Nvidia (NVDA 6.18%) stock is seeing a massive artificial intelligence (AI)-fueled surge in 2023, rising 212% so far thanks to the critical role the company's graphics cards and processors play in training AI models and running inferencing applications.
In fact, Nvidia's role in the proliferation of this technology is so profound that Morgan Stanley recently singled out the company as a proxy for AI stocks. That's not surprising, as it wouldn't have been possible to train popular large language AI models such as ChatGPT from Microsoft (MSFT 1.82%)-backed OpenAI without thousands of data center GPUs (graphics processing units) from Nvidia.
However, the AI-driven surge in Nvidia stock has made it so expensive that investors who haven't bought it yet may be scratching their heads and wondering if it is a good investment anymore. And they cannot be blamed, as Nvidia's price-to-earnings ratio stands at a whopping 229, whereas its five-year-average earnings multiple is 73. The sales multiple of 42 further indicates how richly valued the stock is.
While these multiples seem rich enough to scare new investors away from Nvidia stock, do those investors have any alternative if they want to take advantage of the AI boom? In other words, is there a cheaper and better AI stock to buy right now?
Nvidia is expected to clock insane growth thanks to AI
UBS analyst Timothy Arcuri recently labeled Nvidia a "kingmaker" with respect to AI technology. He pointed out that companies are pouring a lot of capital into AI software and are looking to scale up their infrastructure, but they are constrained by capacity. One of the reasons behind this capacity constraint is that Nvidia hasn't been able to produce enough of its data center GPUs that are deployed for training AI models.
Andrej Karpathy, who is currently at OpenAI and was formerly the director of AI at Tesla, recently pointed out that AI companies reportedly need 432,000 of Nvidia's H100 data center GPUs to train their large language models (LLMs). It is worth noting that this particular chip costs $35,000 a unit, which means Nvidia already has a solid revenue pipeline worth $15 billion. This explains why the company has placed orders for more semiconductor wafers so that it can manufacture more of its data center GPUs.
More importantly, the AI gold rush that Nvidia benefits from should last for a long time. According to UBS, the company controlled 6% of the $250 billion data center chip market in 2022. That share is expected to rise to 25% this year. Meanwhile, Mizuho Securities analyst Vijay Rakesh estimates that Nvidia could end up controlling 75% of the market for AI server chips by 2027.
Rakesh adds that Nvidia could generate $25 billion to $30 billion in revenue this year through sales of its AI chips. By 2027, Rakesh said that figure is expected to jump a whopping tenfold to $300 billion. There are two reasons why that could indeed be the case.
First, shipments of AI servers should increase at an annual rate of 22% through 2026, according to market research firm TrendForce, creating the need for more AI chips. It is worth noting that GPUs are the dominant AI chip type, suggesting that Nvidia's addressable market is likely to keep growing, which should help improve the company's volume shipments.
Second, Trendforce said the average selling price (ASP) of Nvidia's data center GPUs is expected to jump further. Its H100 GPUs are 2 to 2.5 times pricier than their predecessors, which were used for training ChatGPT. It is now anticipated that the successor to the H100 could be 40% pricier, indicating that each chip could command almost $50,000. Nvidia, therefore, could enjoy even stronger ASPs in the long run, which could drive eye-popping growth for the company.
NVDA Revenue Estimates for Current Fiscal Year data by YCharts
Moreover, all of this suggests that Nvidia could sustain such outstanding growth as the demand for AI chips increases and it continues to maintain its dominance in this market.
Are there any alternatives?
Now that we have seen how rapidly AI is expected to power Nvidia's growth, it is easy to see why Morgan Stanley treats the chipmaker as a proxy for this space. After all, Nvidia's dominant presence in the AI hardware market is unmatched.
So we have to turn our attention to companies dabbling in AI software to see if there's a better AI stock than Nvidia. Microsoft is one possibility, as the company enjoyed an early-mover advantage in AI software thanks to its investment in ChatGPT developer OpenAI. The company has quickly used OpenAI's algorithms to enhance its other offerings. Microsoft's Azure Cloud, for instance, has been adding customers rapidly since the company gave enterprises access to OpenAI on the cloud for training LLMs.
However, Microsoft's AI-driven growth isn't going to be as solid as that of Nvidia. In fact, AI could take longer to drive material growth for Microsoft, with the company's executives reportedly expecting its annual AI revenue to hit $10 billion. That would be lower than the estimated AI-specific revenue that Nvidia is expected to generate this year. Moreover, Microsoft is expected to generate $235 billion in annual revenue in the current fiscal year, so AI isn't going to move the needle in a significant way for the company right now.
Alphabet (GOOG 9.96%) (GOOGL 10.22%), another tech giant trying to capitalize on AI, has been pulling the right strings on the AI software side. However, the technology is likely to take time to drive significant growth at Alphabet given its huge reliance on the digital ad and search engine market. This explains why Alphabet's growth isn't going to be as solid as that of Nvidia.
Meanwhile, traditional Nvidia competitors such as Advanced Micro Devices and Intel are still trying to cut their teeth in the AI chip market. While AMD's AI accelerators are still being tested by customers and will hit the market later this year, Intel has a revenue pipeline of $1 billion for its AI chips. All this tells us why Nvidia is a top AI stock to buy despite its rich valuation, as it is well ahead of its peers looking to take advantage of this niche.
The good part is that Nvidia's forward earnings multiple of 55 points toward a massive jump in the company's earnings. Nvidia should be able to deliver a big spike in earnings thanks to strong volume shipments of AI chips as well as its terrific pricing power, which is why growth investors can consider buying it even after its outstanding gains in 2023.
