Artificial intelligence (AI) used to be a hot, but mostly empty, buzzword. These days, it is a reality that can reshape industries far removed from Silicon Valley and alter our everyday way of life.
Three Motley Fool contributors identified a handful of AI stocks with promising prospects for the long term. Read on and consider picking up a few shares of Intel (INTC -9.20%), Air Products and Chemicals (APD 0.43%), and Amazon (AMZN 3.43%), which offer a diverse array of AI investment opportunities. Each offers substantial long-term growth potential from a unique angle.
Intel's secret AI sauce: A tasty blend of software and value
Anders Bylund (Intel): Intel makes and designs plenty of chips aimed at the AI market. Starting with beefy general-purpose ships like the Xeon Scalable line, into the number-crunching prowess of Intel's graphics processors, and down to the Movidius series of custom-tailored neural network accelerators, Chipzilla wants to have a hand in every corner of the AI sector.
Critics often argue that high-performance specialists such as Advanced Micro Devices and Nvidia have the upper hand against Intel. Ultra-specialized solutions are right up their alley and both stocks are trading at nosebleed-inducing valuation ratios.
Nvidia and AMD are great companies in their own rights, and I'm not throwing them under the bus here. Despite their lofty stock prices, both may have the potential to rise even higher as the AI boom plays out.
But Intel has a couple of aces up its digital sleeve.
First, Intel insiders see the company's AI software as a significant competitive advantage. The semiconductor titan is putting its back into software solutions, with generous development budgets and a keen understanding of how superior software can quench even the strongest hardware alternatives.
Intel's VP of Data Center and AI Operations, Sandra Rivera, spoke at a technology conference last week and the software advantage came up in that conversation:
Particularly when you get into large scale-out clusters, it really isn't just the device. Process technology always gets you a performance boost, anywhere from 15%, 20%, 25%, 30%. Architecture and design gets you another 20%, 25%, 30% performance boost. But software is the multiplier. You can get 5x, 10x, 20x performance boost through software.
So hardware design and architecture are important, but top-notch AI software can make a real difference. Rivera noted that software development is "the biggest area of investment" for Intel at this point, and I don't hear that from the silicon-focused competition.
Second, many investors have walked away from Intel in recent years. The AI explosion is helping, but we're not talking about a full recovery by any means. AMD's stock price has doubled year to date while Nvidia's stock nearly tripled. Intel's chart has only climbed 30% higher in 2023. And when you look at traditional value metrics such as price-to-sales or forward price-to-earnings ratios, Intel shares are a bargain in this potentially thrilling growth sector.
Thanks to these two advantages, I'm more interested in buying Intel stock than AMD's or Nvidia's high-priced alternatives right now. Chipzilla provides a better risk-to-reward balance in my eyes.
This 20th-century infrastructure provider is still very relevant today
Nicholas Rossolillo (Air Products and Chemicals): Why is an industrial gasses and energy equipment company on a tech stock list? Well, a case could be made that Air Products and Chemicals is a player in the tech world, owing to its hydrogen business and future hydrogen supply development plans. Hydrogen is a new and promising renewable fuel that is attracting lots of investor attention.
But there's more tech to this company than initially meets the eye. Air Products is a supplier to the semiconductor and computing device assembly industries. Industrial gasses are a key ingredient in the complex process of manufacturing chips. Be it rare gas mixtures or more common gas concentrations, demand for these products will be on the rise as the U.S., Europe, and East Asia race to ramp up semiconductor manufacturing capacity in the next five to 10 years.
Besides supplying these chip fabricators, there's another tailwind working in Air Products' favor: skyrocketing energy usage in data centers. More than a decade of cloud development has pushed more high-performance computing work into giant data centers. But the chips in each of these facilities use up massive amounts of electricity. Even some leaders in the semiconductor industry are worried that new generative AI work will ramp up energy consumption in data centers beyond what can easily be supplied. According to estimates, U.S.-based data centers consume a mid-single-digit percentage of total electricity today, but that could increase to a low-teens percentage in a few short years.
Besides more efficient chips, new energy sources (again, like hydrogen) and more efficient industrial equipment could provide part of the solution for data center and AI energy consumption. Air Products could thus be an ancillary supplier to the tech universe via this route.
Shares of the company trade for 29 times trailing 12-month earnings, a premium for a giant, slow-ish-moving industrialist. But Air Products has earned it with many years of steady low-teens-percentage earnings growth as demand for its gasses, energy services, and related equipment steadily tick higher. And with tech manufacturing due for a jump higher in developed markets like the U.S. and Europe for the rest of the 2020s, Air Products' run could be far from over. I recently started buying using a dollar-cost average plan.
Investors are counting Amazon out in the AI races -- that's a mistake
Billy Duberstein (Amazon): Some may be surprised to know that, despite its lead in cloud computing, Amazon's stock is actually farther below its all-time highs than all its major cloud infrastructure competitors.
AMZN Percent Off All-Time High data by YCharts
To be sure, Amazon has not been as high-profile with its artificial intelligence efforts as Microsoft, with its splashy OpenAI investment, or Alphabet, which acquired DeepMind back in 2014.
But investors are foolish to think Amazon will be left behind. Amazon still has leading capabilities, and while Microsoft has an exclusive relationship with OpenAI, Amazon offers a broad range of large language models (LLMs) from other AI start-ups such as Laughing Face, Stability.ai, Cohere, and AI21 Labs. Not to mention, Amazon has its own machine-learning models developed in-house as well. It's still early in the AI races, and it's not clear that OpenAI's ChatGPT LLM will be the sole winner over others for generalized AI.
In addition, Amazon makes proprietary chips -- a result of its 2015 acquisition of Annapurna Labs. So besides offering the widest range of GPU and CPU chipsets from the industry leaders, Amazon also has its own proprietary AI accelerator chips, Inferentia and Trainium. Proprietary chips that can potentially lower costs for customers or do other unique things is a nice addition, and therefore, could also be an advantage for Amazon's cloud offerings going forward.
Finally, many think AI will be most beneficial to companies with proprietary data, on which it can train models no one else can. This should of course benefit Amazon's e-commerce business, which dominates that industry. After all, who has more data on the purchasing patterns of consumers across the country, or online advertising effectiveness, or streaming video preferences, than Amazon? Meanwhile, Amazon will no doubt be able to use AI across its giant logistics delivery platform to great effect, potentially lowering costs and boosting efficiency.
The bottom line is, Amazon is extremely well positioned to benefit from AI, not only as a supplier of AI chips, computing power, and models, but also as a user of AI across its various e-commerce, advertising, and logistics businesses.
