What is a data center? It's a dedicated building filled mostly with clusters of computer servers working in unison to complete tasks, and it's typically managed by tech giants like Microsoft and Amazon (AMZN 3.43%). Millions of businesses rent those data centers to store their data, run their online sales channels, and even develop software. This is called cloud computing.
Those businesses now demand more computing power than ever so they can develop, train, and deploy artificial intelligence (AI) applications using those centralized data centers. That's driving a surge in demand for advanced chips capable of handling AI workloads, and it could be a multiyear opportunity for semiconductor companies that supply the chips that power those servers.
With that in mind, the following four chip stocks could benefit significantly from the AI boom, and you might want to buy them now.
Image source: Getty Images.
1. Advanced Micro Devices
Advanced Micro Devices (AMD 2.37%) is one of the world's largest chip companies. Its hardware can be found in a number of popular consumer products like the Sony PlayStation 5, Microsoft's Xbox Series X, and even the infotainment system inside Tesla's electric vehicles. Not to mention, AMD's chips are a go-to choice for computer and gaming enthusiasts.
But the company also has a strong presence in the data center segment, and it will begin shipping its newest MI300 chips later this year that are specifically designed for AI applications. The MI300A variant will combine processor chips (CPUs) and graphics chips (GPUs) to create what AMD says is the world's first accelerated processing unit (APU) for data centers.
That particular chip was chosen by the Lawrence Livermore National Laboratory to run its upcoming El Capitan supercomputer, which will be the most powerful in existence when it comes online in 2024 with two exaflops of performance (it can calculate 2 quintillion floating point operations per second).
Since a significant portion of AMD's revenue comes from consumer-related segments like gaming and personal computing, the company's growth has slowed over the last 18 months amid broader economic weakness. As a result, its stock trades 33% below its all-time high. However, with a new range of data center AI chips almost ready to ship, now could be a great time to invest in AMD ahead of what could be a major growth cycle.
2. Amazon
Amazon is the world's largest e-commerce company, and its Amazon Web Services (AWS) business is also the world's leading cloud computing platform. It normally buys chips from top providers like AMD and Nvidia (NVDA 6.18%) to power its data centers, but it knows AI could drive a major demand cycle and it wants a slice of the action.
Amazon has been working on its own chips since 2018 when it launched Inferentia, which is designed to help complex machine-learning models make faster predictions. It was followed by the Trainium chip in late 2020 which, as the name implies, is aimed at training advanced AI models. Amazon's e-commerce empire was built on cheap prices, and it's taking the same approach to chips -- the company says its Trainium hardware could offer savings of 50% compared to other GPU infrastructure.
But amid the demand wave for AI chips this year, Amazon is working on building its reputation as a true competitor in the hardware space. It just made a blockbuster $4 billion investment in one of the AI industry's leading start-ups, Anthropic, and that company will be using Amazon's chips to train its future models. That could attract troves of other AI developers, especially if Amazon helps them save money.
AWS has suffered from slowing growth recently. In the second quarter of 2023 (ended June 30), the cloud platform grew its revenue by just 12% year over year -- far slower than Microsoft Azure (26%) and Alphabet's Google Cloud (28%). Making headway in new AI businesses like hardware will be critical to bolstering the growth of AWS going forward, and investors have an opportunity to buy Amazon stock now ahead of that push.
3. Axcelis Technologies
Axcelis Technologies (ACLS 3.06%) is an out-of-the-box addition to this list because it doesn't actually produce any chips itself. Instead, the company provides ion implantation equipment to chipmakers that is critical to the fabrication process.
In the recent second quarter of 2023 (ended June 30), Axcelis told investors it was beginning to see an increase in demand from customers producing AI chips in the memory (DRAM) space. Plus, it's experiencing strong growth from customers making silicon carbide chips used in electric vehicles (EVs). Many EV manufacturers are also developing autonomous driving technologies, which means those cars could eventually become platforms for powerful software requiring more computing power than ever before.
While some chip companies have experienced a slowdown lately, Axcelis has grown its revenue by 24.2% in the first half of 2023. Plus, the company has amassed an order backlog for its equipment worth $1.2 billion, which is equivalent to more than one full year's worth of revenue.
Axcelis stock has doubled this year, yet it's trading at a price-to-earnings (P/E) ratio of just 26, which is a 12% discount to the Nasdaq-100 index. Therefore, for all of the company's success, investors can still buy in now at a relatively cheap price.
4. Nvidia
This list wouldn't be complete without Nvidia, which has an estimated 90% share in the market for data center chips designed for AI workloads. The company has captivated investors recently with its explosive growth, sending its stock price more than tripling in 2023 so far.
Every tech giant with a cloud computing platform has been purchasing Nvidia's AI chips hand over fist, as have start-ups like OpenAI. In the recent fiscal 2024 second quarter (ended July 30), Nvidia's data center revenue surged by 171% year over year to $10.3 billion. The company's CEO Jensen Huang says there is more than $1 trillion worth of existing data center infrastructure that needs to be upgraded to support accelerated computing and AI, so there might still be plenty of growth ahead.
But Nvidia's dominance isn't just about hardware. The company has built an ecosystem around its GPU chips, with its Compute Unified Device Architecture (CUDA) at the center. Developers use CUDA to build AI applications on top of Nvidia's GPUs, and it allows them to accelerate tasks like training models.
CUDA can't be used with other hardware. Therefore, even though competitors like AMD and Amazon have launched their own AI chips, cloud providers can't move away from Nvidia without considering the needs of developers who might prefer staying within its software ecosystem.
Nvidia stock is expensive compared to the Nasdaq-100 right now, with a sky-high P/E ratio of 104. But measured against Wall Street's estimates for the company's earnings next year, its forward P/E is a more reasonable 26. As a result, there might still be gains on the table for investors willing to hold Nvidia stock for at least a couple of years.
