NVIDIA (NASDAQ:NASDAQ: NVDA) is currently in an interesting place. As a deeply cyclical stock, it has reacted more violently to different phases of the economic cycle, as you can see from the following chart. At the start of the year, the stock was trading in bubble territory with a P/E of almost 90x. Since then, its stock prices have plunged more than 51% since the start of the year and more than 57% from its recent peak. On the other hand, its close counterpart AMD (AMD) “only” lost m41 YTD. And the losses of NVDA and AMD are much worse than those of the technology sector represented by QQQ.
At the same time, such a large price correction has deflated the bubble to a large extent and there are a lot of positives to be said for the stock. First, due to the large price corrections, the valuation is much more reasonable now (around 47x P/E as we will see later). Second, the company’s fundamentals remain solid and growth prospects robust once the contraction phase ends. It plans to introduce a slew of new GPU, CPU, DPU and robotic processors in the near future. It just recently reported record data center and gaming volumes. And such a large volume can help significantly improve cost absorption once demand picks up.
The rest of this article will weigh the pros and cons more closely. But overall, my general conclusion is that long-term investors like me can wait for a better opportunity, while swing investors can prepare for a trading window between $125 and $200 (which translates to a P/E of around 25x to 40x) in the next 1 or 2 years.
The positive points
As mentioned above, there are a lot of positives to be said for the stock. Here I will just highlight a few highlights that are less talked about by other SA authors.
First, NVDA enjoys enormous capital allocation flexibility, and the company is in a very strong financial position even though demand has recently slowed. The following table provides a summary of NVDA’s maintenance and growth CAPEX expenditures (details of these concepts can be found in my previous article here) over the past few years on a quarterly basis. As a result, the company systematically reinvests more than its maintenance and CAPEX needs. Its CAPEX spend has been consistently above $200 million since 2020. And it peaked at $480 million in 2020 and currently sits at $433 million, very close to the top.
Second, the beauty of NVDA is that it doesn’t need to spend too much on maintenance CAPEX. And therefore, a large part of its CAPEX expenses are growth CAPEX. Its maintenance CAPEX (approximate by the total amount of depreciation) has averaged $186 million since 2018. And its total CAPEX has averaged $223 million. As a rough approximation, the difference of $37 million per quarter is on average the amount of growth CAPEX it reinvests to fuel growth. Therefore, growth CAPEX represents approximately 20% of total CAPEX expenditure. As a result, the result of its owners is much higher than its accounting result because the growth CAPEX must be added to the result of its owners, which brings us to 3rd item below.
Third, its P/E is lower than on the surface once its owners’ profits are taken into account, instead of accounting profits. The following graph shows NVDA’s true economic earnings against its accounting EPS using a more rigorous method, as detailed in my previous article here. The method used a method detailed in Greenwald’s book titled Value Investing. As can be seen, NVDA’s owners’ profits have consistently exceeded its book profits. And the ratio between them has averaged 174% since 2017. Its current OE/EPS ratio is 132%. Its current P/E is around 47x on a book basis. And after correcting for growth CAPEX, its P/E based on owner income is “only” around 35.6x.
Now the negatives.
First, there is no doubt that the stock is currently in the contraction phase of the economic cycle, as you can clearly see from the chart below. I detailed the cyclical nature of business and the fundamental reasons for the cycle in my previous article. The main points are summarized below:
- Quarterly year-over-year growth in NVDA’s operating revenue over the past 10 years shows two complete cycles over the past ten years, each lasting approximately 3.5 to 4 years. The first expansion cycle peaked in early 2014 and peaked again in early 2018, resulting in a cycle length of approximately 4 years. And the second cycle peaked in early 2018 and peaked again in mid-2021, lasting about 3.5 years.
- The bottom of the previous contraction occurred in mid-2019. If history is any indication, I expect the current contraction phase to last around 4 years as well, which translates to sometime later in 2022 and early 2023.
And the consensus estimates seem to reflect that same view, as you can see in the next chart below. Earnings revisions for the past 3 months for the year ahead paint a fairly pessimistic picture. No one expects EPS growth, and a total of 37 analysts have revised EPS lower. Consensus EPS in 2024 and 2025 is now between $4.6 and $5.6, around 32% and 39% lower than it was just 6 months ago.
Second, its current valuation is more reasonable but still high. As mentioned above, the recent correction has brought NVDA’s valuation back into a more reasonable range both relative to its historical valuation and relative to its peers. Namely, it is currently valued at around 47x P/E on a book basis or around 35.6x on an owner’s earnings basis, compared to around 90x at the start of the year. But such a valuation is still expensive. As you can see, that’s a 34% premium to AMD’s 35.3x P/E, and more than a 100% premium to AMD’s 22.0x P/E. NASDAQ 100.
Final thoughts and other risks
To reiterate, my general conclusion is that long-term investors like me can wait for a better opportunity given its still high valuation and growth prospects. While for short-term investors, there could be some interesting swing-trade opportunities ahead.
Given the deeply cyclical nature of the business and also its stock prices (see chart below), I expect a trading window between $125 and $200 over the next one to two years. These prices translate to a P/E in the range of around 25x to 40x based on consensus EPS estimates of between $4.6 and $5.6, as mentioned above. As you can see from the chart below, the stock is currently 57% off its recent high. And over the past two decades or so, the stock has undergone corrections as large as the current one 3-4 times (depending on how you count). And the last three times happened in 2003, 2008 and 2011. And all of them took around 2-3 years for the price to reach the pre-correction level. If the story is telling, that’s how long I expect the title to spend in this downturn as well.
Finally, the risks. Besides the high cyclicality, NVDA also faces many other risks. These risks include macroeconomic risks, high inflation, supply chain disruptions, etc. NVDA’s products also have considerable exposure to cryptocurrency mining activity. And its high-end, high-margin products are particularly sensitive to the possibility of a recession. All these risks have been detailed by other SA authors and also by myself previously. Here I will only focus on two risks specific to the projected trading window in this article.
First, it is never easy to time the cycle precisely even if the cycle has repeated regularly in the past. Second, as you can see from the chart below, the cycle can sometimes be so deep that the maximum appreciation can occur at the bottom of the cycle, as indicated by the green line below. And vice versa, bottoming out can occur at peak prices, as indicated by the red lines below. Long-term investors and swing traders should be aware of these pitfalls.