Artificial Intelligence (AI) is all the rage right now. OpenAI’s ChatGPT went viral late in 2022, attracting big investment from perennial software giant Microsoft and prompting internet search leader Google parent Alphabet to respond with its own AI services. An artificial intelligence arms race has begun.
This is great news for computing hardware stocks, many of which bottomed out in October 2022 and have been off to the races ever since. A new bull market is coming, and semiconductors, computing equipment, and AI are laying the groundwork for it. That’s why three Fool.com contributors think Advanced Micro Devices (AMD -2.08%), Aspen Technology (AZPN -0.65%), and Super Micro Computer (SMCI 3.47%) are still a buy for the long haul.
The new leader in data center processors?
Nicholas Rossolillo (AMD): Former underdog chip designer AMD just had an “epic” 2022. Led by its Epyc processors, the company’s data center segment nearly doubled in size from $3.69 billion in revenue in 2021 to $6.04 billion in 2022. With new chips like the Instinct MI300 (an “accelerated computing” processor that combines a CPU and a GPU) coming out later in 2023 to address advanced computing and AI needs, the future looks bright for AMD’s top moneymaking segment. Data center surpassed the client PC and laptop segment in both sales and operating profit earlier in the year, and the gaming segment in the fourth quarter.
As incredible a run as AMD has had in this department, there’s still a massive opportunity for it to expand even further. At the moment, incumbent Intel is still the leader in data center processors — at least in terms of size. CEO Pat Gelsinger alluded on the earnings call that cloud operators (which use data centers to run their applications and where most AI heavy lifting is done these days) still have 95%-plus Intel chips installed. In terms of dollars and cents, that meant $19.2 billion in data center and AI revenue for Intel in 2022.
Intel believes that its large installed base gives it lots of opportunity to earn more revenue in the coming years when data centers eventually need to replace older chips — and I believe that’s true. However, that also means Intel has the most to lose in this battle. To wit, Intel’s data center segment declined 15% in 2022. AMD is clearly in the technological lead right now, so it is the better growth option. With advanced AI workloads a high priority, scrappy AMD could further “chip” away at Intel’s large data center business in the coming years.
AMD stock has rallied some 50% off of its mid-October 2022 lows, so the path forward will be much more bumpy. Though data center chips are still expanding, the company is still struggling in its client segment as the industry burns off excess inventory of PCs and laptops. Nevertheless, the new AMD aimed more squarely at big enterprise computing is still highly profitable on a free-cash-flow basis, and its profit margin could be in store for a rally in the second half of 2023. Shares trade for 28 times expected 2023 earnings per share. For a company with both revenue and profit growth potential ahead of it, I think it’s a fair value for investors who want to own AMD — the new leader in AI processors — for the long haul.
Here’s the mastermind behind AI of Things
That’s one of the keystones in Aspen Technology’s business model. The company makes software and cloud-based services that help businesses make more efficient use of their industrial machinery.
AIoT plays a central role in the fully integrated aspenONE platform, covering all elements of production from engineering and design to asset management and materials. The idea is to measure everything with sensors in the real world, analyze the data through automated software, and make decisions based on the outcome. That’s the essence of AIoT.
This software suite, recently renamed AspenTech DataWorks, helps companies ingest and process a wide range of structured and unstructured data, ready to feed into Aspen’s full range of data analysis tools. Thanks to the magic touch of AI systems, DataWorks can turn time-consuming manual data entry into highly automated workflows. Processes that used to require dedicated data management departments can be set up in minutes and then run hands-free for years.
It’s no surprise that Aspen’s AI-powered solutions are popular with world-class enterprises. The company counts some of the planet’s largest oil, pharmaceutical, consumer goods, and food businesses among its customers. The long-term target is a self-optimizing industrial plant where every process is measured, analyzed, and adjusted in real time — all by Aspen’s automated AI tools.
That industrial utopia is still a long way away, but that only leaves more room for business growth over time. Meanwhile, your average analyst expects Aspen to deliver 11% top-line growth in 2023 and a five-year compound average growth rate of 17.5% on the bottom line. The stock has certainly been kind to its owners recently, doubling the S&P 500 index’s total returns over the last three years.
As Aspen’s AI strategy finds traction in the ever-changing marketplace, I expect more market-beating stock returns from the best AI company you’ve probably never heard of.
This AI stock just tripled its earnings yet trades at a single-digit P/E multiple
Billy Duberstein (Super Micro Computer): It’s difficult to know who will win the AI wars, but one thing that is certain is that artificial intelligence takes a ton of computing power to run. That means data centers that are filled with servers consisting of the latest, most sophisticated chips.
But one big complication is that these chips take enormous amounts of electricity to run. And they also generate lots of heat that needs to be cooled — adding yet another layer of cost.
Super Micro Computer is a server assembler that may not sound like a great business, but actually has everything one could ask for to benefit from this wave of innovation. Super Micro’s model is a bit different than other server assemblers in a couple of important ways.
First, its servers are modular, built upon its “building block” system, in which server components are somewhat interchangeable across different configurations. This allows for rapid mass customization, which AI data center operators may find attractive for application-specific systems.
Second, these building block servers allow certain components to be swapped out or refreshed, instead of customers having to replace an entire server system, saving even more on costs.
Third, Super Micro is based in Silicon Valley, and works closely with all major CPU and GPU vendors. When combined with its modular designs, those close relationships allow Super Micro to achieve very fast time to market.
And finally, Super Micro boasts the most energy-efficient servers on the market, due to its designs that efficiently distribute computing resources, and contain the company’s proprietary liquid cooling systems.
As electricity costs and other inflationary forces spiked last year, the focus on total costs of ownership brought more and more customers to Super Micro. As a result, revenue and earnings accelerated; on its recent earnings call, management noted accelerated computing and GPU platforms for AI systems were growing very fast, and now make up over 20% of revenue.
Even as the broader market fell, Super Micro’s stock did surge over the past year. But its rise wasn’t due to hype about future profits that may or may not happen, as is the case with other AI stocks. Rather, Super Micro’s P/E ratio fell, as its 200% earnings-per-share growth was actually larger than the stock’s 124% gain.
The market clearly expects Super Micro, trading at just 8.2 times earnings, to give some of those earnings back, perhaps due to cyclical factors.
Yet while management guided to a sequential decline in the current quarter, there is usually negative seasonality in Super Micro’s March quarter. For the full fiscal year ending in June, Super Micro still expects 30% year-over-year growth. Encouragingly, founder, chairman, and CEO Charles Liang also anticipates the company can grow at least 20% in fiscal 2024, as it continues to take market share due to the increasing needs for energy-efficient high-performance computing.
It’s not often you’ll find an AI-related stock with that type of growth that trades at a single-digit P/E ratio. Even after its massive rise, investors shouldn’t hesitate to buy Super Micro shares at these prices.