For much of the last 19 months, the bulls have been in control on Wall Street. All three major stock indexes have catapulted to multiple record-closing highs this year, with the rise of artificial intelligence (AI) being the primary catalyst.
The excitement surrounding AI has to do with the capacity for software and systems to learn over time without human intervention. This machine-learning ability should allow AI-driven software and systems to become more proficient at their assigned tasks and learn new skills, thus giving the technology utility in most sectors and industries.
Although estimates vary wildly for AI, the analysts at PwC released a report last year ( Sizing the Prize ) claiming the technology could add $15.7 trillion to the global economy via increased production and consumption-side benefits, by 2030.
No company has benefited more from the sky-high euphoria surrounding artificial intelligence than semiconductor titan Nvidia (NASDAQ: NVDA) .
Image source: Getty Images.
Until recently, Nvidia's operating execution had been virtually flawless
The reason investors have flocked to Nvidia above all other AI companies is its hardware. Its H100 graphics processing unit (GPU) rapidly became the standard chip used by businesses running generative AI solutions and training large language models (LLMs) in high-compute data centers. Nvidia was responsible for all but 90,000 of the 3.85 million GPUs shipped for use in data centers last year, per TechInsights.
What's more, the company's hardware is in such high demand that it and its next-generation chips are backlogged. When demand for a good or service overwhelms supply, it's perfectly normal for the price of said good or service to climb. A substantial increase in the price of Nvidia's H100 GPU has boosted the company's adjusted gross margin by 13.7 percentage points over the last five reported quarters.
CEO Jensen Huang has also overseen a major investment in ongoing innovation that should help Nvidia retain its GPU advantages . The company's next-gen Blackwell chip is aimed at accelerating capacity in six areas, including generative AI, while using less energy than its predecessor chip. Meanwhile, the Rubin platform, which will run on the all-new Vera processor, is set for delivery sometime in 2026.
These catalysts have helped to lift Nvidia's market cap by $2.4 trillion since the start of 2023, translating into a gain of 695% for investors as of the closing bell on Aug. 13.
While things have seemingly been picture-perfect for Nvidia and its shareholders, my prediction is that the going is about to get significantly tougher for Wall Street's AI darling.
Nvidia is about to get crushed by competition -- but likely not in the way you think
Every publicly traded company faces headwinds and/or competitive pressures. Despite maintaining a near-monopoly in AI-GPUs for data centers, Nvidia is liable to be crushed by competitive pressures over the coming two years.
Some of you might be thinking I have no clue what I'm talking about given the well-defined advantages Nvidia's H100 and upcoming Blackwell chips pack over other key hardware players -- and I don't disagree with you. There's a really good chance Nvidia's willingness to spend aggressively on research will allow the H100, Blackwell, and possibly even Rubin platform to remain at the top of the pack, in terms of computing ability.
The problem for Nvidia is that computing capacity represents only one of the factors businesses are considering when building out their AI-accelerated data centers. That capacity is, undeniably, important...but it's not everything .
Despite efforts by global chip-fabrication leader Taiwan Semiconductor Manufacturing , its chip-on-wafer-on-substrate (CoWoS) capacity is still insufficient to cover Nvidia's needs. CoWoS is a veritable necessity for packaging the high-bandwidth memory needed in high-capacity data centers. In other words, Nvidia's chips might be faster, but the company can't meet all of its orders or deliver anytime soon.
When you're talking about first-mover advantages in generative AI and LLMs, at least some businesses aren't going to wait to fill valuable hardware "real estate" in their data centers. Advanced Micro Devices (NASDAQ: AMD) has been ramping up production of its MI300X AI-GPU, which has a median price point of around $15,000, compared to the H100, which comes in at around $30,000 per chip. AMD's chip might not have a computing advantage over the H100, but at roughly half the cost and with far less of a backlog, it has AMD sitting pretty .
AMD isn't the only external competitor vying for data center real estate. With reports suggesting Nvidia's Blackwell chip will be delayed by at least three months due to design flaws and supplier constraints, tech juggernauts Samsung and Huawei are entering the picture with AI chips of their own.
But wait -- there's more.
Nvidia's top four customers by net sales, which are all members of the "Magnificent Seven," are developing AI-GPUs for their data centers, too .
More than likely, these chips don't have a chance of outperforming the H100 or Blackwell on the basis of computing capacity. But that's not going to stop Microsoft , Meta Platforms , Amazon , and Alphabet from complementing the chips they have purchased from Nvidia with internally designed chips that will ultimately be substantially cheaper and easier to access. Once again, we're talking about Nvidia's hardware losing out on valuable data center real estate.
Image source: Getty Images.
The final piece of the puzzle
On top of my forecast that Nvidia's stock can be clobbered by competitive pressures over the next two years, the company will also have to contend with three decades of undefeated history when it comes to next-big-thing innovations, technologies, and trends.
Over the last 30 years, no shortage of game-changing technologies and trends have graced Wall Street. The advent of the internet, genome decoding, business-to-business commerce, housing, China stocks, nanotechnology, cryptocurrency, 3D printing, blockchain technology, cannabis, augmented/virtual reality, the metaverse, and now AI are just some of these growth-powering trends.
With the exception of AI, every other buzzy innovation, technology, or trend in the last 30 years navigated a bubble in its early stages. This early-innings bubble is evidence that investors regularly overestimate how quickly new technologies and trends are going to be adopted at the consumer and/or enterprise level. When lofty targets aren't met and the initial euphoria fades, the bubble eventually bursts .
There's nothing to suggest that artificial intelligence is a mature technology. In fact, most businesses lack a clearly defined game plan as to how AI will be used to grow their sales and improve their bottom lines. This is all the evidence we need that euphoria surrounding AI has stepped well past the bounds of reality.
If and when the artificial intelligence bubble bursts (which history suggests it will), no stock will be hit harder than Nvidia.
Although long-term investors in Nvidia are still going to be up a boatload if Nvidia retraces 50%, 60%, or even 80% , the next two years are shaping up to be challenging for Wall Street's AI darling.
John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool's board of directors. Randi Zuckerberg, a former director of market development and spokeswoman for Facebook and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool's board of directors. Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Sean Williams has positions in Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Meta Platforms, Microsoft, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy .
Most Popular
Comments / 0