There's one stock that everyone wants to talk about: Nvidia (NVDA +5.93%). And for good reason. The stock has gained nearly 1,000% in less than two years, meaning a $10,000 investment in late 2022 would be worth almost $100,000 today.
While nobody can predict what will happen, there are two paths forward for Nvidia. One path is promising; the other is downright terrifying. Let's dive in and see which is more likely.
Image source: Getty Images.
The bull case: Nvidia is Apple circa 2004
The bull case for Nvidia goes something like this: Artificial intelligence (AI) is the story of the next 20 years. It will bring changes to the way people interact on a level equal to or exceeding those brought on by the smartphone in the mid-2000s. Therefore, the best historical comparison for Nvidia is Apple in 2004.
NASDAQ: NVDA
Key Data Points
Back then, Apple's hot new product was -- wait for it -- the iPod. Remember that? The company wouldn't debut its first iPhone for another three years.
Nevertheless, Apple was already well on its way to becoming the world's largest company. In 2004, its stock grew by more than 200%, making it the top performer in the S&P 500.
From 2004 until today, its market cap has surged at a compound annual growth rate (CAGR) of 36%. A $10,000 investment made 20 years ago would be worth $4.8 million today.
AAPL total return level data by YCharts.
It would be incredibly difficult for Nvidia to repeat that performance, but it is possible. AI is a fascinating new technology, and 20 years is a very long time, giving the company plenty of runway to live up to the lofty expectations surrounding its stock.
But what about the bear case?
The bear case: Nvidia is Oracle circa 2000
This is the bear case for Nvidia: AI is a promising technology but overhyped. Due to elevated valuations in AI stocks, a stock market bubble has formed, and when it pops, Nvidia's stock will feel the pain. In short, Nvidia today is like Oracle in 2000.
To recap, the 1990s were full of hope for the internet. In particular, the stock market went gaga for internet stocks. Valuations soared, not just for brand-new companies but also for legacy tech companies that were cashing in on the internet craze.
In particular, Oracle, whose business revolved around facilitating the transition from mid-century mainframe computers to modern servers for enterprise clients, saw its stock skyrocket in the late 1990s.
Unfortunately for Oracle, expectations ran far ahead of actual sales. At its peak in 2000, its stock traded with a price-to-sales (P/S) multiple of 24, far above its lifetime average of 5.
ORCL PS ratio data by YCharts.
When the dot-com bubble burst in 2001, Oracle experienced an 85% drawdown. The stock has taken over 20 years to recover its pre-bubble high. Those who bought in the summer of 2000 and held through 2020 would have made almost no money on their investment.
It's a lesson that even companies with the most promising technological innovations and solid business models don't necessarily make great investments.
So which is it?
On the one hand, Nvidia's P/S ratio has a scary resemblance to Oracle's in the run-up to its crash in 2001.
ORCL PS ratio; data by YCharts.
That should give pause to any Nvidia bull. Simply put, the stock is so expensive now that some moderation must occur.
Either the company's sales will continue to grow like a weed to bring down the ratio, or the stock price will have to come down. Factors such as increased competition, market saturation, or a slowdown in AI and semiconductor spending could decrease Nvidia's P/S ratio.
My gut tells me it's the latter, although the party could continue for some time. Either way, I recommend caution with Nvidia stock.
Unlike many of its mega-cap peers, the company lacks a well-diversified business model. Its stock is performing well because AI and semiconductor spending has gone through the roof, but that trend may taper off -- particularly if the economy deteriorates or enters a recession.
Moreover, if the demand for AI and semiconductor technology were to decline, Nvidia's stock could be in for a rough ride -- perhaps as bad as Oracle's 23 years ago.
