Nvidia: Seeing Growth Near-Term (Upgrade), Doubling Down On Skepticism Long-Term

Summary:

• NVIDIA Corporation continues to experience high demand and strong financial performance, and has a reasonable valuation, leading to a bullish outlook in the near-term.
• However, given its high, unsustainable margins, competition from alternative solutions (with 2-10x lower TCO) could lead to a substantial decrease in pricing power and market share in the long term.
• Investors should hence monitor the potential risks and consider taking profits in the medium term as revenue may eventually reach its peak.
• Analog examples include profits of oil companies in function of oil prices, or how Intel’s profits have shrunk in part due to AMD’s rise. Nvidia is facing competition from both Intel and AMD.
• Other risks include the unsustainability of its FCF/operating margin expansion (without which profits wouldn’t have increased as much), and lower demand if Nvidia’s customers don’t generate enough profits from those GPUs.

Investment Thesis

In the near term, NVIDIA Corporation (NASDAQ:NVDA) clearly continues to see incredible demand, which has and continues to propel revenue, earnings, and the stock prices. So, with no catching up to demand in sight for the next 1–2 years (according to Nvidia), and given an increase in earnings that has outpaced even the stock growth, clearly the only valid call for Nvidia would be a bullish one. Although with a $2T market cap, the high-growth multi-bagger gains have already been made, but that is reflected in the lowered valuation that warrants the bullish outlook.

That’s the near-term. For long-term investors, however, this simply isn’t a stock that one can just buy and forget about, even if GAI is a new generational tech trend. The reason is because Nvidia’s margins (both gross and operating) are so incredibly high that any alternative/competitive solution has a TCO of at least 2x (50%) less. Why spend over $50B (all its customers collectively) on Nvidia hardware when other products could deliver the same for far less than $25B? So, as competitors, notably Intel Corporation (INTC) and Advanced Micro Devices, Inc. (AMD), as well as cloud service providers with their in-house designs, ramp their own products, in principle market dynamics should necessitate Nvidia to drastically lower prices and/or concede market share.

Nvidia is capitalizing on the GAI gold rush, which in the near term has no signs of slowing down (and has allowed Nvidia to charge such high prices), and this GAI trend is clearly moving faster than any market dynamics could, but that rush might not last forever and/or might eventually go through a commoditization/competitive phase where profits (and hence the stock) will inevitably shrink.

Similarly, one might liken Nvidia’s earnings power to the oil industry. At times of high demand and short supply, oil prices (and hence profits) will be high. But as Nvidia and its competitors inevitably catch up to demand, those prices might shrink a lot.

Background

Previous coverage in January already detailed quite extensively the bearish case for the stock. Here, that will be extended and concretized with a real-world example.

Nevertheless, what has changed compared to prior coverage is a firmer acknowledgment that the continued revenue and earnings beats, with earnings expanding far beyond the revenue growth due to increased leverage, have changed the investment profile in quite a short time. This is combined with continued visibility on forward growth (even if obviously the comps will become tougher, the growth will remain strong, especially compared to other Big Tech). As well as due to the quite stagnant stock price, which had been quite flat for much of 2023, despite those aforementioned beats.

Nevertheless, admittedly, this is a bit of a late call given the strong YTD rally in 2024. From the previous coverage:

In the picture of the overall thesis, the main point here hasn’t been that the stock is overvalued, it is rather the risk that Nvidia is currently capitalizing as basically the only scaled supplier of chips powering the LLM/AI explosion.

That article’s valuation section also noted the 30x run rate P/E multiple, which given the continued expected growth (as it has been widely known for quite some that that Nvidia is a supply constraint), perhaps should have led to a more explicit bullish rating.

In that light, the previous coverage was perhaps too focused on the long-term (likely at least 3+ years from now), which might be less relevant given how quickly the financials and tech (“GAI gold rush”) are currently moving relative to any competitive dynamics. Even despite the strong YTD stock rally, recognizing this provides a bullish rating in the near-term, while remaining skeptical about the long term.

Thesis crux

As explained previously, Nvidia is using its dominant position to charge prices far above anything that could ever be possible in a healthy, competitive market. This has recently been once again demonstrated in the following tweet, which shows an example where Intel’s Gaudi2 accelerator has an IaaS (cloud) price that is literally over 10x lower.

Gaudi2: $10.42 / hour.

H100: $110.15 / hour.

As publicly available MLPerf results readily show, this is very far in excess of any performance advantage Nvidia has, with Intel in fact claiming its upcoming Gaudi3 will exceed Nvidia’s H100/200 in performance. This is the simple downside risk that Nvidia investors face. Let’s say for simplicity that the H100/200 is 2x faster than Gaudi2. That still leaves a monstrous 5x TCO (total cost of ownership, i.e., price per hour for similar performance) advantage in favor of Intel. This means that to be competitive, in principle Nvidia would have to cut its prices by on the order of 5x.

Now, if Nvidia really did that, then that would be the difference between about $100B of data center revenue in 2024, generating up to $80B in gross profits, or $20B in revenue instead and maybe $12-14B in gross profits. That, in turn, could be the difference between a $2T or $350B market cap. So, in terms of operating margin and earnings power, basically, all the leverage that Nvidia has achieved over the last year that has propelled the stock would evaporate, and then some.

Basically, Nvidia’s entire market cap is based upon its excessively high chip prices (as detailed in previous coverage, Nvidia charges per chip 2x more than what an entire TSMC wafer with ~70 chips costs, although an H100 does include HBM memory and other components).

To be sure, this example is likely a bit exaggerated, at least for 2024, as the data center business also includes its networking business and generated “only” $18.4B in the last quarter. Although its data center segment (GPU + networking) has been rising at about $4B per quarter, so at that rate it would indeed approach the $100B run rate in a few quarters.

In general, though, whatever the precise scale, the observation that competitive solutions are literally up to 10x cheaper is what matters.

For now, Nvidia maintains its scale and incumbency position advantages and has been able to satisfy the vast majority of the market’s AI hardware demand. But as competition ramps, surely customers will start weighing their options, for why would anyone spend $5B on Nvidia hardware when $1-2B of Intel or AMD hardware would yield exactly the same performance and capabilities?

For example, a company such as Meta Platforms, Inc. (META) would (in principle) certainly would see its shareholders ask them to go with the cheapest option, improving (in this example) the company’s FCF by up to $4B. With Meta having announced to acquire 300k Hopper GPUs in 2024, likely worth on the order of $10B, the example certainly isn’t exaggerated.

Earnings results

They are available on Seeking Alpha and have been widely covered.

Valuation

The stock traded for around $400-500 for much of 2023 after the initial upside outlook, with the rally towards $800 only occurring quite recently. Since the earnings have increased by far more over the last year, this shows that much of those earnings have only ‘served’ to compress the (forward) valuation multiple (growing into its valuation), which has dropped to ‘just’ over 30x currently, which is quite reasonable.

As such, with reasonable visibility over the next 1-2 years for continued growth, as the CEO said it would continue to be supply constraint during that time, the valuation warrants a bullish rating.

While it may seem a bit like conceding being wrong on the stock, upgrading to bullish only after a nearly 100% rally since last year’s Q1 results, as just described this is purely due to the valuation and increased certainty that this isn’t just a one-off spike in demand, but instead a sustained trend of increased demand. (Since the previous coverage in January had already noted the 30x run rate P/E multiple, this indeed should have been the conclusion back then already, before the ~50% rally.)

Even bulls would likely concede that they wouldn’t have expected Nvidia’s data center segment to (suddenly) grow at around a $4B per quarter clip, for several quarters, as the entire business had previously only been around the size of $4B, with quite modest growth. That certainly classifies as a black swan event.

In hindsight, this continued growth was the risk-reward investors had to pay a premium multiple for. But given the uncertainty as a black swan event, it would make sense for some to (prefer to) stay on the sidelines.

Operating leverage risk

In an extension of the valuation discussion, though, with a market cap approaching $2T, clearly the major stock gains have already been made over the last few years. In fact, a substantial portion of the ongoing rally (from $120 to $300) already occurred prior to Nvidia’s upside outlook nearly a year ago (although that was due to the stock being down from its previous high, bottoming in late 2022).

As another analyst detailed, one of the reasons for the strong earnings growth that has allowed for this performance (beyond simply growing into its valuation), has been the increase in leverage (operating/income margin).

The expansion of free cash flow margins, for me, was actually the biggest takeaway from Nvidia’s fourth quarter earnings release: while Nvidia’s revenues soared 265% year over year in FQ4’24, Nvidia’s free cash flow grew twice as fast (+546% Y/Y) which resulted in a massive improvement in the company’s free cash flow margin. The firm’s FCF margin expanded to a massive 51%, showing 22 PP growth year over year. This takeaway is significant because Nvidia has managed to become more profitable and now retains a much larger share of its revenues as free cash flow than last year.

With a nearly doubling of the FCF margin, this alone could be attributed for much of the stock performance over the last year and months. But clearly, the risk is hence that this isn’t sustainable, even disregarding the above discussion about a potentially ~5x decrease in pricing power in a bearish scenario. As such, any stock gains going forward will likely be mostly a function of continued (more linear) earnings growth from higher revenue, which will slow down on tougher comps, combined with any changes in the valuation multiple.

In other words, any forward stock gains will likely be smaller, although for which the market is now (indeed) demanding a smaller (less risky) valuation (P/E multiple). Still, admittedly even with tougher comps ahead, Nvidia will likely remain (among) the fastest-growing Big Tech for the next few years or so.

End-user application risk

With some already extrapolating growth trending towards $200B revenue, one other risk that does remain is that all this spending by Nvidia’s customers at some point would also have to start generating some returns for them (not just for Nvidia). The point is that Nvidia already had a data center business prior to GAI, which was driven by AI, so seemingly the majority of its revenue is currently driven by that one application class. However, the leader in GAI, OpenAI, likely still only has revenue in the single-digit billions, with a valuation that is lower than Nvidia’s entire annual revenue. So clearly, there is a lot of upfront spending on GAI, which may perhaps subside if it does not yield appropriate returns for Nvidia’s customers.

On the other hand, other Big Tech than OpenAI, such as Meta, Alphabet Inc. (GOOG), (GOOGL) Amazon.com, Inc. (AMZN), and Microsoft Corporation (MSFT), are clearly capable of spending this money, and will likely continue to do so give their strategic imperative to lead in this area. Still, a graph plotting Nvidia’s revenue as % of Big Tech capex showed a correlation at about 60%, which is very high (even if it might be somewhat inflated as Nvidia has other customers too).

Investor Takeaway

Near term

Despite the continued stock rally because its increase has been less than its earnings over the last year, this means the stock has become much less risky (although a bit more than before the YTD rally). This is confirmed by, indeed, the forward valuation multiple compressing to ‘just’ around 30x, although with the caveat that this prices in growth for another full year (except for any earnings beats, which are likely as Nvidia has only provided next-quarter guidance). This in turn means any stock upside (without multiple expansions and aside from any 2024 beats) does rely on further growth in 2025 (as that is the nature of a forward valuation multiple). Nevertheless, with Jensen Huang’s comment that he doesn’t fully except supply to catch up to demand even next year, this most likely largely de-risks the stock.

While the declined valuation multiple (despite continued growth visibility) makes the stock a lot more investible, without any further increase in operating leverage (which seems unlikely with the FCF margin now over 50%), unless the data center business keeps growing unboundedly (which is also doubtful, but to some extent possible), the ‘price’ of this lowered stock risk is also the likely lower returns going forward, with the market cap nearing $2T.

In other words, despite the strong growth, this isn’t a growth stock anymore, but this might actually be a reason to invest in the stock, as growth stocks are known for their possible large sell-offs, which given the de-risked valuation seems unlikely. Still, even as the YoY comps are becoming tougher, for the near future Nvidia will likely remain (one of) the fastest-growing Big Tech, so the forward returns might remain quite decent.

Overall, the skyrocketing revenue stacked with strong leverage expansion, combined with (comparatively) only modest stock growth over the last year, has in quite a short time drastically changed the stock outlook and risk profile, in the near term.

Long term

In the longer-term of many years, though, the risks previously detailed remain as valid as they were: that Nvidia’s financials are currently moving much faster than any market dynamics could, doesn’t invalidate those. Specifically, the risk is the combination of substantial pricing power loss and market share erosion as the competition ramps, both from AMD/Intel as well as from internal chip design efforts such as by Amazon, Google, and Microsoft. That 80+% data center margin simply isn’t sustainable in a competitive market.

As proof, one real-world example was shown where the Intel offering currently delivers an over 5x improvement in TCO. That is effectively the difference between spending $1B instead of $5B on AI infrastructure. As Nvidia’s earnings show, with so much money being made (at the expense of its customers’ potential earnings/FCF), there should be zero reasons not to expect its customers to investigate alternative solutions to decrease their own spending, which would in turn reduce Nvidia’s revenue, earnings and accrued leverage. In this (perhaps extreme) example, that would be the difference between earning just $20B instead of $100B in data center revenue, and that is before any decline in market share that would have led Nvidia to decrease its pricing in the first place.

In practice, for any new or existing shareholder, while as mentioned this might not be relevant for at least a few more years, recognizing this does mean one might expect (and should monitor for) an eventual top in revenue in the medium term, and hence consider taking at least some profits in due time. This would be similar to how for example declining oil prices would move the price of such stocks downwards, or even how Intel’s revenue and stock has seen downward pressure in the wake of AMD’s rise. There is no reason this couldn’t to Nvidia, and given its margin profile as described, there are likely even more reasons to expect this to happen.

As such, the skeptical/bearish long-term stance for Nvidia Corporation remains valid, although as mentioned, it might still be a few years before the revenue/stock top.

---

Analyst’s Disclosure: I/we have a beneficial long position in the shares of INTC either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

---