There seem to be two passions on Wall Street right now: anything related to artificial intelligence (AI) and companies doing stock splits. What you may not know is that these two trends overlap at three companies.

A stock split is a tool available to publicly traded companies to change their stock price and the number of shares outstanding. Keep in mind that these changes are purely superficial and stock splits do not affect a company’s market capitalization or underlying operating performance.

Although there are two types of stock splits – forward (reducing notional value) and backward (increasing notional value) – most investors tend to focus on companies that conduct forward stock splits. Because successful companies outperform their peers in terms of innovation and performance, forward stock splits are almost always conducted from a position of operating strength. The same can rarely be said of publicly traded companies that conduct backward stock splits.

In the last six-month period, 13 notable companies have announced or completed a stock split, all but one of which were forward splits. Three of these completed or anticipated stock splits are companies leading the artificial intelligence revolution.

Image source: Getty Images.

Nvidia, Broadcom and Super Micro Computer have all announced historic stock splits

The excitement surrounding the rise of AI is palpable. According to analysts at PwC, AI is expected to add $15.7 trillion to the global economy by the end of the decade. This enormous market suggests that this technological leap forward could produce several big winners.

On May 22, Hardware Guide NVIDIA (NVDA 6.53%) began with the announcement of a 10-for-1 split – the largest split in the company’s history. Shortly after the split was implemented after the close of trading on June 7, Nvidia stock reached a market capitalization of $3.46 trillion, briefly making the company the most valuable publicly traded company in the world.

Nvidia’s H100 graphics processing unit (GPU) has become the standard in high-performance enterprise data centers that want to train large language models (LLMs) and run generative AI solutions. A virtual monopoly on the “brains” behind AI-accelerated data centers, coupled with exceptionally strong pricing power due to the scarcity of AI GPUs, has driven Nvidia’s stock price higher.

Specialist for AI networks Broadcom (AVGO 5.07%) was next in line to join the Class of 2024 stock split club. The company’s board of directors approved a 1-for-1 split (its first ever) on June 12, effective after the close of business on July 12.

Broadcom’s networking solutions (e.g., the Jericho3 AI Fabric) are known for connecting large numbers of AI GPUs in enterprise data centers to reduce latency and ensure that companies get as much computing power out of those chips as possible. In other words, Broadcom is helping to further accelerate the split-second decision-making required by AI-driven software and systems.

Finally, customizable rack server and storage specialist Super-microcomputer (SMCI 4.89%) joined the club by announcing its first-ever stock split, also at a 1:10 ratio, on August 6, which will take effect after the close of trading on September 30.

Super Micro’s server solutions are in high demand as companies build the physical infrastructure needed to train LLMs, run generative AI solutions, accelerate quantum computing, and more. Super Micro more than doubled its revenue to $14.94 billion in fiscal 2024 (which ended June 30), and is widely expected by analysts to deliver 75% revenue growth to over $26 billion in the current fiscal year.

Two of these three AI stock splits are not worth buying

But despite the euphoria surrounding stock splits, not all leading companies involved in the AI ​​revolution are worth buying.

The harsh reality of the next big innovations is that they have a history of disappointment. Over the past three decades, every exciting innovation, technology or trend that attracted investors with big dollar signs eventually succumbed to an early-stage bubble.

The reason why bubbles keep forming on Wall Street when new technologies or trends emerge is because investors (both professional and private) have a habit of overestimating how quickly new innovations will be adopted and used. Investors are constantly over-optimistic and fail to take into account that innovations, technologies and trends all take time to mature.

A clear sign that artificial intelligence is far from being a mature technology is the lack of concrete plans from tech companies on how they plan to use AI to increase their revenues and boost their profits. While some companies may already have rough expectations, the vast majority of companies lack a real AI game plan. This all but guarantees that the AI ​​bubble will burst at some point in the future.

If the story is true and the AI ​​bubble bursts, no company would be hit harder than Nvidia. Most of the stock’s value increases and recent revenue growth have been driven by the shortage of AI hardware and AI GPU. Once that shortage eases, which is expected as external and internal competition for data center real estate increases, Nvidia stock could see a long way to fall.

Super Micro Computer is in a similar situation. Almost all of its recent revenue jump has been due to data center expansion. If the AI ​​bubble bursts, or if companies realize they don’t know how to get a positive return on their AI investments, Super Micro’s future server orders and its stock will suffer.

Additionally, Super Micro’s rack servers are equipped with Nvidia’s leading H100 GPUs, which are currently on backorder. The longer supply shortages persist, the more likely it is that Super Micro will not be able to reach its potential.

Image source: Getty Images.

This is the only AI stock split stock worth buying right now

Of the three companies traded on Wall Street as part of the artificial intelligence (AI) stock split, Broadcom is the crème de la crème from an investor’s perspective.

While Broadcom has benefited greatly from increased demand for its AI-driven networking solutions, the company already had extensive and diverse revenue streams—and a huge backlog—before artificial intelligence became the hottest topic on Wall Street.

Broadcom, for example, is a leading supplier of wireless chips and other accessories used in next-generation smartphones. Telecommunications companies are spending heavily to upgrade their mobile networks to 5G download speeds, triggering a multi-year device upgrade cycle that has increased demand for Broadcom’s wireless solutions.

Broadcom has its fingers in many other areas beyond the smartphone sector. The company offers software for financial services, cybersecurity solutions, and optical products for next-generation vehicles and industrial equipment, to name a few.

Broadcom’s management team has also used acquisitions as a means to expand the company’s product and service ecosystem and increase its profits, including the acquisition of cybersecurity solutions provider Symantec in 2019 and the $69 billion deal to buy VMware in November 2023. The VMware purchase is particularly important to advance Broadcom’s strategy of helping enterprises meet the challenges of adapting private, hybrid and multi-cloud environments.

The key point here is that the potential bursting of the AI ​​bubble will not turn Broadcom’s world upside down the way it would for Nvidia and Super Micro Computer. While Broadcom would almost certainly see a decline or even slowdown in AI networking revenue, many of its other sales channels would be unaffected.

At 24 times forward annual earnings, Broadcom is admittedly not as good a deal as it was two years ago. But with earnings per share expected to grow 18% per year over the next five years, the valuation offers room for earnings per share growth.