AI – killer robots or a massive opportunity?

June 08, 2023

Dmitrijs Brizgalovs

BluOr Bank Investment Portfolio Manager

ChatGPT – a chat bot that is designed to generate human-like responses on given prompts or questions – has generated more excitement about artificial intelligence than any previous iteration of AI technology.

Since ChatGPT became widely available, I too have been fascinated with its ability to generate human-like conversations. My first conversation started with a simple question: “what is AI?” And right before my eyes ChatGPT started typing: “AI, or Artificial Intelligence, is the development of intelligent systems that can perform tasks and make decisions that typically require human intelligence.” Awesome!

Then of course, I followed with a question that is on everyone’s mind: “Will the AI take over the world?” To which ChatGPT responded: “No, the idea of AI taking over the world in a malevolent or dominating way, as often portrayed in science fiction is not supported by the current understanding and capabilities of AI technology.”

Okay, there will not be a killer robot world takeover anytime soon. Or maybe it’s what it wants me to think? I’ll leave this for another discussion.

As I grew in confidence about the level of response, I began to test its true colors.

Being a professional investor, I am very interested to know how I could profit from an increasingly AI driven world so naturally the next thing I did was ask ChatGPT its thoughts on the market and possible investment recommendations. Sadly, all I got was a generic answer that I could have read in a finance related textbook without any specific recommendations and with a disclosure in the end that investing is complicated and you should consult an investment professional. Maybe ChatGPT is still building its investment positions and does not want to tell anyone yet. I will keep on asking, but at least for now it looks like my job is safe. However, looking ahead I am convinced that AI will dramatically change how we live and work.

Sam Altman, the man behind OpenAI and therefore ChatGPT when recently speaking about the future effect of AI at the US Senate hearing said, "There will be an impact on jobs. We try to be very clear about that." Additionally, according to Goldman Sachs estimates, the global need for 300mn full-time jobs may be reduced to AI-powered automation and overall roughly two-thirds of current jobs are exposed to some degree of AI automation.

Should we be afraid? Or should we be excited?

I think we should be cautiously excited. In the past, every new advancement in our civilization has increased the capabilities and the efficiency of what can be achieved, thus increasing our living standards. Advancements in agriculture, the printing press, the Industrial Revolution, electricity, the Internet are some major examples. And now you can carefully add AI to this club.

Just like these other innovations, AI will destroy some jobs but it will create others. Additionally, it will make us more productive in our current roles by automating routine tasks, enabling us to focus on more complex creative work. There is already lots of anecdotal evidence that by leveraging AI applications people have the ability to simultaneously work in multiple full time jobs. In other words, their productivity has increased dramatically.

This all sounds very promising! But if you are an investor you might be thinking how can I invest my capital to benefit from this new technology?

If we were to look at the broad AI ecosystem, we can distinguish between three types of layers: applications, models, and infrastructure.

Applications are the end products that are either integrated into already known platforms and products or are standalone apps that people and organizations use to leverage the new AI capabilities (Microsoft Office 365, Grammarly, GitHub Copilot, Google Assistant).

Models are algorithms that power these applications. They are the brains behind the curtains. They are trained on a vast amount of data and based on the inputs provided they generate output that tries to mimic human intelligence and creativity (ChatGPT, Bard, Dall-e).

Infrastructure is the backbone of the whole system where all the computations actually happen. This can additionally be split into cloud services (Amazon AWS, Google Cloud, Microsoft Azure) and hardware design (Nvidia, Intel) and manufacturing (TSMC, ASML).

As you can see, an investor can play this AI game from multiple angles. Several well-known companies are already integrated in this ecosystem and a lot more new companies will be created that will try to benefit from this strong new trend. Why do I feel so sure about this? Because this has already happened before.

Despite the ongoing battle with persistently high inflation, high interest rates, bank failures, possible US recession and debt ceiling crisis, S&P 500 Technology index that measures the performance of US largest technology companies has rallied around 27% this year while the whole S&P 500 index only gained 9%. Moreover, if you strip out technology stocks from the index, the performance further drops to just 3%. The difference is very significant!

It is safe to say that currently there is hype and craze around all things AI. This does look a little familiar. The same market dynamics could be observed back in the late 90’s when the internet started to take off. During that period, investor interest in the technology companies was so strong that the capital flew freely to both large corporations as well as startups that had no track record of success. It was all about the story and the allure of potential outsized returns sometime in the future. The hype was so powerful that basically all you needed to do was to put “.com” after your company name and you instantly became a paper millionaire. This type of unsustainable behavior eventually formed the now infamous “Dot Com Bubble”.

I thought that it would be a good idea to ask our good friend ChatGPT about what percentage of companies during the “Dot Com Bubble” where actually profitable, and its response confirmed the previously stated point: “Estimates suggest that a significant percentage, possibly over 70%, of the companies during the Dot Com bubble were not generating profits. Many of these companies relied on speculative investments and anticipated future growth rather than immediate profitability.”

We all know what happened in the end. The bubble popped, as it always does, and technology stocks crashed more that 80% from their peak in March 2000 to the bottom in November 2002.

Asset bubbles are inevitable. They are part of financial markets. From time to time, they form when something new and exciting comes along and people lose their minds by forgetting the fundamentals, only focusing on the story, the hype, and the promise of enormous returns.

Remember crypto? Remember all the generated interest, speculative buying, startups popping everywhere funding their ideas from ICOs, and super star like personalities? This all culminated in November 2021 when Bitcoin was trading around 56 thousand USD. This bubble, just like all of its predecessors, popped and already year later in November 2022 Bitcoin was trading around 15 thousand USD crashing more than 70%.

Coming back to today and seeing similar trends starting to emerge around AI it is reasonable to start asking ourselves if we are letting our imaginations run away from us again. The valuations of some of the companies responsible for this year’s rally are approaching very elevated and questionable levels.

One of the biggest winners of this AI race is Nvidia. It has had a massive rally appreciating almost 120% this year. Its Price/Earnings ratio is 177 and Price/Sales is 29. Let’s put that in perspective – Price/Sales ratio of 29 implies that the company has to pay back 100% of its revenue to shareholders as dividends for the next 29 years. That assumes zero cost of goods sold, zero expenses, zero taxes paid. If that sounds unreasonable to you then join the club. Investors are betting big that AI disruption will generate unprecedented growth rates to justify these out of this world valuations.

Despite this, I think we still have more ground to cover. We have not yet hit the part of the mania with IPOs and companies adding AI to their name to generate excitement. The bubble seems to be forming, but this trade looks far from finished so the market still has more room to run before the eventual repricing kicks in.

George Soros, one of the most successful investors of all time, has famously said: “When I see a bubble, I rush in to buy it.” It is not an absurd strategy and people have made loads of money this way. The hard part is to know when to get out before it becomes messy. Investors that invest in bubbles need to be honest with themselves and be modest enough to ask themselves: “will I know when will be the time to pack my things and go, keeping in mind that market timing is very hard, even for the best of the best?”

We are living in a very exciting time where AI will and is changing the world and a lot of great investment opportunities will be created along the way. But as prudent investors, let us be mindful to keep a critical eye on valuations and their actual implications before they get so out of hand that they cease to make sense in both the physical and artificial realm. Because eventually when the bubble pops, as it always does, you better not be on the wrong side of this trade.

Just remember, hot technology companies are not the only game in town for potential outsized returns. That is the great thing about financial markets –opportunities are everywhere, you just need to look hard to find them. And sometimes the best returns come from sectors where no one is looking.

The Latvian version of this article originally appeared in the June 2023 issue of Forbes Latvia.