The "Magnificent Seven" stocks have been setting the standard for market-defining growth investments for more than a year now. Led by Nvidia (NASDAQ: NVDA), the designer of specialized chips for accelerated processing of artificial intelligence (AI) systems, these seven megacaps have gained an average of 85% in 52 weeks.
But stock groupings like the "Magnificent Seven" are always subject to change. Some of the current members aren't pulling their weight and could be up for replacement someday soon. Apple is only up by 10% over the same period, while Tesla actually fell 11% lower.
At the same time, a handful of stocks are bubbling up from below. In particular, Advanced Micro Devices (AMD 2.37%) looks ready to take Apple's or Tesla's spot in the "Magnificent Seven" society fairly soon.
Here's why I think AMD belongs in the "Magnificent Seven" discussion in 2024.
The shifting sands of AI supremacy
Nvidia may look like the undisputed ruler of the AI-chip hill so far, but the company's position is far from guaranteed.
For example, some of the world's largest and most powerful supercomputers are hard at work training the next generation of eye-opening AI systems. Nvidia's A100 or H100 accelerators are found in five of the 10 beefiest supercomputers nowadays.
At the same time, AMD's Instinct MI250X accelerator chips power two systems on that top-10 list -- including the dominant performer. In other words, system builders chasing extreme performance aren't strictly limited to Nvidia solutions.
Both companies are rolling out new generations of their AI accelerator lines right now; you should see their names in the next few versions of the elite-level supercomputer rankings. So far, performance benchmarks for the H200 and Instinct MI300X accelerators are limited to tests performed by Nvidia and AMD themselves, and the companies criticize each other for handpicking favorable benchmarks or parameters. Nothing new under the sun, right?
But it's fair to say that AMD offers comparable raw performance, which can turn into a massive advantage if its pricing is significantly lower. And then you have to consider the power requirements of each solution. Anyone can win performance comparisons by throwing more hardware at the problem, but the AI-building users of these muscular systems may pick hardware based on expected electric bills, cooling systems, and other real-world practicalities.
And AMD has a history of putting up a good fight on those terms. I can't wait to see the actual wattage measurements for Nvidia's H200 and AMD's MI300X accelerators.
AMD's spirited bid for AI accelerator leadership
So AMD isn't sitting on its hands, letting Nvidia snag every AI-training system contract without a fight. AMD's current solutions come with more and faster memory while Nvidia's generally claim the raw number-crunching crown, but you can find realistic use cases for either one. And again, I'm looking forward to the next round of AI accelerator upgrades finding end-market customers in 2024.
In all fairness, AMD is more than a few steps behind at the moment. The company recorded a bit more than $400 million in data center graphics processing unit (GPU) sales in the fourth quarter of 2023. Nvidia doesn't specifically break out its AI accelerator sales in financial reports, but the company's overall data center revenues quintupled to $18.4 billion in the same reporting period.
So AMD has some catching up to do, but the company wields a mighty AI accelerator design and relies on the same third-party manufacturing solutions as Nvidia. If the balance of AI power shifts away from Nvidia and toward AMD this year, people might start including the runner-up AI chip expert instead of Tesla in their "Magnificent Seven" discussions.
This dashing challenge from AMD is the main reason I'm not buying Nvidia stock right now. Instead, I recently cashed in some Nvidia profits and reinvested it in an undervalued growth stock instead. AMD's stock is also a little rich for my blood right now, but the company is a serious challenger to Nvidia's AI supremacy.
