Is AI a Bubble?

Brit Ryle

Posted March 1, 2024

Searching for the AI Bubble

Is AI a bubble? It’s tough to get a straight answer. Wall Street is probably happy to let individual investors think AI stocks are a bubble…

That way you stay on the sidelines and watch these AI stocks move higher and higher until you can’t stand it anymore…

Because Wall Street knows that the so-called “fear of missing out” (FOMO) will eventually wear you down. And that when the average investor succumbs to FOMO and starts buying in, that’s a good sign that it’s time for them to get out.

These tech CEOs, hedge fund managers, and investment bankers – they’re part of the same club. They went to the same boarding schools, sit at the same tables at $ 5,000-a-plate fundraisers and they all have each other on speed dial.

So you can bet your boots they knew Nvidia (NASDAQ: NVDA) was going to beat earnings last week. Maybe not to the penny. And they probably don’t have a copy of the spreadsheet from the CFO tucked in a drawer…

But they knew.

The prices for Nvidia call options (a bet that the stock would move higher) got pushed so high it would take a 10% post-earnings move — a $170 billion dollar addition to Nvidia’s valuation — just to break even. (Not a very good bet, unless you maybe have some reason to believe there’s a big beat coming.)

At its highs the day after earnings, Nvidia’s valuation had increased by $280 billion. Score one for Wall Street!

The irony is that many individual investors have gotten into some AI stocks (a good thing too, more about that in a minute). I expect you were aware that it was earnings time for Nvidia last week. It was like the buildup to the Super Bowl or Apple earnings back in the day. it was all the media could talk about!

You probably remember what the gist of the nonstop coverage was, too. “How can Nvidia meet such high expectations?” “Nvidia is priced to perfection ahead of earnings.”

I’ll admit: I even started to wonder if Nvidia could pull another rabbit out of its hat. 

All it takes is a little sliver of doubt…

AI Stocks for the People!

I mentioned Apple a minute ago. It serves as a pretty good example of a stock that individual investors recognized as a good investment early on. It was a fan favorite long before Warren Buffett got on board in 2016. 

So if you’ve bought into some AI stocks, good. There’s a long road ahead. 

As I wrote last week, AI is in the infrastructure phase – paving the roads that AI will run on.

That is, data centers like Microsoft Azure (NASDAQ: MSFT), Amazon Web Services (NASDAQ: AMZN), Google Cloud (NASDAQ: GOOG) along with data center REITS like Equinix (NASDAQ: EQIX) and Digital Realty Trust (NYSE: DLR) are adding super fast servers to handle the data demands of AI. That means Nvidia chips and servers from companies like Super Micro Computer (NASDAQ: SMCI)

The AI server market itself generated $30 billion in revenue in 2023. Foxconn Chairman Liu Yangwei says that the AI server market will hit $150 billion in 2027. 

Next up, the applications and services that run on these data centers. Obviously, with its big investment in OpenAI and control over ChatGPT, Microsoft is in a top position. I’ve written several times that Amazon’s position as a host for AI applications is underappreciated. For instance, Pro Trader Today stock (NASDAQ: AI) runs on Amazon Web Services (AWS), and jumped 27% on good earnings yesterday. 

Snowflake (NASDAQ: SNOW) and MongoDB (NYSE: MDB) are big data services that run on the Big 3 – AWS, Azure, and Google Cloud. We’ve written up Snowflake already, and my co-editor Christian DeHaemer says yesterday’s sell-off is a buying opportunity.

Mark? What Are You Doing With Those Chips, Mark?

Meta (NASDAQ: META) CEO Mark Zuckerberg boasts that he will have 350,000 of Nvidia’s best chip, the H100 at $30,000 each, by the end of this year. Throw in chips from other companies (AMD (NYSE: AMD) and it’s estimated Meta will own 600,000 GPUs (graphic processor units, the kind of chip ideal for AI).

What’s he gonna do with all those chips? 

Meta is a really interesting case. Because the company won’t use Azure or AWS – it’s building its own AI-worthy data centers. Plus, Facebook, Instagram, and WhatsApp give Meta a massive base of potential users for the AI applications it rolls out. 

What’s Meta gonna do with all those chips? Probably make a lot of money…

So back to the question: is AI a bubble? The quick answer is: not yet. 

Because there’s not really a very big difference between an “emerging investment trend” and a “bubble.” An emerging investment trend occurs when some new product or innovation spurs investment by companies to take advantage of the innovation, which in turn spurs investors to buy stock, with the expectation that revenue and profits will grow. 

A bubble occurs when those expectations become unrealistic. 

For instance, we might not think that something as grounded in reality as oil production had bubble potential, but it did. When Saudi Arabia flooded the market with oil in 2014-2015, the Shale Oil bubble popped. Oil prices cratered, huge sums spent on leases became unsustainable, over 100 U.S. oil companies went bankrupt…

It was unrealistic expectations for oil prices that made the Shale Revolution a bubble. 

At some point, price will be the thing that exposes unrealistic expectations for AI (notice how I didn’t call AI a bubble!). “Regular” economy stocks like CarMax (NYSE: KMX) and Coca-cola (NYSE: KO) have started dabbling with AI, using it for some customer service and marketing. 

The uses will expand, and companies will see productivity gains and improved profit margins. The stock market will go higher. Then at some point, the payoff of AI will fade. Companies will cut AI spending to maintain profit margins…

That’s when we’ll find out there were some unrealistic expectations for AI… and it’s most likely a few years from now.

Have a great weekend and I’ll talk to you Monday,

Briton Ryle
Chief Investment Strategist
Pro Trader Today