Listen to the audio version of this article (generated by AI).
Picture two crews of laborers, swinging sledgehammers under a desert sun, racing toward each other from opposite ends of a continent never crossed by rail.
Skeptics back East called it overbuilding. The railroad men, they believed, were laying track faster than the country could possibly need it, betting fortunes on demand that hadn’t shown up yet.
Then the final spike went in at Promontory Summit. Within a decade the freight volume crossing that line made the doubters look foolish. The infrastructure came first. The economy that needed it came roaring in right behind.
I bring this up because something almost identical just happened in the AI trade, and if you only watched the headlines this week, you missed it.
The violent selloff that rattled AI stocks over the past couple weeks wasn’t the beginning of the end. It was a buying opportunity dressed up as a crisis, and I can show you exactly why.
A lot of investors are out there right now convinced the AI Boom just cracked. They watched the semiconductor sector drop six, seven, eight percent in a single session and concluded the music had stopped.
Here’s the proof: Micron (MU) reported the biggest beat-and-raise I have ever seen from that company, and the stock jumped 15% after hours. That single print reversed an entire selloff that had been building since Broadcom (AVGO)‘s earnings kicked off this wave of “extremely high expectations meets extremely high expectations” fatigue.
Beat-and-raises hadn’t been enough for the market lately. Micron came to the table and said: here’s a beat-and-raise for you, market, what do you think about that? The market said: bought.
Everybody is screaming “bubble” right now, but the data says this bubble hasn’t even started inflating yet.
In this week’s episode, I’m going to walk you through why the fundamentals never broke, why the technical setup says we’re nowhere near a real top, and the exact screener I’m running to find the highest-torque dip buys in this market right now.
Stick around to the end, because the full breakdown on this week’s Being Exponential goes even deeper:
The Fundamentals Never Moved
So let’s run through what’s actually happening underneath the noise.
SpaceX (SPCX) isn’t just launching rockets anymore. It’s renting out compute to Anthropic, to Google, and now to Reflection AI as well. That tells me SpaceX has probably sold out something like 90% of its existing compute capacity, which means it has to build more.
More Colossus data centers.
More orbital infrastructure.
We’ve spent the past couple years calling this a hyperscaler race among four titans — Microsoft (MSFT), Amazon (AMZN), Meta (META), and Alphabet (GOOGL). Now SpaceX is forcing its way into the picture, and most of the market still hasn’t priced that in.
If SpaceX and Tesla (TSLA) merge, as widely speculated, that’s even more capital flowing straight into the AI infrastructure buildout.
We’re talking hyperscaler spending of $700 billion to $800 billion this year. Add SpaceX into that mix and next year could push past $1 trillion. Run the math out to 2028, 2029, and 2030, and you get estimates climbing toward $1.1 trillion, $1.2 trillion, $1.3 trillion in those out years. That’s not a slowdown. That’s convergence.
Qualcomm (QCOM) just raised its long-term data center revenue target to more than $15 billion by 2029, with $5 billion targeted for 2027. That’s not a one-year story. That’s three straight years of structural growth baked into the guide.
And underneath all of it, there’s a shift happening from generative AI to agentic AI — a shift from training workloads to inferencing workloads, which is a shift from GPU-heavy demand to CPU-heavy demand. That’s exactly why Intel (INTC), Arm Holdings (ARM), and AMD (AMD) have been getting a wave of upgrades and rising estimates lately. The capex isn’t slowing. It’s redistributing.
The Technical Tell
Now here’s where the dot-com comparison actually matters, and where most people get it wrong. I ran an analysis on pure price action: how far the Nasdaq 100 has traded above its 200-day moving average during this AI Boom versus during the dot-com boom of the late 1990s.
From 1995 through most of 1998, tech stocks rallied steadily, trading only 10% to 15% above the 200-day average. It wasn’t until the parabolic phase — late 1998 through early 2000, after the Fed cut rates following the LTCM crisis — that the Nasdaq 100 went vertical, eventually trading more than 50% above trend by March 2000. That’s when the boom became a bust.
Right now, throughout this entire AI Boom, the Nasdaq 100 has averaged roughly 10% above its 200-day moving average. As of today, it’s sitting around 13%. We have never gone vertical. We have never even gotten close to that 50% danger zone.
History tells me every boom enters a go-vertical phase before it busts, which means that phase is still ahead of us, not behind us. The most money in the dot-com era was made in those final, overheated years — 1998, 1999, 2000. I think we’re still working toward that stretch, not past it.
What I’m Watching, and What I’m Buying
To be clear, I’m not waving off every risk.
The K-shaped economy — the gap between Wall Street’s fortunes and Main Street’s — is something to monitor closely, along with politics over the next two years.
But on the encouraging side, inflation is dropping fast, oil has fallen to around $70 a barrel, and the 10-year Treasury yield has eased to 4.4%, which should support borrowing and consumer spending in the months ahead.
So here’s my playbook…
Screen for stocks up more than 50% year to date or more than 100% over the past year. Narrow that list to names that have pulled back 10% to 20% in the past two or three weeks. Then require that they’re still holding above their 200-day moving average. That combination (high momentum, healthy pullback, intact uptrend) is where I’m hunting for high-torque buys right now, because the momentum that was is the momentum that will be.
We’re not in the ninth inning here. We’re in the sixth or seventh. Buy the dip!
P.S. For the full conversation, including more on the SpaceX-Tesla speculation and Luke’s live read of the chart, watch this week’s full episode of Being Exponential. And be sure to subscribe to Being Exponential on X (formerly Twitter) for more exclusive content.
