Is Nvidia a good investment for the next 10 years?

Why it might be time to take profits from this tenbagger

Due to the omnipresent AI hype, the chip provider Nvidia has become a mainstream stock. Even among non-techies, word has spread that the company is benefiting greatly from the triumph of artificial intelligence. Many retail investors now have Nvidia stock in their portfolios just as naturally as Apple or Microsoft.

And the success proves the happy Nvidia shareholders right: driven by outstanding results in H1 2023, the Nvidia stock has already gained over 200% this year. Nvidia shareholders who have been with the company for more than 5 years are even happy about a tenbagger. Congratulations!

Nvidia share price

Overall, Nvidia is currently valued at an enterprise value of over $1,100 billion, is one of the 10 most valuable companies in the world and thus weighs almost as much as Amazon. And this despite the fact that Nvidia generates just 6% of the e-commerce giant's sales!

In order to better understand how this share price rallye was achieved, it is perhaps good to remember that such erratic movements in shares are not unusual in a hype around new technologies, but actually occur regularly.

What is a Tech Hype?

In the context of new technologies, the term "hype" refers to intense attention and enthusiasm around a particular product or enabling technology. This attention and enthusiasm often exceeds what is justified by the actual features or value of the technology.

Hypes can be triggered by a variety of factors, including media coverage, word of mouth, marketing campaigns, and social media. It is important to note that tech hype is not necessarily a negative thing. Hype can help raise awareness of a new idea or product and can often lead to more people trying it out or looking into it.

However, hype also regularly leads to inflated expectations, which are then disappointed when the product, idea or event does not live up to the hyped (i.e., inflated) expectations.

The Hype Lifecycle

The hype lifecycle is a concept developed by IT consulting firm Gartner. It describes the adoption and impact of technologies and innovations over time.

The hype cycle consists of five phases:

The Hype Lifecyvle

1. Technology Trigger: A new technology or innovation emerges and generates significant media interest, even though there are few usable products or proven applications yet. The trigger of the current AI hype was the unveiling of ChatGPT in November 2022. Since then, there has been almost daily news in the field of Generative AI (GenAI), a veritable gold rush has broken out in the industry.

2. Peak of Inflated Expectations: Expectations for the great mass of such AI tools are huge, and the advertising for the products just entering the market and the rising stock prices of "AI stocks" continue to fuel them. Although serious AI experts warn against exaggerated expectations, the masses are jumping on the bandwagon. There are amazing first successes e.g. in AI-based generation of texts and images, but also many failures like the hallucinations of ChatGPT and Co. which will eventually lead to skepticism. This phase may be ahead for many GenAI applications starting in 2024.

3. Trough of Disillusionment: The technology is not adopted as quickly as expected and is often seen as a failure. Even in the generative AI space, many will ultimately be disappointed that while AI can provide great assistance and bring efficiency gains, it ultimately can't perform miracles either. Most manufacturers of AI products will give up in the merciless competition, while others will move on and improve their products or services. Countless software vendors currently have GPT-based AI chatbots in development, for example. The added value of these tools for the user is far from universally proven. Technologies such as the NFTs and the Metaverse, which were being hyped over the last years, are currently going through this valley of disappointment.

4. Slope of Enlightenment: The Technology is becoming better understood and implemented in this phase. Companies are learning how to leverage the technology for their business. I see quite a few cyber security technologies and their stocks in this phase right now. The hype is over, the stocks have crashed in 2022, but the security technologies are mission critical and should have a great future ahead of them. In general, this is a good period to invest in formerly hyped technologies.

5. Plateau of Productivity: The Technology is widely accepted and used. The market is becoming more stable and its size is now quite assessable. Leading players have emerged. I would see the current world of social media in this phase. At least before Elon Musk caused a shake up of this industry. ;-)

The AI hype around Nvidia

Nvidia will actually benefit more than almost any other company from the quantum leap in AI. The leading AI chip maker is certainly the best-known "shovel seller" in this gold rush. Yet it is by no means certain how much AI gold there actually is to be discovered, let alone how many shovels Nvidia, of all companies, will be able to sell.

But seriously, Nvidia is known for its GeForce, Quadro, and Tesla GPUs, as well as the Tensor cores built into its newer GPUs that are optimized specifically for AI and machine learning. Nvidia has also developed an extensive software library for AI and machine learning called CUDA, which is widely used and an important advantage for Nvidia.

How big is the market for AI chips?

There are different forecasts for the size of the AI chip market in the coming years. According to Statista the AI chip market is forecast to grow from $20 billion in 2022 to $165 billion in 2030, with a compound annual growth rate (CAGR) of about 30%. Other market researchers predict that the AI chip market will reach about $228 billion U.S. dollars by 2032.

In other words, it's a big market, indeed a very big market, but no one can really seriously predict the size of the longer-term TAM (Total Addressable Market) for Nvidia. Neither can Nvidia's market share.

Looking at longer-term analyst estimates for Nvidia, the company is reportedly expected to reach $147 billion in revenue by 2028.

So currently, the majority of analysts believe that Nvidia will be able to solidify its monopoly-like position in the booming AI chip market for years to come.

Currently, Nvidia earns dream margins on its GPU chips. Nvidia sells its H100 AI chip for $25,000 to $30,000, while Raymond James analysts estimated manufacturing costs at just over $3,000.

As a result, Nvidia's net margins are likely to literally explode to over 40% in the current fiscal year. After earnings per share of $1.74 in the last fiscal year, profits are likely to multiply to about $10 per share in the current year.

But that's not all: with their series of numbers, the analysts assume that the abnormally increased profitability can be increased even further in the coming years and believe in net margins of about 50% and that for many years.

I do not believe in this best-case scenario. If only because I can't recall any comparable situation in the tech markets where a hardware manufacturer has been able to achieve such margins in an attractive mass market over an extended period of time.

Even if Nvidia fans see it differently: There is no comprehensible reason for me why competitors should not emerge in the coming years that will take certain market shares from Nvidia or at least nibble at the margins.

The competitive situation in AI chips

Here is a quite lengthy overview of Nvidia's "market companions", which are currently often underestimated:

AMD: Advanced Micro Devices (AMD) is known for its graphics processing units (GPUs), which are used in many AI applications because they are capable of performing many parallel calculations, which is particularly useful for these types of tasks. This makes AMD a direct competitor to Nvidia.

In recent years, AMD has increased its presence in the GPU market, competing with Nvidia in certain segments, especially gaming GPUs. AMD, like Nvidia, has its own software and hardware infrastructure for AI and machine learning, including the ROCm (Radeon Open Compute) platform and the MI series of GPUs designed specifically for AI applications.

Intel: Intel offers several types of hardware for AI, including CPUs, FPGAs (Field Programmable Gate Arrays), and specialized AI chips such as the Nervana Neural Network Processors. Many investors have already written off Intel due to its disastrous performance in recent years. I think a comeback is quite possible, even if there is little concrete evidence for it at the moment.

IBM: IBM has developed the Power9 chip, which is optimized for AI workloads. IBM, like Google, has also developed a fundamentally new quantum computing technology that could play an important role in AI systems in the future.

Google: Google has developed its own AI chips called Tensor Processing Units (TPUs) that are specifically optimized for machine learning and deep learning.

OpenAI: The ChatGPT manufacturer, funded by Microsoft among others, is also considering developing its own AI chips in the future, according to a Reuters report.

Microsoft: According to unconfirmed media reports, Microsoft will soon present its own AI chips, which are to compete directly with Nvidia in the training of large language models. Again, the goal is obviously to reduce the current dependence on Nvidia and the directly related GPU costs.

Apple and Alibaba have also developed their own AI chips to meet their specific machine learning and cloud infrastructure needs. And then, of course, there are new young competitors like Graphcore, which is an emerging British chip company that has developed the Intelligence Processing Unit (IPU) chip, specifically designed for AI applications.

Even though the detailed competitive situation is unclear, it is pretty obvious that the competition for AI chips will increase in the coming years. It is unlikely that Nvidia will remain a quasi-monopolist in this market in the long run.

The chip industry is and will remain extremely cyclical even in the AI era. And it will also become price-sensitive again in the medium term. For this reason, I would warn against extending the steep growth and, above all, the fantastic margins that Nvidia will achieve this fiscal year and probably also in the next fiscal year, too far into the future.

The valuation of the Nvidia share

Nvidia stock is overvalued for my taste at 34x trailing 12-month sales or >100x trailing 12-month cash flow.

Expectations for the current fiscal year FY24 are spectacular at nearly $55 billion in revenue, or over 100% year-over-year growth (see analyst forecasts here).

But the market has long priced in this stellar performance. Even if Nvidia revenue doubled again in the next two years (which is roughly in line with optimistic analyst estimates), the EV/sales ratio would still be in double digits by 2026 if the stock price stays the same.

That's way too much - especially in the highly cyclical semiconductor industry. Nvidia’s sales are non-recurring and deserve a correspondingly lower valuation compared to SaaS vendors with their business model based on regular subscriptions. Nvidia shares are currently much too expensive.

I think you can compare Nvidia's situation today quite well to Cisco in the late 90s. Cisco was the leading "shovel seller" in the Internet hype at the time and seemed untouchable with its superior networking technology. The stock found its way into the portfolios of countless U.S. retail investors at the time and soared to breathtaking highs in 2000. Then the dot-com bubble burst. Cisco's stock lost more than 90% of its value and still has not returned to its 2000 price, even though the company has dominated its industry ever since.

In my opinion, another Nvidia stock crash - similar to 2022 - is only a matter of time. I do not expect a good return on Nvidia stock over the next 10 years.

Nevertheless, Nvidia is an outstanding company and I caution against speculating on falling prices. After all, no one can predict when this hype will end and the stock will come back down to earth.