Wall Street Left This Company for Dead. Then It Became the Trade of the Decade.
Listen to the audio version of this article (generated by AI).
In the summer of 2019, when rows of unsold Model S and Model X sedans baked in the California sun, Wall Street thought Tesla (TSLA) was cooked.
The company was running out of road and Tesla’s bears knew it. Margins were shrinking, as was its cash flow, so more and more people lined up to short TSLA stock.
But what they could not see was the assembly line inside the building, where Tesla would produce a mass-market sedan called the Model 3. That model would eventually send Tesla stock up 10-fold, then 20-fold, then 30-fold, from where it traded that summer.
This time, we’re seeing that dynamic playing out across five stocks at once, and how they will pivot is what you need to understand before the market figures it out on its own.
The gap between what these companies are earning and what their stocks trade for has become one of the widest gaps this AI bull market has produced. And gaps this wide do not last. Every one of these five names looks, on the surface, like a company in some kind of trouble: an ad giant burning cash, a carmaker whose core business is bleeding margin, a turnaround story everyone has heard before, a neocloud getting whipsawed on a chart, and a rocket company sitting near a 52-week low.
Semiconductor stocks have fallen roughly 20% over the past month while semiconductor earnings estimates have climbed more than 10% over that same stretch, and the earnings we walked through on this week’s episode of Being Exponential make that gap almost impossible to justify with weaker fundamentals.
The stock market is currently pricing five of the best AI infrastructure businesses on Earth as though their best days already happened. Here is what the numbers actually say, stock by stock, and why I recommend leaning into the weakness rather than running from it:
Alphabet: The Cloud Business Nobody Priced In
Alphabet Inc. (GOOGL) just posted the kind of cloud number that used to take a decade to produce. Its cloud division grew at 25% to 35% in the years right after the pandemic. Last quarter, that same business put up 82% growth on more than $20 billion in quarterly revenue, an acceleration of 50 percentage points in about a year. That kind of jump on a business already running at a $100 billion annualized clip does not happen in the history of capitalism, because it has not happened before now. Operating margins in the segment climbed to 35% from 20% a year earlier, and the backlog sits at $514 billion. The market is fixated on the fact that free cash flow turned negative because Alphabet needs the debt markets to fund its buildout. I recommend a different fixation: a company this strong can handle debt, and I recommend buying the dip.
Tesla: The Fremont Lot, Version Two
Tesla’s current business genuinely struggles right now. Prices are getting cut, margins are getting crushed, and profits are falling, because the company is buying volume with lower prices rather than earning it with demand. That is today’s Fremont parking lot. But the bulls are not pricing the car business; they are pricing robotaxi, Optimus, and energy, the way an earlier generation of bulls priced the Model 3 ramp instead of the Model S slowdown. This quarter did not hand bulls the volume numbers or the timeline they wanted on robotaxi, so the stock stays dead money in the short term. Longer term, the growth drivers Elon Musk keeps teasing eventually show up in the income statement, the way the Model 3 eventually did, and that is why the stock remains a name for patient investors rather than short-term traders.
Intel: The Turnaround Finally Has Numbers
Intel Corp. (INTC) delivered $16.1 billion in quarterly revenue, up 25% year over year and nearly $2 billion above its own guidance, because its AI compute business finally scaled into real profitability. Data center AI revenue hit $6.3 billion, up 59%, with segment margins swinging to 39.5% from 16.1% a year earlier, a 23-point improvement in four quarters. Foundry losses narrowed from a 71.7% operating margin loss to 36.2%, and custom ASIC revenue nearly tripled. Intel is proving itself a legitimate foundry alternative to Taiwan Semiconductor Manufacturing Co. (TSM) and a real accelerator competitor to Nvidia Corp. (NVDA) and Advanced Micro Devices Inc. (AMD) in the same quarter. The turnaround story has arrived, and I recommend treating this print as the evidence, not the promise.
Nebius: The Dip-Buy Setup of the Week
Nebius Group N.V. (NBIS) benefits from two separate votes of confidence in a single week. Alphabet said on its earnings call that demand for compute continues to outstrip supply even after tripling its own capacity, and that it plans to buy more compute from neoclouds such as Nebius rather than build every gigawatt itself. Separately, Nvidia disclosed a 10% stake in Nebius, a signal from the company that understands AI supply bottlenecks better than anyone. The chart backs up the fundamentals: the stock bounced off its 100-day moving average on a nearly oversold reading, setting up for a bullish crossover. I recommend buying this dip.
SpaceX: A Line in the Sand at $100
SpaceX (SPCX) trades down roughly 50% from its highs since its recent public listing, and the valuation remains rich by any conventional measure. But the business behind it, bringing terrestrial compute online over the next two to three years and orbital compute online in the years after that, could add hundreds of billions of dollars in revenue on each front, building toward $1 trillion in revenue within five to seven years. A business of that scale merits a premium multiple, which is why I recommend punching long at $100 if the stock gets there, a level that marks the point where the risk-reward tilts sharply in your favor.
Where the Real AI Wealth Is Being Created
The lot outside Fremont looked empty of promise in the summer of 2019. But we all know how that turned out.
This week’s earnings tell you the same thing about five stocks the market currently treats like they have nothing left to prove. Watch this week’s full episode of Being Exponential for the complete breakdown of the technical setups, the bear-case rebuttals, and where the next leg of this divergence resolves.
But here is the part of the Tesla story most investors miss entirely: the biggest money was not made by the public shareholders who bought the dip in that Fremont parking lot. It was made years earlier, by the venture investors who backed Tesla while it was still a private company nobody outside Silicon Valley had heard of. Those early backers then made 10-fold, 50-fold, and even 100-fold their money before the stock ever traded on the Nasdaq.
That is exactly what I want to show you during my free 2026 AI Megadeal Event on Thursday, July 30, at 1 p.m. Eastern.
I recommend zooming out from the five public stocks we just walked through, because the biggest fortunes in this AI boom are forming one step earlier than a single share of GOOGL, TSLA, INTC, NBIS, or SPCX. I’ll explain why I believe AI is creating a new generation of buyout opportunities, walk through the exact framework I use to identify them, and share the one company I believe best represents this trend today.
This event is free to attend, but you must reserve your seat to get an invitation.
The lot outside Fremont taught an entire generation of public investors to look past a bad quarter and toward the ramp coming next. The engineers who backed Tesla before it was popular knew this, and it is why their returns dwarfed everyone else’s.
It’s my goal to help you put that same playbook to work, before the rest of Wall Street catches on.
