution example from earlier, a few years after the train was invented, the first modern inter-city railway was built between Liverpool and Manchester in the UK in 1830. By the mid-1840s, the UK economy was coming out of a rough patch of slow growth as well as political and social unrest. The manufacturing industries were starting to grow once again and people were looking to take out their money from government bonds (gilts) and invest them in assets which promised better returns. Existing railway companies were the right candidate that ticked all the boxes for people with capital as they moved ever-increasing amounts of cargo and people. New media such as newspapers made it extremely easy for newly founded railway companies to promote themselves as foolproof venture and provided the means for the general public to invest in them. Shares could be purchased for a 10% deposit, with the railway company holding the right to call in the remainder at any time. Thousands of investors on modest incomes bought shares while only being able to afford the deposit. Many families invested all their savings in those lightly regulated railway stocks thinking they will only go up. But we all know that gravity is a b*tch and whatever goes up needs to come down. As with other bubbles the Railway Mania became self-promoting cycle based on purely over-optimistic speculation. As the sector got saturated with a large number of companies and the initial excitement settled investors started to slowly realize that railways were not all as lucrative and as easy to build as they had been led to believe. Investment in those empty promises stopped virtually overnight leaving many companies without funding which ultimately led to investors loosing their money.

History remembers many bubbles of such scale: the Southsea bubble, tulip mania, the tech bubble, the housing bubble which lead to the Great Financial Crisis of 2008.

All this brings us to the questions: Is AI becoming a bit of a bubble? People are very excited to not only witness but also take part in the next big technological revolution. What is very special about AI is that everybody with a smartphone or a computer can experience how powerful AI is and what it is capable of. Simply sign up for OpenAI's ChatGPT or ask the Google Assistant on your phone to "Talk" and shoot your question. Any topic any complexity - answers will be very informative and insightful. It is extremely accessible and very potent. People are rightfully impressed which leads them to believe that they an easily make money by investing in AI companies. Unfortunately, making good investment decision very rarely mixes well with excitement and other human emotions.

Lets stop here for a second and do a very quick analysis of Nvidia, by far the biggest AI chip manufacturer in the world. Nvidia's stock price has gone up 1965% over the course of the last 4 years, growing its market cap from $165 billion in February of 2020 to $3.4 trillion today. The company is currently trading at the staggering 29 times SALES and 50 times profits, with 87% of all the company's revenue coming from its Data Center segment, aka "AI chips.". This really begs the question, "Are AI chips liquid gold of the "tulips" of the 20th century?".

Now, people will say that this massive overvaluation is based on the fact that the company grew its revenue from ~$11B in 2020 to ~$63B in 2024 which equates to an impressive compound annual growth rate of 40%. However, people forget that for a company to be trading at such a high price relative to the cash it generates investors are assuming that this massive growth will continue for the foreseeable future. Here is what this means: in order for the company's price to match the long term average P/E of the S&P500 of 17x, Nvidia needs to continue growing with 40% per year for the next 3 years(without its stock price going up or down) in order to justify its current valuation.

This means that buying the company at its current price locks you in with a 0% return for the next three years unless you believe Nvidia's profits will grow at a faster rate than 40% per year.

Something else worth mentioning is that massive companies, and believe me, Nvidia with its $3.19T market capitalization is one of them, cannot grow at a fast pace for a long time purely because sectors that have massive growth always lure in new competitors, and Nvidia already has plenty of them. Unless Nvidia has some form of a hidden durable competitive advantage, which I currently do not see, I simply cannot justify its extremely high price.

I am aware that one company, however, does not create a bubble in the finance markets but at the same time I don't believe I could analyse the entire AI industry and still write this essay in an article-style format which was my initial intention. Also, the whole idea of piece of writing is to give you food for thought and which is why I will throw some numbers that can be interpreted in more than one way.

My good "friend" Gemini ( Google's Android Assistant) was kind enough to provide me with the following information:

There have been no new AI-related ETFs in 2021, 5 or so in 2022, 10 to 15 in 2023 and 20 to 30 new AI-related ETFs in 2024. Clearly the demand for diversified AI-related funds is growing rapidly.
Out of the 500 companies comprising the S&P500 73 of them are tech stocks. Before 2020 the tech sector returns were all over the place and were a lot more volatile compared to how it has been recently. In 2021 the S&P 500 grew by 27% and the tech sector grew by 33% which are very similar figures. Then in 2022 both S&P500 and the tech stocks went down by 20% and 10% respectively. 2023 - S&P grew by 18% and tech companies grew by 35%. 2024 - S&P grew by 15% and tech stocks completely knocked it out of the park by going up by 40%. When do you think ChatGPT came into existence? Late 2022.
When I asked Gemini if the AI bubble is real it's answer was rather interesting. It leaned more on the "it is possible" side and even went as far as giving a few names of companies that are considered to be overpriced at the moment. Those were, surprise, surprise ... NVIDIA, META, Microsoft, Amazon, Tesla, Apple, Google.
Average S&P500 price-to-earnings ratio over the last 150 years is 19.1. Currently we are sitting at above 29 according to GURUFocus . The S&P 500 Information Technology Sector P/E is even more overprice and sitting at 36.36 as of 25th January 2025. Remember the tech bubble of the 2000? Have a little look at this chart:

My stance on this

I will keep section short as I realize you have had to read a total of 2595 words to get to this point of the essay, which is starting to turn into a tiny book and trust me this was not my intention when I started writing it.

I truly believe there is a little bit of truth in all of those cases and future reality will incorporate excerpts of all three. History will repeat itself, I believe, and some professions will go while others will emerge. Stock prices of certain companies will normalize once the hype around the future of AI settles. People will just learn to embrace the strength of their own invention and simply use it to push the boundaries of human evolution.

We just have to wait and see.

P.S. I encourage you to get in touch if you think I've left out something crucial that relates to the subject, if you have some helpful critique you feel compelled to offer, or even if you would want to talk about a particular section of the article.

My email address is located in the website's "Contact me" section.

Page updated

Google Sites

Report abuse