Three decades ago, the Internet began to enter the mainstream, setting off a chain of events that forever changed the growth trajectory of American business.

Wall Street has been waiting, sometimes impatiently, for the next big trend that would rival the Internet for business. After a long wait, the artificial intelligence (AI) revolution seems ready to answer that call.

AI gives software and systems the autonomy to take over tasks normally done by humans. The key to AI’s long-term success and the source of its seemingly limitless power is the ability of AI-driven software and systems to learn without human intervention. Machine learning gives AI the potential to perform existing tasks more competently and learn new skills.

No company has benefited more directly from the rise of AI than the semiconductor giant NVIDIA (NVDA -0.21%).

Image source: Getty Images.

Until recently, Nvidia’s operational ramp-up was running smoothly

At the beginning of 2023, Nvidia had a market capitalization of $360 billion, making it one of America’s most influential technology companies. On June 20, 2024, less than two weeks after the completion of its historic 10-for-1 stock split, Nvidia’s market capitalization reached an intraday high of $3.46 trillion. Investors have never seen a market leader gain more than $3 trillion in value in less than 18 months.

The catalyst behind this potentially one-off move is the company’s AI graphics processing units (GPUs), which have become standard in high-performance enterprise data centers. Semiconductor analysts at TechInsights estimate that Nvidia was responsible for all but 90,000 of the 3.85 million GPUs shipped to enterprise data centers in 2023.

Because demand for the company’s chips exceeds supply, Nvidia dramatic increase the selling price of its superstar AI acceleration chip H100. Within five quarters, the company’s adjusted gross margin increased by around 13.7 percentage points to 78.4%.

The company’s hardware’s focus in enterprise data centers has also driven ongoing innovation. In March, Nvidia unveiled its next-generation Blackwell platform, capable of accelerating computing capacity in a number of areas, including generative AI solutions, while consuming less energy than its predecessor. In June, CEO Jensen Huang announced the debut of its Rubin GPU architecture, powered by a new processor called “Vera.” Rubin is scheduled to launch in 2026.

The final piece of the puzzle for Nvidia’s exemplary production ramp-up so far has been the capacity increase of its suppliers in order to meet the strong demand. For example, the world’s leading chip manufacturer Semiconductor manufacturing in Taiwan (TSM 1.56%) has increased its chip-on-wafer-on-substrate (CoWoS) capacity required to package high-bandwidth memory in AI-accelerated data centers.

Nvidia is no longer error-free

This seemingly textbook “recipe” as the leader of the hottest trend on Wall Street briefly helped Nvidia outperform Microsoft And Apple to become the largest publicly traded company. But after a rough few weeks for Nvidia and the stock market as a whole, it’s become clear that Nvidia is just as fallible as any other company.

To maintain its historic lead, Nvidia had to execute its plans flawlessly. The company had to be able to sell all of its hardware, charge a premium price for its products and software – such as the CUDA platform that helps developers build large language models – and maintain its competitive edge by bringing its next-generation GPU architecture to market on time.

Unfortunately, reports surfaced this past weekend that Nvidia has informed many of its top customers (all members of the “Magnificent Seven”) that shipments of its Blackwell chip will be delayed by at least three months. This would push delivery from an expected arrival date later this year to the first quarter of 2025.

According to various reports, the delay is due to possible design flaws at Blackwell as well as capacity constraints at Taiwan Semiconductor (TSMC). Even though TSMC is effectively doubling its CoWoS capacity, this is nowhere near enough for Nvidia to meet demand from enterprise customers.

Blackwell’s delay is the first domino that shows Nvidia isn’t perfect. It also opens the door for Nvidia’s competitors to succeed.

On July 30 Advanced micro devices (AMD -1.50%) delivered second-quarter operating results that were welcomed with open arms by Wall Street and investors. AMD’s data center segment revenues increased 115% compared to the same period last year and 21% quarter-over-quarter (i.e., compared to what was reported for the quarter ending in March). AMD attributed this outperformance to the expansion of AI GPUs.

In particular, AMD’s MI300X is significantly cheaper than Nvidia’s H100. Even though the H100 offers a number of computational advantages over the MI300X, the delay in delivery of the H100, combined with the now delayed Blackwell chip, gives AMD’s hardware a lot more shine.

What’s more, all four of Nvidia’s top customers have developed AI chips for use in their data centers. Even if Nvidia maintains its compute lead, it will lose valuable data center space as its top customers choose to install their internally developed (and low-cost) chips.

Image source: Getty Images.

History suggests that Nvidia’s sell-off will intensify

To make matters worse, there hasn’t been a single major innovation, technology or trend in the last three decades that has escaped an early-stage bubble. Aside from the advent of the internet, investors have seen early bubbles burst in genome decoding, business-to-business trading, housing, Chinese stocks, nanotechnology, 3D printing, blockchain technology, cryptocurrency, cannabis and the metaverse.

The problem with breakthrough innovations and technologies is that investors overestimate their adoption and utility. No matter how large the addressable market, it takes a while for new technologies to change the growth landscape for the American economy.

For example, although most leading companies are investing a lot of money in AI-driven data centers, many lack a clear plan on how artificial intelligence will help them increase their revenues and make more money. Investors saw the same story just a few years ago with the rise of the metaverse and before that, blockchain technology. All innovations need time to mature – no exceptions!

Over three decades, market leaders have consistently declined 80% or more with each new trend. On a high-low basis, the leading companies behind internet/networking, 3D printing, genome decoding, cannabis, cryptocurrency, and blockchain technology all plunged 90% or more before bottoming out.

The silver lining for Nvidia is that the company has several established segments beyond its AI GPU operations that can provide a better foundation than what other market leaders had when their respective bubbles burst. Nvidia’s GPUs, used in gaming and crypto mining, combined with its virtualization software, should prevent a total collapse.

Nevertheless, the story is clear: large A pullback is in order once the euphoria surrounding the next big trend fades. Blackwell’s delay is the first domino to fall, and it strongly suggests that the Nvidia sell-off will intensify in the coming weeks, months, or quarters.