Artificial intelligence (AI) is expected to become big business in the long run, with Bank of America estimating that the proliferation of this technology could add $900 billion to the global economy by 2026. That figure is expected to balloon to a whopping $15.7 trillion by 2030.
The investment bank also adds that AI models are going to become way more powerful than they are right now. More specifically, BofA points out that AI models could become 1 million times more powerful as compared to OpenAI's chatbot within the next decade. But this could be possible only with the help of powerful chips with massive amounts of computational power and efficiency so that complex AI models can be trained, and inferences can be run using those models to provide responses in real time.
This explains why the demand for AI chips is expected to increase at a breathtaking pace, generating annual revenue of $309 billion in 2030 compared to just $10 billion in 2021, according to Verified Market Research. Nvidia (NVDA 6.18%) is one of the best ways to tap into this lucrative opportunity as the chipmaker currently dominates the market for AI chips.
However, Dutch semiconductor bellwether ASML (ASML 2.04%) could also turn out to be a top AI chip play, and it could even turn out to be a better AI stock than Nvidia. Let's look at the reasons why.
ASML is the reason why Nvidia is dominating AI chips
A microchip contains numerous electronic circuits with miniature electrical switches called transistors, and Nvidia has packed billions of transistors into its AI chips, which gives them immense computational power. At the same time, the die size of Nvidia's chips has been shrinking, which means that those transistors are packed more tightly. Closely packed transistors mean that electrons will have to travel a smaller distance in chips manufactured using a smaller process node, thereby making them more energy efficient.
This explains why Nvidia's A100 data center graphics processing units (GPUs), which are based on a 7 nanometer (nm) manufacturing process, are 3 to 6 times faster than the 12 nm V100 GPUs. The chipmaker's latest generation H100 Hopper GPUs are based on a custom 5 nm manufacturing node, and they are reportedly 4 times more powerful than the A100 GPUs for training AI models.
The terrific demand for these chips helped Nvidia deliver outstanding guidance when it released its fiscal 2024 first-quarter results last month. The chipmaker sees revenue jumping an impressive 64% year over year in the current quarter to $11 billion, driven primarily by the exponential growth of its data center chips that are being deployed by cloud service providers, internet companies, and enterprises to power their AI infrastructure.
However, Nvidia wouldn't have been able to shrink the size of its chips -- and make them more powerful and power-efficient -- without ASML. That's because the Dutch company holds a monopoly in the market for EUV (extreme ultraviolet) lithography machines, which enable chipmakers and foundry partners to shrink the sizes of their chips.
According to ASML, EUV lithography allows chipmakers to "pack in ever more and tinier transistors to make the chips more powerful, faster and energy efficient." More specifically, the Dutch giant's machines enable semiconductor companies and foundries to manufacture advanced chips based on 7 nm, 5 nm, and 3 nm nodes.
Not surprisingly, there is a massive demand for ASML's EUV machines given the race to shrink chip sizes for applications such as AI. The company had a sizable backlog worth 39 billion euros at the end of the first quarter of 2023, which substantially exceeds its full-year revenue forecast of 26 billion euros, a 25% jump over its 2022 revenue.
More importantly, Wall Street analysts estimate ASML's revenue growth to accelerate impressively in a couple of years.
ASML Revenue Estimates for Current Fiscal Year data by YCharts.
Furthermore, ASML's top line growth is expected to drive healthy earnings growth as well.
ASML EPS Estimates for Current Fiscal Year data by YCharts.
ASML should be able to sustain such impressive levels of growth for a much longer time given the growing demand for AI chips, which will create the need for more of its EUV machines. As such, analysts are expecting the company to clock 30% annual earnings growth over the next five years. Interestingly, ASML's bottom line is expected to grow at a faster pace than that of Nvidia over the next five years, as the latter's earnings are forecast to increase at an annual rate of just under 20%.
And looking at the company's valuation, it's clear that it is a top AI stock to buy right now.
The valuation makes buying this stock a no-brainer
ASML stock is trading at 39 times trailing earnings and 11 times sales. Those multiples make it way cheaper than Nvidia, which is currently trading at 205 times trailing earnings and has a price-to-sales ratio of 38.
The discussion above makes it clear that ASML could be one of the biggest beneficiaries of the AI boom given the role its machines play in helping the likes of Nvidia make advanced chips. That's why investors looking to take advantage of this hot technology trend may want to buy ASML before it jumps higher following its 33% gains this year.
