Dividend Stocks

3 AI Stocks to Buy as Nvidia Sputters

We all know by now that compound interest is hard to visualize. As early as 1256 AD, storytellers began recording the wheat and chessboard problem. It’s a story where a king agrees to reward a subject with one grain of wheat on the first square of a board, two on the second square… and continues doubling the amount until the 64th square is reached. 

The ruler quickly realizes his mistake. By the 32nd square, he owes roughly 279 metric tons of wheat. And if we add up everything to the 64th square, the king would theoretically owe his subject 1,600 years of modern wheat production! 

Now, what happens if we apply the same principle to computing power? How crazy can things get?

That’s exactly what’s happening today. Since the 1970s, the number of transistors on microchips has roughly doubled every two years, which has increased computing speeds by about the same factor. This observation has become known as Moore’s Law, and it’s why your typical graphing calculator now has more processing power than the entire Apollo 11 moon mission once did. 

It’s hard to overstate the importance of this exponential growth. Today’s large language models (LLMs) essentially brute-force their way through terabytes of training data, which has only been possible with the latest generation of GPU processors.  

But that’s only the start. 

That’s because computing speeds are still doubling every 18 to 24 months. The latest GPT models will be obsolete by next year. And the next-generation models will be outdated soon after. 

In fact, researchers believe that if we keep doubling computing speeds at current rates, we could begin reverse-engineering human brains by the end of the decade! So, rather than try inventing a new AI, researchers could theoretically simulate all 86 billion neurons in a human brain and see what happens. 

As Jeff Remsburg wondered in this space earlier this week: What could possibly go wrong?

The Road to AGI 

These issues are extremely concerning to InvestorPlace’s global macro expert, Eric Fry. In his latest presentation, scheduled for this Thursday at 1 p.m. Eastern (sign up here), he outlines how artificial general intelligence (AGI) will make AI smarter than humans, and how this moment is approaching far faster than people realize. If computing speeds continue doubling as they have, we could soon see smarter-than-human AI emerge as soon as 2029. 

So… then what happens? This is also something Eric talks about in depth in The Road to AGI Summit. We need to prepare for this eventual outcome. And his recommendation is to get ahead of the curve financially. Go here to get more details and save your seat. 

In the meantime, our writers at InvestorPlace.com also have been examining AI stocks that they believe will put you on the right side of this divide. 

When No. 2 Becomes No. 1 

I’m usually not a fan of the No. 2 player. Whenever there’s a head-on competition between two companies, the larger one will often earn slightly higher profits, which gives them a slightly larger R&D and marketing budget. Over time, this advantage grows larger, which is how you end up with companies like Nvidia Corp. (NVDA) dominating rivals like No. 2 Advanced Micro Devices Inc. (AMD) in GPUs.The historical (and sometimes depressing) truth of capital markets is that winners typically keep winning. 

However, AMD has finally found a path to become No. 1. And it’s doing that by quietly expanding in an industry without Nvidia as competition. 

In 2022, AMD bought Xilinx in a record $50 billion deal, giving the chipmaker access to the server chip market. And by combining this opening with its recently-launched Zen microarchitecture, AMD was able to create a solid position against rival Intel Corp. (INTC). The upstart has since doubled its server market share to 24% and used that know-how to break into the PC market as well. 

In an update at InvestorPlace.com this week, Yiannis Zourmpanos highlights how AMD’s latest lineup of processors is further extending this momentum. He now calls AMD one of his top “buy-and-hold forever” picks

The surge in data center revenue emerged from Instinct MI300 GPU shipments and EPYC CPU sales. The ramp-up in Instinct MI300 GPU shipments significantly boosted data center revenue, reflecting the high demand for AMD’s high-performance computing solutions. Further, double-digit increases in EPYC CPU sales further enhanced data center revenue… 

The client segment also grew, with revenue up 49% YOY to $1.5 billion. This increase was driven by AMD Ryzen processors and early shipments of the next-generation Zen 5 processors. 

Its strong revenue growth across all segments puts AMD among the top stocks to buy and hold. 

Though Nvidia will continue to dominate the GPU side of AI chips, we see AMD solidifying its position in the CPU side of that equation for years to come. 

The “Aramco” of Healthcare 

In 2006, British mathematician Clive Humby quipped that “data had become the new oil.” Information had become so important to business intelligence that having more of it had become a competitive advantage. 

AI has taken that truth to a whole new level. Large language models require even more amounts of training data than previous generations of business intelligence applications, and that demand is growing exponentially. Some speculate that OpenAI’s GPT-4 uses almost three times more data than GPT-3

That’s where Faisal Humayun’s latest 7X potential stock comes in. 

Tempus AI Inc. (TEM) is a Chicago-based healthcare company that has amassed the world’s largest library of clinical and molecular data. The company holds 200 petabytes of data and serves 95% of top pharma companies through data licensing. So, much like its Saudi Aramco counterpart, Tempus is the dominant player when it comes to controlling the lowest-cost “feedstocks” for downstream use. 

The company recently went public in an IPO that valued shares at $6 billion, and it’s already reporting some immediate financial gains. As Humayun explains in an InvestorPlace.com update this week

For Q2 of fiscal year 2024, Tempus AI reported revenue growth of 25% year-over-year to $166 million. Importantly, data licensing revenue increased by 40% YOY. Further, the company achieved positive adjusted EBITDA of $12.7 million quarter-over-quarter (QOQ). 

Even more importantly, he sees greater fundamental gains to come as Tempus AI builds its internal abilities to harness this healthcare data. 

With significant investment in research & development, Tempus AI is positioned to accelerate growth. In June, the company received U.S. Food and Drug Administration (FDA) approval for its “Tempus ECG-AF device that uses AI to help identify patients who may be at increased risk of atrial fibrillation.” 

Together, that has analysts forecasting better than 30% revenue growth rates through 2027. Though we might not know exactly which company will create the next blockbuster drugs, Tempus is an AI superstar that should benefit, regardless. 

The New American Superstar 

Years ago, Intel was the American chipmaker to buy. The firm’s vertical integration meant they could pour excess profits into more R&D, which kept them well ahead of the competition. CPUs were the workhorses of the computing industry, and Intel was the stablemaster. 

Today, the rise of AI has turned that equation on its head. Large language models require specialized chips, which have boosted demand for Nvidia GPUs and AMD’s power-efficient CPUs. A whole industry of other chipmakers and suppliers have emerged to house these processors. 

Among these winners is Micron Technology Inc. (MU), a company that Louis Navellier and his team describe as a chip stock to ride the 2024 AI wave: 

Micron makes the world’s leading High-Bandwidth Memory 3E, or HMB3E, architecture. 

It provides higher bandwidth than previous generations of memory, while being smaller and using less energy. Nvidia uses the HBM3E architecture in its latest graphics processing units, including the new H200. 

In other words, Nvidia’s GPUs might be fast at calculations, but much of its capacity is currently wasted. As The Economist notes, some GPU cores may sit idle as much as half of the time as they wait for data to process

Micron’s latest ultrafast chips attempt to reduce these bottlenecks by storing and transferring data faster than before. According to Nvidia, this 50% improvement in speeds is enough to increase overall performance 3X. Louis and his team also note that Micron is experimenting with AI chips that could power the next generation of AI PCs. 

Of course, it’s important to note that Micron lacks the same “moat” that AMD and Tempus have. Most of Micron’s revenues are still derived from commodity-like DRAM and NAND memory, which are interchangeable with those produced by rivals. Micron’s return on invested capital (ROIC) has averaged 9.8% over the past-5 years, roughly the same as the S&P 500. 

However, not only are insatiable needs from AI driving a cyclical upswing in industry-wide demand, but they’re also providing firms like Micron with the funds to develop novel chips and finally break out of their commodity markets. Analysts expect ROIC to rise to 15% next year, and 18% in 2026. 

The Winding Road to AGI 

Imagine yourself in 1999. The internet was still mostly accessed through broadband, Amazon was still a startup, and Wi-Fi had just been invented two years before. 

What if someone told you then that, within a decade, people would be using “tablet or smaller sized devices” to connect to the internet? Or that by 2019, machines could regularly pass the Turing test of intelligence, where a conversation with them is indistinguishable from talking with a human? 

Many people would have called them crazy. 

But that’s precisely what computer scientist Ray Kurzweil predicted would happen in The Age of Spiritual Machines, a bookpublished in 1999 that laid out his forecasts for the 21st century. By mapping out the rate of computational improvements, he was able to make some startlingly accurate predictions. 

Not every forecast came true, of course. Even he admits he got several dozen predictions wrong. But Kurzweil – who is now a top AI researcher at Alphabet Inc. (GOOGL) – has gotten far more correct than the typical futurist. 

That’s why one of his “crazy” predictions for 2029 needs to be taken seriously. In just five years, Kurzweil believes computing power will become so fast that AI will surpass human intelligence. 

That will have enormous consequences for us all. Super-intelligent AI could theoretically make even smarter versions of itself… or we might find a way to harness its power to extend human thought… the possibilities are endless. 

After the market rout earlier this month, it may seem like AI hype is fading… but we are already well on The Road to AGI. We are in a window where the next great trade is setting up right now, and Eric Fry says Wall Street is sleeping on it.  

And if you set yourself up for that now, you have the potential to build more wealth than you ever imagined possible — all thanks to the world-defining power of AGI. 

That’s why Eric is hosting a free event on Thursday, August 22, at 1 p.m. Eastern. The Road to AGI Summit isopen for anyone to attend — but it’s critical that you register ahead of time to reserve your spot 

At that event, Eric will deliver a three-part “future-proof” blueprint to everyone in attendance. Plus, he’s bringing a brand-new stock idea for The Road to AGI that he’s excited to share. Again, this is ALL FREE for attendees.  

Register now for The Road to AGI Summit on August 22. 

The opinions expressed in this article are those of the writer, subject to the InvestorPlace.com Publishing Guidelines.

On the date of publication, the responsible editor did not have (either directly or indirectly) any positions in the securities mentioned in this article.

Thomas Yeung is a market analyst and portfolio manager of the Omnia Portfolio, the highest-tier subscription at InvestorPlace. He is the former editor of Tom Yeung’s Profit & Protection, a free e-letter about investing to profit in good times and protecting gains during the bad.

Newsletter