2 Trillion-Dollar AI Stocks Facing 95% Plunge - Wall Street’s Boldest Bear Calls for 2025

Wall Street's boldest bears just painted targets on two AI giants—and their predictions could wipe out nearly all shareholder value.

Analysts see catastrophic downside in names everyone thought were untouchable. We're talking about companies that soared to trillion-dollar valuations on AI hype now facing potential 95% collapses.

The Numbers Don't Lie

These aren't minor corrections. The most pessimistic forecasts suggest these stocks could surrender virtually all their pandemic-era gains. When analysts start throwing around 95% downside targets, you know the fundamentals must be screaming.

AI Bubble Meets Reality

Revenue multiples stretched beyond recognition. Adoption curves slowing faster than expected. Competition eating margins alive. The classic signs of sector overheating are all there—yet most investors keep buying the dip like there's no tomorrow.

Wall Street's contrarian playbook is clear: when everyone's crowded into the same trade, even the mightiest can fall. And these analysts are betting that AI's day of reckoning is closer than anyone wants to admit.

Because nothing brings down overvalued tech stocks faster than old-fashioned mathematics finally catching up to the hype.

Image source: Getty Images.

Nvidia: Implied downside of 44%

The first trillion-dollar AI stock you might be shocked to see among possible plunge candidates is the world's largest publicly traded company,(NVDA -2.62%).

Bullishness on Wall Street for Nvidia is nearly universal. Out of the 65 analysts who've weighed in on the company, a combined 59 have it rated the equivalent of a buy or strong buy, as of September 2025, with another five chiming in with the equivalent of a hold rating.

The reason for this Optimism has to do with Nvidia dominating the AI-graphics processing unit (GPU) space. It began with the Hopper (H100) and has continued with the next-gen Blackwell and now Blackwell Ultra GPUs. No external chip companies are remotely close to matching the compute capabilities of Nvidia's AI hardware, which has led to a sizable backlog and enviable AI-GPU pricing power.

Nevertheless, Seaport Global analyst Jay Goldberg stands out as Wall Street's lone wolf when it comes to Nvidia. He has a sell rating on the company with a price target of just $100, which WOULD imply downside of up to 44%, if accurate.

Goldberg had a laundry list of warnings in his latest research note that maintained his and his firm's $100 price target, including:

• Concerns about a sequential quarterly slowdown in data center sales growth.
• Unjustified optimism surrounding agentic AI, which has a rather small commercial market at the moment.
• Nvidia's China exposure, with delays in shipments of AI chips to the world's No. 2 economy giving external competitors in China the opportunity to gain ground.

But in my opinion, Goldberg might be missing the two biggest threats to Nvidia stock.

First, we haven't seen a next-big-thing investment trend avoid an eventual early stage bubble-bursting event in more than 30 years. Investors consistently overestimate the early stage adoption and/or utility of game-changing technologies, which leads to lofty expectations not being met. Given that most businesses have yet to optimize their AI solutions or generate a positive return on their invested capital, there's a very high probability, based on historical precedent, of an AI bubble forming and bursting.

The second concern for Nvidia is that many of its top customers by net sales are internally developing AI-GPUs for their data centers. Even though these chips won't be sold externally, and they're not a match for Nvidia's hardware on a compute basis, they are notably cheaper and not backlogged. There's a good chance Nvidia may lose out on valuable data center real estate going forward, or at the very least contend with delayed upgrade cycles from its top customers.

Image source: Tesla.

Tesla: Implied downside of 95%

But when it comes to implied downside, electric-vehicle (EV) maker(TSLA 0.98%), which integrates AI solutions into its EVs, takes the cake.

Unlike Nvidia, Tesla has its fair share of skeptics. More than 10% of the 45 analysts covering the company in September rate it as the equivalent of an underperform or sell. This includes GLJ Research founder and analyst Gordon Johnson.

Earlier this year, Johnson reduced his price target on Tesla to just $19.05 per share. Keeping in mind that Tesla stock closed out the previous week at nearly $396 per share, it implies downside of roughly 95%.

Although Tesla has been profitable in each of the last five years, Johnson, a longtime Tesla bear, has a number of critiques to offer.

To begin with, he's been critical of Tesla's ancillary projects beyond EVs and energy generation and storage products. For instance, the hype surrounding Optimus robots has added to Tesla's valuation but provides minimal value to the Tesla brand and little hope of any near-term sales.

Gordon Johnson also pointed to Tesla's operating structure as a reason to be skeptical of its stock. While most "Magnificent Seven" members are thriving off of high-margin software sales, the bulk of Tesla's sales are coming from lower-margin hardware where it possesses less pricing power. The more than half-dozen price cuts Tesla has undertaken on its fleet of EVs over the last couple of years is evidence that its competitive edge in the EV space has shrunk.

Johnson has been critical of Tesla's mind-numbing valuation, as well. Whereas most automakers are valued at high-single-digit or low-double-digit price-to-earnings (P/E) multiples, Tesla shares are valued at an estimated 234 times forecast earnings per share in 2025. Worse yet, more than half of Tesla's pre-tax income has originated from non-innovative and/or unsustainable sources, such as automotive regulatory credits and interest income on its cash.

Though a decline to $19.05 per share seems highly unlikely, there are other reasons to believe Tesla stock can head notably lower. Specifically, CEO Elon Musk hasn't given investors much reason to be optimistic of late. Cybertruck sales have been a disappointment and the robotaxi launch event in Austin, Texas, was underwhelming.

These disappointments speak to a broader issue with governance at Tesla. Namely, Musk has a habit of overpromising and underdelivering. Whether it's promising Level 5 autonomy is "one year" away for 11 straight years or projecting 1 million robotaxis would be on public roads by 2020, Musk consistently misses the mark.

The issue from an investment standpoint is that these lofty promises are being baked into Tesla's stock. If these unfulfilled promises are backed out, it becomes very possible for Tesla shares to lose a significant percentage of their value.