Last week, the iconic Dow Jones Industrial Average and benchmark S&P 500 surged to record highs, while the growth-powered Nasdaq Composite came within a stone's throw of eclipsing its all-time closing high from November 2021. Though growth stocks are most directly responsible for this move higher in the broader market, the more specific trend that's spurring this rally is excitement surrounding artificial intelligence (AI).
AI uses software and systems to undertake tasks that would normally be done by humans. The key ingredient is machine learning, which allows AI-driven software and systems to effectively learn over time and become smarter and more effective at their tasks.
The beauty of AI is that it isn't just grounded in the technology space. It has application in virtually all sectors and industries, which is why the analysts at PwC believe it can add an estimated $15.7 trillion to global gross domestic product by the turn of the decade.
Image source: Getty Images.
However, not every Wall Street analyst is on board with the AI hype train. Based on the bottom-of-the-barrel price targets provided by select analysts, three of the highest-flying and/or widely owned AI stocks could plunge between 43% and 88%.
Nvidia: Implied downside of 43%
The first artificial intelligence stock that could quickly lose its luster, at least according to one Wall Street analyst, is the infrastructure backbone of the AI movement: semiconductor company Nvidia (NVDA 6.18%). According to analyst Gil Luria of D.A. Davidson, Nvidia is worth $410 per share, which as of its close at $721 on Feb. 9 implies 43% downside.
Though Luria acknowledges that Nvidia's A100 and H100 graphics processing units (GPUs) are the undeniable go-to for high-compute data centers, he cautions against getting swept up by the short-term euphoria. Luria's January 2024 note that set the $410 price target anticipates that Nvidia's growth trends "will start reverting to the trend line within the next two to six quarters."
Every next-big-thing trend over the last 30 years has worked its way through an initial period of unsustainable hype. While this doesn't mean that companies in the AI space won't, eventually, be successful, it makes it highly unlikely that enterprise demand can match or surpass the already lofty expectations of investors.
The bigger issue for Nvidia is that it's set to face mounting external and internal competition. Advanced Micro Devices and Intel are both rolling out GPUs designed to challenge Nvidia's top-notch hardware. AMD is ramping up production of its MI300X AI-accelerated GPU, while Intel expects to bring its Gaudi3 generative AI software chip to market sometime this year.
With regard to external competition, many of Nvidia's largest customers from a sales perspective are developing their own AI-focused chips. This includes Microsoft and Meta Platforms, which are the respective No. 1 and No. 2 in spending with Nvidia.
Perhaps the nail in the coffin for Nvidia could be that increasing production of its A100 and H100 chips is liable to hurt its gross margin. Much of the company's sales growth in fiscal 2024 (ended in late January) was the result of AI-GPU scarcity. When Nvidia boosts its own output, it'll be inadvertently cannibalizing its pricing power.
This outlier price target for Nvidia might prove realistic, based on what history tells us about early-stage bubbles in next-big-thing trends.
Super Micro Computer: Implied downside of 66%
A second supercharged AI stock that at least one Wall Street analyst believes could implode, based on its closing share price of $740.29 on Feb. 9, is Super Micro Computer (SMCI 8.89%). Senior equity research analyst Mehdi Hosseini of Susquehanna recently defended his $250 price target on Super Micro on CNBC, which if accurate would translate into implied downside of 66%.
Shares of Super Micro are up 751% over the trailing year and a scorching-hot 160% in the six weeks since 2024 began. These phenomenal gains are being driven by the company's energy-efficient and customizable servers that are proving integral to enterprise-operated AI-accelerated data centers.
Super Micro has a very close-knit relationship with Nvidia (it uses Nvidia's GPUs in its high-end servers), which also helps to explain its astronomical gains.
But even with its full-year sales forecast coming in well ahead of its own prior estimates in fiscal 2024 (the company's fiscal year ends June 30), there are potential headwinds that could disrupt this highflier.
As noted, investors have a terrible habit of overestimating the adoption of new technologies and innovations. Super Micro Computer was in this situation a little over a half-decade ago as one of the expected beneficiaries of rapid growth in cloud computing.
However, Super Micro's growth ultimately fell well short of what optimists had initially expected. While this hasn't been the case early on in the AI cycle, there's a precedent that bubbles are common with next-big-thing trends.
Something else to consider is that Super Micro's growth is currently constrained by AI-GPU scarcity. Even though Taiwan Semiconductor Manufacturing is increasing its chip-on-wafer-on-substrate capacity, there's the real possibility Super Micro Computer's growth will be held back by supply constraints.
And Super Micro's ability to defend its intellectual property (IP) could be called into question. While it's having no trouble riding Nvidia's coattails for the time being, my Foolish colleague Nicholas Rossolillo pointed out in early January that a majority of the company's enterprise computing and storage systems haven't been granted patents or lack IP protections. This suggests competition could eventually siphon away market share from Super Micro Computer.
A Tesla Model S charging. Image source: Tesla.
Tesla: Implied downside of 88%
The third AI stock that could absolutely plunge, based on the prognostication of one Wall Street analyst, is electric-vehicle (EV) manufacturer Tesla (TSLA -1.11%). Longtime Tesla bear and GLJ Research founder and CEO Gordon Johnson has a $23.53 price target on the world's largest automaker by market cap. This implies a jaw-dropping 88% downside from where shares closed this past week.
Johnson's oddly specific price target was arrived at by placing a 15 multiple on his earnings estimate for Tesla in 2025 and applying a 9% discount rate to arrive at the current price.
Although Tesla isn't as directly an AI company as Nvidia or Super Micro Computer are, its vehicles are increasingly reliant on technology and AI in a fiercely competitive industry. For instance, Tesla's full self-driving software is an example of advanced technology that uses cameras, ultrasonic sensors, and radar to steer a vehicle away from potential obstacles.
Despite being the only pure-play EV maker that's currently profitable on the basis of generally accepted accounting principles (GAAP), Tesla does look ripe for a breakdown.
Since the start of 2023, it has slashed the sales price of its four production models (the 3, S, X, and Y) on more than a half-dozen occasions. CEO Elon Musk has been crystal clear that Tesla's price strategy is based on demand for the company's EVs. Ongoing price cuts suggest that increasing competition and rising inventory levels are taking their toll.
The proof of this can also be seen in Tesla's operating margin. Since the September-ended quarter of 2022, the company's operating margin has been more than halved to 8.2%. With price cuts continuing into 2024 and days of supply remaining high, relative to previous years, there's the real possibility of continued operating-margin contraction.
What's more, Tesla is generating a sizable percentage of its pretax profit from unsustainable sources, such as interest income and regulatory tax credits. This isn't something you would expect to see from a high-growth company with a big premium.
Lastly, Musk has arguably proved to be more of a risk than a benefit to Tesla. Putting aside his encounters with securities regulators, Musk has frequently made innovative promises he has yet to keep. If those unfulfilled innovations are backed out of Tesla's valuation, Johnson's seemingly lowball price target becomes realistic.
