3 Top Stocks to Profit From the Coming AI Explosion

AI is taking the investing world by storm at an interesting time. Tech giants have been announcing tens of thousands of layoffs in recent months as the economy begins to slow. Meanwhile, new AI services like ChatGPT (Microsoft's new toy) and Alphabet's Bard and investment in AI start-up Anthropic are receiving billions of dollars' worth of attention. Perhaps Meta Platforms (CEO Mark Zuckerberg summed it up best, stating 2023 would be the "year of efficiency." Human workers are out, and big-tech, super-efficient AI systems are in.  

But we humans can strike back by investing in the companies that are enabling this AI in the first place. Here's how. 

The data center infrastructure spending spree is still in effect

A data center build-out arms race has been in effect for years now. Big tech, in particular, has been spending heavily to support cloud computing and AI workloads. Meta specifically reported a massive surge in capital expenditures (or capex, spending on property and equipment) in support of Facebook, Instagram, and WhatsApp, as well as its Reality Labs VR segment in 2022.  

It's not just Meta in heavy-spend mode, though. All the cloud software giants are gearing up for another decade of cloud growth. Microsoft may have just fired the starting gun on the next race with its $10 billion investment in ChatGPT creator OpenAI, spending which is not yet included in the 2022 CapEx table below.

Data source: YCharts. *Amazon capital expenditures include an undisclosed sum on e-commerce fulfillment centers not related to AWS cloud data centers.  

But what of 2023 being the year of efficiency? Much of big tech's spending cutbacks will be on employees, office space, and other variable operating expenses. Meta is a perfect example of how this will play out. At the end of 2022, the social media giant reported total employees of 86,482, which doesn't include most of the 11,000 job cuts it has announced. However, 2023 capex (mostly on data centers and related technology) is expected to be $30 billion to $33 billion. That's lower than the previous forecast for $34 billion to $37 billion, but the upper end of that guidance would still exceed the more than $31 billion in capex in 2022.

Indeed, the "year of efficiency" is good news for the machines.

How to profit from the rise of AI

For every AI system put into operation (be it ChatGPT, Alphabet's counterattack Bard, or a user recommendation via Facebook or Instagram), there's a semiconductor behind the scenes working away.

In fact, with the commercialization of AI, a lot more chips will need to be employed. Nvidia (NVDA 0.76%) CEO Jensen Huang has been discussing this for years, indicating that the market for inference (when AI is asked to complete a task) will be much larger than that for AI software training. Back in the autumn of 2020, Huang had said that the vast majority of inference work was done by CPUs (like those made by Intel (NASDAQ: INTC)). But with AI complexity only increasing, the tech world has been shifting toward the GPU, Nvidia's specialty.

Nvidia is a pricey stock at this juncture (90 times trailing 12-month earnings, or 44 times one-year forward expected earnings) after nearly doubling in price from October 2022 until this writing. But if you believe the battle for AI supremacy is only heating up, Nvidia could still be a top way to profit over the next decade. Meta recently said it is pausing some data center build-outs as it finalizes a new architecture built for AI. Details are scant at this point, but this new construction almost certainly favors heavier use of GPUs.  

Advanced Micro Devices (AMD 0.69%) should be another top beneficiary. The company has been gobbling up market share of the data center CPU space from Intel for years, and the trend will likely continue. AMD recently announced new versions of its fast-growing Epyc CPUs, named Genoa. A new CPU-GPU combo unit called the MI300 -- an APU (accelerated processing unit) -- will also come out in the second half of 2023.

Paired with its early 2022 acquisition of enterprise chip designer Xilinx, AMD has made a hard pivot away from consumer electronics to address the fast-moving data center industry. At just 19 times one-year forward expected earnings, AMD looks like a top semiconductor stock buy right now.  

Then there's Marvell Technology Group (MRVL -1.86%), which also made a few acquisitions in recent years to boost its network connectivity chip lineup. New processors and accelerators are needed to process AI workloads. But next-gen infrastructure that carries massive amounts of data within the cloud and then delivers the final product to the user is also a key part of modern AI-powered applications. 

Marvell is a leader in this high-throughput networking space. As companies like Meta design their data centers around new AI, Marvell will most certainly be an ingredient in their recipes. Shares trade for just under 16 times one-year forward expected earnings.  

During the last earnings call, Zuckerberg highlighted Meta's new focus on "delivering compounding earnings growth even while investing aggressively in future technology." It's great news for investors that big tech is shifting its focus to profitable growth, but semiconductor companies have already been focused on this all along. As the AI battle brews among Meta, Alphabet's Google, Microsoft, Amazon, and others, chip stocks could be the best way to profit in the years to come.