Artificial intelligence (AI) is widely considered to become the biggest paradigm shift of the next decade. Not surprisingly, it also emerged as a major investment theme.
While many companies with AI exposure have seen a surge in share prices, none of them grabbed as much public attention as Nvidia (NVDA 3.37%) has. With its robust hardware-software ecosystem, Nvidia already emerged as a dominant force in the still-nascent AI (and generative AI) market. Shares are already up by 190% so far this year, placing the company firmly in the $1 trillion market-capitalization category.
At least two compelling reasons remain to view Nvidia as a long-term buy even at the currently elevated share prices. But there are also two solid reasons you might want to sell the stock in July 2023. Let's assess.
A GPU upgrade cycle in gaming
Although no longer Nvidia's largest revenue source, its gaming business is still sizable and accounted for almost 31% of the company's revenue in the first quarter of fiscal 2024 (ending April 30, 2023).
Nvidia's revenue and earnings took a hit in fiscal 2023, as weak PC demand resulted in a significant inventory buildup of graphics cards. But this headwind might soon resolve.
According to Jon Peddie Research, demand for PC gaming hardware could recover by 2025, despite recessionary and inflationary pressures.
The company is positioned to benefit from this demand recovery. Its recently launched 40-series RTX graphics cards (code-named Ada Lovelace, an improvement over the RTX-30 series) have been witnessing solid demand. Around 29% of its installed base had shifted to RTX series graphics cards, while the remaining were using the older generation GTX series card (as of March 2022).
Since the new RTX series cards are superior to GTX cards in cost efficiency, computing power, and the more-immersive and realistic gaming experience (due to ray tracing and AI-enabled graphics), the subsequent chip upgrade cycle will be a major tailwind for the company.
Nvidia estimates its addressable market in gaming to be $100 billion. With the annual run rate of its gaming business close to just $9 billion, there is still significant scope for future growth.
The data center opportunity
The data center segment accounted for 56% of the company's total revenue in fiscal 2023. The launch of ChatGPT has made generative artificial intelligence (AI) mainstream, and data centers have been racing to adopt this technology.
But since most of the $1 trillion worth of data center infrastructure is based on central processing units (CPUs) and dumb network interface cards (NICs), many data centers will have to opt for accelerated computing hardware and software to run generative AI applications.
With its full-stack solution (specialized chips such as A100 and H100 GPUs and Grace CPUs, its specialized CUDA software package to optimally program the chips, and domain experts for accelerated computing), the company is in a good position to benefit from this trend.
Nvidia's comprehensive hardware-software ecosystem continues to power AI applications, with the most notable being generative AI applications that run on large language models (LLMs). The company's offerings have been used extensively to train LLMs on huge data sets.
Nvidia is expected to see a surge in demand for its offerings from enterprises focusing on training LLMs with their internal data. And inference (LLMs providing analysis and insights for new data sets based on previous training) is also expected to be a major tailwind for the company's AI-optimized solutions.
Why investors are concerned about this stock
Despite Nvidia being widely regarded as a top-notch AI pick, investors should also be aware of the two major risks.
Nvidia is currently at a price-to-earnings (P/E) ratio of 220, dramatically higher than the median semiconductor industry multiple of 24.2. The sky-high valuation assumes that the company will continue to maintain its dominance in AI and accelerated computing for several years, no matter the macro environment, supply chain challenges, and competition.
The valuation also assumes near-perfect execution by Nvidia's management and includes no margin of safety. If the company misses its earnings estimates, the stock may see a dramatic correction.
The company is also exposed to significant geopolitical risk. The U.S. government seems to be considering extra restrictions on the export of AI chips to China. Nvidia managed to comply with the previous restriction of limiting the computing power of chips by selling lower-spec data center A800 and H800 chips in the Chinese market.
But the new export controls would further limit the sale of these chips. With China accounting for 20% to 25% of Nvidia's data center revenue and over 20% of the company's total revenue, these additional controls may hurt profitability.
While investors should definitely not ignore these challenges, Nvidia has built a strong moat around its high-performance AI chips with its CUDA software package, 4-million-member developer base, and partnerships with all major cloud providers. So even if the competition develops similar chips, they will find it difficult to replicate the company's full-stack approach to solving AI workloads.
According to a Bloomberg Intelligence report, the generative AI market is expected to be worth $1.3 trillion by 2032. Given the scope of the market opportunity and Nvidia's well-established technological superiority in processing generative AI workloads, it makes sense for retail investors to pick up a small position in this stock even at elevated price levels.
