Hi friends!
There has been much talk about AI and almost as much about a supposed AI bubble. But is there a bubble? This will be a short and hopefully sweet post in which I share my thoughts quickly about bubbles and if AI is the new bubble.
How do you recognize a bubble?
You may wonder how you will recognize if there is a bubble if this is not one? It’s quite simple actually, because we have just seen a bubble. Let’s take the crypto bubble we have just seen. Mind you, I’m not claiming that Bitcoin won’t return to all-time highs over time. I have really no clue whatsoever about crypto in general. But I take it as an example of a bubble. The characteristics are easy to apply.
First, you have a broad mass movement of people investing in the asset. People talk about it at parties how their crypto did so well. As they say, once the shoeshine boy or the taxi driver starts talking about it, you are in a bubble.
Secondly, if there is a bubble, the bears will be attacked fiercely by the bulls. They will be called stupid, not understanding the new paradigm, the sea shift, the new future. They will feel isolated, sanely thinking in a world that seems to have gone crazy. Early investors will beat their chest and say how smart they were investing early. They will be seen as the new gurus. But often they were only lucky, not smart or skillful.
Third, many other players will want to profit from the bubble by jumping on the train. Think about companies taking some portion of their balance sheet in crypto. Their stock prices shot up just because of that fact alone. Think of the thousands of ICOs (initial coin offerings) of all sort of crypto coins that were completely worthless but made some people very rich.
The fourth sign that there is a bubble is when new rules pop up to “rationalize” the new paradigm. During the dotcom bubble, “eyeballs” were all that mattered, for example. Also think about NFTs, for example, which were not just glorified jpegs, but “a new paradigm in art.” Now, I don’t claim that there is no value in the NFT system, but real change takes years, often even decades, not a few months.
Fifthly, you can also know for sure there is a bubble if it is already more advanced. You see that some players turn out to be frauds. I don’t have to tell you about FTX and many others in the crypto world who were not exactly striping the t’s and dotting the i’s. From about 2009 to 2015, there was a biotech bubble. It’s one of the reasons why Elizabeth Holmes found fertile ground for her Theranos scam.
There have been many bubbles. SaaS stocks were also in a bubble because many investors, myself included, though the pandemic conditions would last for much longer and change our society more profoundly than they really did.
Is AI The New Bubble?
Many investors have asked me what they should do with their Nvidia position. They are happy with the big jump in the share price but also worried about a potential bubble. The stock is up 169% year-to-date.
Nvidia’s rocket stock price caused many to shout: "AI BUBBLE!" I will be clear here from the start. There is no AI bubble, at least not yet. That doesn't mean it couldn't come, of course, but up to now, valuations are definitely not crazy at all. More about that later in this article.
Many bears thought Nvidia would drop hard after earnings, and, granted, the stock was not cheap. While Nvidia had a good quarter, the quarter didn't justify the 25% surge. But what really blew everyone's mind was guidance. The revenue consensus for Q2 stood at $7.18 billion, which would have meant 19.5% revenue growth, which is not bad already. But Nvidia guided for $11 billion in revenue, up 65% year-over-year.
The result is that Nvidia's stock has actually become cheaper, despite the stock price surge. It now trades at a forward PE of 42. Is that expensive? I have often said valuations are tough and depend on your time horizon.
I love Peter Lynch. He introduced the PEG ratio or price/earnings/growth. If you look at the forward PE, Nvidia, with its updated guidance, now trades at 42 times forward PE. Revenue growth is expected to be around 60% now. That means a PEG lower than 1, which Lynch always called cheap. On top of that, Nvidia now trades around its long-term rolling forward PE. This chart goes all the way back to 2016.
While there is much difference between the valuations, the peak valuation was quite a bit higher than now. Let’s take the chart of Nvidia’s stock price over the same period.
As you can see, the stock price has been volatile too, but it’s up 4,670%. Every $10,000 invested in Nvdia in 2016 is now worth $468,000. That’s the stuff dreams are made of.
Attentive readers will have noticed something. Peter Lynch didn't apply the PEG to revenue growth, as I just did, but to earnings growth. But I took revenue growth to be conservative. Earnings are expected to grow by over 100%, so the 'real' forward PEG is even much lower.
Many people just look at the stock that has gone up so much and therefore think this is an AI bubble, but I don't see it yet. Let's look at the three most prominent AI players:
Yes, Microsoft is expensive. But it's not bubbly expensive. Most people only look at the stock price. They see that Nvidia's stock is up almost 250% from the bottom and call it a bubble without looking at the fundamentals.
If we apply the characteristics of bubbles to AI, I don’t see the bubble yet.
ChatGPT is a topic of conversation now, but I have not heard any conversations in my daily life between people who are normally not interested in investing about which hot AI stocks you should buy. Have you?
I also don’t see Nvidia bears being attacked, ridiculed and isolated. To the contrary, there were many bearish takes about why Nvidia was “so ridiculously overvalued.” That’s a good sign of a healthy market.
Yes, many companies now name AI during their earnings calls. But I have not seen craziness in the stock price resulting from that. In 2017, during the previous crypto bull market, a tiny drinks company, Long Island Ice Tea, changed its name to Long Blockchain Corporation. The stock shot up 432% in a single day. I don’t see such outrageous things now.
I have also not seen new metrics popping up. If stocks suddenly start moving on the amount of VRAM a company can produce on its GPU chips, that would be a red flag.
As far as I know, I have also not seen any fraudulent constructions in AI.
That doesn’t mean that AI doesn’t have the potential to become a bubble, but the potential alone is not enough. C3.ai, which has the ticker AI, is down 27% in the last two days on very mediocre results and that’s very healthy.
So, to answer the question of many, what I'm going to do with my Nvidia position? Nothing. I’m not selling, I’m not adding. Of course, to each their own. If you want to lock in some gains, that's fine. But, please, don't believe the "it's only a bubble" argument.
These articles are free for 14 days. After that, they go behind the paywall. So make sure you subscribe to read them for free!
In the meantime, keep growing!
AI/IoT is the next wave of semiconductors and this new wave of the long history of semis I think is just getting started. Volatile, yes. But I’m holding and will be watching for those red flags you mentioned in the coming years.