VIRT - Important Warning About Artificial Intelligence Stocks

2023-10-08 09:00:00 ET

Summary

• Artificial intelligence stocks are surging.
• Some investors have made little fortunes already.
• But we are staying away from them, and here is why.

Co-produced by Austin Rogers.

The latest buzzword on Wall Street is artificial intelligence , or "AI."

More specifically, "generative AI." This type of AI refers to the ability of a software program like ChatGPT to generate a certain type of content such as written text, images, or even video based on a given prompt.

It can also be used to generate huge swathes of computer code, says NVIDIA Corporation ( NVDA ) CEO and co-founder Jensen Huang, which puts the world at "the tipping point of a new computing era." In this new era of computing, Huang says that effectively anyone can become a programmer. Just plug in the specific prompt, and the AI software will generate all the code.

And ChatGPT isn't the only example of a generative AI platform to emerge this year. Google ( GOOGL ) rolled out its latest version of Google Bard, and Microsoft ( MSFT ) just made Bing Chat (powered by the same technology as ChatGPT) available for limited use.

How did this AI revolution happen? To borrow the words of Ernest Hemingway to describe how a character in one of his novels went bankrupt: "slowly, then suddenly."

This AI revolution will require certain specialized computer chips, which will obviously benefit the world's leading chipmakers, foremost among them being NVDA. Hence why the stock has surged from around $300 billion in market cap to nearly $1 trillion today, gaining 212% year-to-date.

That's how NVDA was the first chipmaker to breach the $1 trillion in market cap!

This Cambrian Explosion of AI technologies is expected to cause AI investment spending to roughly double from 2021's $33 billion to $64 billion in 2025. Few doubts at this point that AI will have profound social, economic, and political impacts in the future and knowledge of this is causing quite a stir among investors.

Though there are undoubtedly riches to be made from AI, this setup sounds to us like the makings of a bubble.

Take NVDA, for example. The company is clearly at the forefront of the AI Cambrian Explosion with its specialized chips and supercomputer. But everyone knows this and has bid NVDA's stock up at an extremely rapid rate.

And even after a huge increase in its earnings guidance this year, NVDA still trades at a forward P/E ratio of about 46x, which is significantly higher than its 5-year average of 39x.

In the next twelve months, NVDA is expected to generate as much free cash flow ("FCF") as it has in the last four years combined. And yet, the stock still trades at 45.5x FCF. That means that if NVDA paid out 100% of its FCF to shareholders, it would take 45.5 years to get back one's invested dollars at today's price.

And it's not just NVDA that's catching the bid from investors who want to own the newest and coolest tech. It's all kinds of companies that label themselves as "AI" or "automation" companies. See, for example, these exchange-traded funds, or ETFs, that all aim to gather baskets of such companies:

• ARK Autonomous Technology & Robotics ETF ( ARKQ )
• iShares Robotics and Artificial Intelligence Multisector ETF ( IRBO )
• Global X Robotics & Artificial Intelligence ETF ( BOTZ ).

Stocks that proclaim to specialize in "automation," "robotics," or "artificial intelligence" are far outperforming the S&P 500 (SP500) this year -- by double, triple, or more.

Is this an artificial intelligence stock bubble (or mini-bubble) in the making?

We think it could be.

This potential bubble in AI stocks appears to be replacing a mini-bubble in certain high-quality consumer staples stocks like Mondelez ( MDLZ ), PepsiCo ( PEP ), and General Mills ( GIS ), each of which had recently run up to all-time highs as investors piled into defensive stocks ahead of a widely anticipated recession.

As the air comes out of consumer defensive stocks, it seems to be inflating AI stocks.

A Peter Lynch-esque Approach

We are not momentum investors or fad-chasers.

When it comes to potentially world-changing technologies like AI, we tend not to invest directly in the players involved in creating and disseminating the technology. Why? Here are a few reasons:

• Competition : Attention and enthusiasm attract investment dollars from a wide variety of sources, as lots of companies want to get in on the action and capture a portion of this burgeoning industry. We saw this during the crypto craze when some companies that were completely unrelated to crypto changed their name to include something crypto-related and saw an immediate pop. More companies trying to get in on the action means more competitors vying for customers.

• Unpredictability : The direction toward which brand-new technologies will go is inherently unpredictable, and therefore picking winners is extremely difficult.

• Timing : Even if you successfully identify long-term winners, buying them in the midst of the bubble is often treacherous. Take the example of MSFT. Obviously, MSFT was a huge winner from the invention and dissemination of the Internet, but if you'd bought the stock at its peak in 2000, it would have taken 16 years for you to get back to even after the dot-com bubble popped.

Generally speaking, we like the approach of Peter Lynch, the famed investor in the 1980s who was a self-proclaimed "technophobe" and largely avoided tech stocks during the dot-com bubble.

Lynch's assertion was simple: "Users of technology are the biggest beneficiaries of high-tech."

Why? Because users of new technologies get their benefits without investing heavily to create them or competing fiercely to capture market share. Quoting Lynch from his book, One Up On Wall Street (pg. 142):

Instead of investing in computer companies that struggle to survive in an endless price war, why not invest in a company that benefits from the price war - such as Automatic Data Processing?

Lynch had a good point. Human resources software provider Automatic Data Processing ( ADP ) did run up during the dot-com bubble and see a decline after it popped, but not nearly to the degree that tech stocks ( QQQ ) did. As such, measuring from the beginning of April 1999, right in the midst (but not the peak) of the dot-com bubble, ADP handily outperformed tech stocks over the next two decades.

This despite the QQQ outperforming ADP over the last 1-year, 3-year, 5-year, and 10-year periods.

Just as we explained in our article discussing why we don't invest directly in robotics and automation stocks, we prefer investing in the users and facilitators of AI rather than the developers and enablers of it.

An example of an AI developer is C3.ai, Inc. ( AI ), an AI software company that is up almost 150% year-to-date but has never turned a profit. An enabler would be NVDA with its specialized chips and supercomputer products on which AI is operated.

What about the users and facilitators?

An example of a user of AI would be Medtronic ( MDT ), a medical device maker that is partnering with NVDA to use AI for various medical uses, such as the detection of polyps during colonoscopies. MDT has its own AI software program called GI Genius, and the partnership with NVDA is meant to (quoting Geoffrey Martha from the fiscal Q4 conference call ) "allow third-party developers to train and validate AI models that can eventually run as apps on the GI Genius platform."

Facilitators of AI would be the data center REITs that house the servers used for generative AI tasks. The two biggies in the U.S. are Equinix ( EQIX ) and Digital Realty Trust ( DLR ).

Both of them have gotten a slight bounce from the sudden interest in AI, but their roles in facilitating the myriad future uses of AI could still be underappreciated.

Our AI Game Plan

So, what is our preferred way of investing in artificial intelligence? Who are the users and facilitators of AI that we believe will benefit most from the explosion of this new technology?

Candidly, it is difficult for us to find a company or industry that won't be affected by AI, most of them in a positive way.

• Software Programming : As NVDA CEO Jensen Huang pointed out recently, AI should dramatically increase the speed and efficiency of incremental innovation and productivity in software programming. Much of it will soon be handled by AI.
• Graphic Design : Much the same could be said for graphic design. Rather than taking the time to manually create graphic designs, no matter how simple or rote, AI can allow graphic designers to simply input prompts for AI software to create designs instantly.
• Banking : AI could improve customer service as well as decrease risk in the loan portfolio through applying more extensive risk management techniques than non-AI software algorithms.
• Consumer-Facing Businesses : Whether it's a brick-and-mortar retailer or an online-only brand, AI can significantly increase the efficiency and customer satisfaction of customer service. We all know that those phone trees could be improved! And imagine an AI chatbot that could actually solve the issue you called about instead of waiting 30 minutes for a human representative to become available.
• Energy : AI should improve predictive models for the very difficult to predict prices of oil and gas, assisting energy companies in capital allocation decisions such as where, when, and how much to drill, where to put pipelines, etc.

We are bottom-up value investors, not trend chasers. We don't plan to make any investments based on AI factors alone.

But we do look forward to seeing how the companies in our portfolio use and take advantage of new technologies to grow their revenues and increase their profits.

One example of a portfolio holding that is already using and marketing its own AI products is a financial services company called Virtu Financial ( VIRT ). Among other financial services, VIRT recently (two years ago) rolled out a few data analytics products that utilize machine learning.

To quote VIRT's Head of Analytics, Erin Stanton, "Over the next one to three years, integrated access and directed machine learning and artificial intelligence will be key drivers in Transaction Cost Analytics ((TCA))."

In other words, VIRT's AI products are going to become increasingly important to their customers. This exposure to AI isn't the sole reason we like VIRT, but the flexibility and diversity of VIRT's capabilities, which include AI products, is a big reason we like the company.

Takeaway

Artificial intelligence seems at this time to have the potential to be as world-changing as the Internet was. But just like it wasn't necessarily a good idea to buy Internet stocks in the early days of that boom, it also isn't necessarily a good idea to buy AI stocks in the midst of its current boom.

We suggest that a better way to think about AI is how it will affect the companies we already like and own. We refuse to chase trends, but we are always happy to see new technologies increase the efficiency and profitability of our high-yield dividend stocks.

For further details see: