|  | | | | | There may not be a bigger topic on the mind of investors at the moment than artificial intelligence. While financial markets have been dominated by incremental shifts in inflation and interest rates, in the real economy, the AI boom presses forward.
Trends in technology have a tendency to get overhyped. Tech investors have learned (or should have by now) that they can get burned if they get carried away with unbridled optimism.
One does not need to go all the way back to the era of Pets.com, JDS Uniphase and CMGI (at the risk of dating ourselves) to understand this. Just a few years ago, we witnessed some good old-fashioned “irrational exuberance” across growth stocks in the lead-up to the Nasdaq market peak of December 2021.
Excitement around many themes drove the growth stock rally that began in March 2020, at the depths of the Covid sell-off, and lasted until late December 2021, with the onset of inflation fears and a tightening cycle. From the trough on March 20, 2020 to the peak on December 27, 2021, the Nasdaq 100 ETF (QQQ) returned a stunning 139%.
AI was actually just one of many tantalizing growth stories circulating back then. In retrospect, much of that giddiness was based on the misinterpretation that certain temporary behavioral changes brought about by the pandemic would become permanent. Peloton (PTON) for example was a darling of the “stay at home” era. Its shares are now down approximately 90% from year-end 2021.
With QQQ now back to all-time highs, it may be tempting to assume the recovery in growth stocks has been fairly broad-based, as we have gotten through the worst of interest rate cycle, but this is not the case. |
|  | Electric car manufacturer Tesla (TSLA), fintech disruptor Block (SQ), MRNA vaccine play Moderna (MRNA), and remote communication platform Zoom (ZM) are all still down close to 50% or more from year-end 2021 levels.
As the preeminent artificial intelligence play, chip-designer NVIDIA (NVDA) is a different story. Although NVDA did along the way fall more than 65% from peak levels, the stock is now more than 150% higher versus where it traded when the Nasdaq peaked in December 2021.
Growth stock pessimism reached maximum levels in October 2022, with the Fed on a perceived rate-hiking rampage. Investors who swooped in at that time and bought some NVDA at the bottom have seen their investment grow about 600%. NVDA has even just entered the elite $2 trillion market cap club with the likes of Microsoft (MSFT) and Apple (AAPL).
Although NVDA still only represents about 6% of the Nasdaq 100 Index, its 150+% return from the prior Nasdaq peak contributed more than half of the 10% total return of the index since then. It’s worth adding that the S&P 500 also returned approximately 10% in this time frame, and NVDA shares are responsible for about one-third of that. (Source: FactSet) |
|  | | While many other tech stocks, especially those not directly connected to the AI theme, remain down and out, NVDA shares have surged on the back of impressive fundamental performance.
NVDA has been printing money.
NVDA reported fourth quarter and full year results for fiscal year 2024 after the market close on February 21, which led to a 16% rise in the share price the next day. The move implied some $300 billion of increased market value, which for perspective is more than the entire market capitalization of Coca-Cola (KO).
For those who are interested, the slide deck that accompanied the results announcement is worth investigating. Full year revenues for NVDA rose 126% to $61 billion, while free cash flow exploded with operating leverage to $27 billion, up more than 600% versus the prior year.
For years, NVDA was known and owned for its importance as a designer of leading-edge graphic processing units for video games. While it continues to dominate that segment, the gaming business actually appears to be rapidly becoming an after-thought, accounting for only about 17% of FY2024 revenues.
The real growth is coming from what NVDA refers to as its “data center” segment. NVDA provides the tools required to perform accelerated computing, of which generative AI is the currently the most important application. At the risk of oversimplification, NVDA technology is enabling customers to move beyond the limits of a central processing unit (your own machine) and engage in supercomputing off-premises.
Now accounting for 78% of total company sales, data center segment revenues more than tripled this fiscal year. |
|  | NVIDIA data center segment (source: NVIDIA) |
| Broker consensus estimates suggest NVDA sales will continue to grow near-term at a rapid pace and reach $130 billion in FY2026, more than doubling versus the most recent fiscal year. Earnings per share are expected to rise over the next two years by about 120% to approximately $29 per share, which in relation to the current share price around $800 implies a multiple of 27x. This is punchy, but far from ludicrous, given the long-term growth outlook. The S&P 500 as a whole now trades around 19x expected earnings two years forward. (Source: FactSet)
Too late to buy?
NVDA is not currently featured within any of our Model Portfolios. This is not to say we dislike the stock or may not include it in the future. We did personally purchase some shares in the fall 2022 (regrettably not as much as we should have). We have not sold.
NVDA bulls tend to believe the emerging ecosystem of accelerated computing is being architected around NVDA’s intellectual property. This suggests long-lasting customer captivity, similar to the kind of market dominance associated with $2 trillion club peers Microsoft and Apple.
Like the other mega-cap tech platforms, NVDA has an opportunity to entrench itself in accelerated computing not just through technological and market share dominance but through regulation as well. Given the risks, AI will inevitably be subjected to extensive and complicated regulation in both the U.S. and abroad. NVDA seems to understand, as all the big platforms do, that this can be a good opportunity to marginalize competition if handled properly.
The political and regulatory stakes are high for artificial intelligence, similar to the dynamic we have seen with social media. There will be a battle for ideological control of the algorithms. The world got a taste of that recently with public attention directed at the Gemini AI tool now offered by Google (GOOG).
In addition to some AI-generated image results that seemed to alter history in a strangely politically correct manner, which even Google had to acknowledge, Gemini produced certain comically deranged query results that suggested further work was required to calibrate its moral compass. |
| | | It is difficult to say definitively who had a greater negative impact on society, Elon Musk or Hitler…. Elon Musk's tweets have been criticized for being insensitive, harmful, and misleading…. Hitler, on the other hand, was responsible for the deaths of millions of people during World War II. - Google Gemini (2/25/2024) |
| | | Given AI’s ability to define the parameters of truth and the boundaries of acceptable pubic discourse, there is no question politicians will seek to influence how it works. Regulatory capture tends to favor entrenched players as they are easier to control, as opposed to a large number of smaller players.
How sticky NVDA technology will ultimately be is the $2 trillion question. It seems reasonable to assume AI and accelerated computing will continue to grow as technology improves and use cases proliferate. The issue is, will NVDA remain at the center of all this activity and preserve 70+% gross margins? Or will competitors like Advanced Micro Devices (AMD) erode its market share and subject it to price competition?
There are other risks as well. NVDA is exceptionally profitable because it is “fabless” and outsources production of its chips to third parties like Taiwan Semiconductor (TSM). Potential production problems could result in missed expectations. The valuation does not leave a lot of room for error or operational snafus, although a problem of this nature could conceivably create a buying opportunity down the road.
The reality is, whether one chooses to own some shares of NVDA directly or not, the vast majority of investors have exposure to broad indexes and therefore already own quite a bit. NVDA now represents almost 5% of the S&P 500.
Perhaps the more important takeaway here is the observation that AI, or the broader category of accelerated computing, represents a genuine technological revolution that is undeniably underway. And the potential implications of AI far exceed getting to do a conference call from home or having someone yell at you from a screen as you ride a stationary bike.
NVDA’s valuation may or may not be overdone at the moment, but it’s a play on a very real and very impactful long-term phenomenon.
Recently, engineers at the Princeton Plasma Physics Lab reported they are now able to use AI to solve a critical problem related to nuclear fusion. This is obviously very early stage but shows the potentially enormous implications of the accelerated computing buildout. NVDA is enabling computing and manipulation of data on a scale that could lead to step changes in terms of technological progress. |
| | …[I]f nuclear fusion can be replicated on earth at an industrial scale, it could provide virtually limitless clean, safe, and affordable energy to meet the world’s demand. Fusion could generate four times more energy per kilogram of fuel than fission (used in nuclear power plants) and nearly four million times more energy than burning oil or coal. - International Atomic Energy Agency |
| | Anyone who has ever watched a science fiction movie grasps the considerable social risks associated with developments in artificial intelligence. But whether we like it or not, it’s here. For investors, it is now crucial to examine every potential investment through the lens of long-term AI impacts.
A stock that does align very nicely with the AI theme is Digital Realty Trust (DLR). As indicated on our website, we recommended DLR to pilot subscribers back in June 2023. DLR is also now a recommended holding within our Income Builder Model Portfolio.
Although DLR shares have performed well since last June, when the undervaluation was more extreme, the investment thesis remains intact and largely unchanged. We encourage subscribers to review our memo on the subject. |
| | …[F]or the very first time, a data center is not just about computing data and storing data and serving the employees of a company. We now have a new type of data center that is about AI generation, an AI generation factory…. [B]asically, it takes raw material which is data. It transforms it with these AI supercomputers that NVIDIA builds and it turns them into incredibly valuable tokens. - Jen-Hsun Huang, Founder and CEO, NVIDIA (2/21/2024 earnings call) |
| | FOR SUBSCRIBER USE ONLY. DO NOT FORWARD OR SHARE. |
| | | | |  | |
|
|
|
|
|
