The 5 AI Infrastructure Stocks Powering the Next Leg of the AI Buildout

Follow Luke on X

Earlier this month, in the Synod Hall at the Vatican, Pope Leo XIV did something no pope has ever done. He stood at a dais beside Anthropic’s Christopher Olah and personally presented an encyclical on AI… He titled it Magnifica Humanitas.

This coincided with the 135th anniversary of Rerum Novarum, the 1891 encyclical that took on the robber barons of the Industrial Revolution. That was no coincidence… the symbolism was deliberate. In 1891, Leo XIII took on the steel and railroad titans of his age. This month, Leo XIV is taking on the tech moguls of ours.

His charge was simple: The men building artificial intelligence have wrapped themselves in the vestments of priests… hoodies and sneakers in place of cassocks, TED talks in place of sermons, a “mission to change the world” lifted straight out of the Gospel… and have appointed themselves the architects of a new era for humanity.

The Pope’s response was that they have built recklessly, with no moral compass, in service of what he calls a “culture of power.” A few hours after the encyclical dropped, a Silicon Valley venture capitalist tweeted, “Bad take from the Pope.”

That exchange, in miniature, tells you everything about the moment we’re living through.

The AI buildout is now large enough, fast enough, and consequential enough that it has summoned the Vatican into the conversation. Trillions of dollars are being routed into compute, chips, satellites, and data centers… and almost none of it is slowing down for a theological debate.

Which leaves investors with a very different question than the one the Pope is asking. He wants to know where we’re going. We need to know who’s building the road…

The timing is everything here. The Gold Rush, for example, made fortunes for those who got in early. Those who showed up late were left holding the bag. And the AI rush is shifting beneath our feet right now.

The first wave was about training. The next wave is about inferencing, orchestration, and the strange new geography of compute itself, including data centers in space.

If you’re still pricing these stocks like the old cycle, you’re going to miss the new one. And the names winning this leg aren’t the obvious ones.

To get a sense of who’s building the road to the AI Boom, let’s run through five of the top AI infrastructure stocks. Just click the video below to watch this week’s episode of Being Exponential, where I break this all down:

AI Infrastructure Stocks: Micron (MU) – The Trillion-Dollar Memory Thesis

Micron (MU) just joined the trillion-dollar club, and Wall Street is finally catching up to what we’ve been saying for months. The memory cycle has historically run two to five years before supply swamps demand. This time is different. Hyperscaler capex is climbing into 2030. That means HBM, DRAM, NAND, and AI storage stay red-hot far longer than the analyst consensus implies. I see a pathway to roughly $200 billion in EBITDA over the next four to five years, which on a conservative 10x multiple gets you to a $2 trillion valuation. The stock is overbought near term, so I’d let it pull back to the $700 to $800 zone before buying. If you already own it, stay in your seat.

Qualcomm (QCOM) – The Sleeper Chip Stock Wakes Up

The next domino has tipped. Qualcomm (QCOM) is no longer just an edge AI story or a smartphone story. With the launch of its AI 200 and AI 250 platforms, it’s a data center inferencing story — and it just landed ByteDance, one of the largest inferencing users on the planet, as a customer for TikTok workloads. Estimates are about to get revised higher, and the stock still trades at only 19 times forward earnings while peers sit at 30 to 50 times. The chart just reclaimed the $230 to $240 breakout zone, which is now a technical floor. Buyable right here.

Planet Labs (PL) – The Data Layer for the Space Economy

Planet Labs (PL) runs a 50% to 60% gross margin earth observation business growing 30%-plus per year. That’s the floor. The ceiling is Project Suncatcher — Google’s plan to launch AI data centers into orbit in 2027, with Planet Labs already at the table. Orbital compute is one of the biggest growth verticals of the next decade, and PL has a real seat. I see a pathway to $300 million to $500 million in EBITDA within five years and a market cap of $30 billion to $50 billion. Don’t chase here. Buy the dip toward $40.

IonQ (IONQ) – The Technical Leader in Quantum

The U.S. government just announced a $2 billion shopping spree across nine quantum names — and IonQ (IONQ), the undisputed commercial and technical leader, wasn’t on the list. Strange. I suspect a bigger tranche is coming. Either way, IonQ ripped 12% on the day, the 200-day moving average is sloping higher for the first time since early 2026, and the chart is in full rebound mode. Forget current revenue. The market cap is $23 billion against a long-term TAM measured in trillions. Buyable here.

Arm Holdings (ARM) – From GPU World to CPU World

The buildout is shifting from training to inferencing, and inferencing needs an orchestration layer — CPUs. Arm (ARM) just launched its new AGI CPU and counts every hyperscaler as a customer. Revenue is set to compound at 30%-plus into 2030, with EBITDA margins scaling from 45% to 50%. Yes, it trades at 116 times forward EBITDA. For a 40% to 50% EBITDA grower at the epicenter of the inferencing shift, that’s defensible. I’d wait for a pullback toward $240 to $250 to add.

One Warning Before You Go

Five AI infrastructure stocks I like in one week is unusual. It also lines up with price action that’s starting to rhyme with late 2021 (speculative tech meltups while the Dow slips). This is not a top, nor is it even a yellow flag. The Pope can debate the soul of the machine, and arguably should. It’s our job is to figure out where the money is going while the debate plays out. The party will continue for now. Just keep an eye on the exits.

For the full breakdown on all five names… including the chart levels, valuation math, and where I’d actually be a buyer versus where I’d wait… watch this week’s full episode of Being Exponential. Also, be sure to subscribe to Being Exponential on X (formerly Twitter) for more exclusive content.

{
“@context”: “https://schema.org”,
“@type”: “FAQPage”,
“mainEntity”: [
{
“@type”: “Question”,
“name”: “What are the top AI infrastructure stocks to watch right now?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Luke Lango highlights five names anchoring the next phase of the AI buildout: Micron (MU) for memory, Qualcomm (QCOM) for data-center inferencing, Planet Labs (PL) for orbital compute, IonQ (IONQ) for quantum, and Arm Holdings (ARM) for the CPU orchestration layer powering the shift from training to inferencing.”
}
},
{
“@type”: “Question”,
“name”: “Why is the AI investing cycle shifting from training to inferencing?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “The first wave of the AI buildout focused on training large models, which favored a narrow set of chip and compute names. The next wave centers on inferencing, orchestration, and new geographies of compute — including data centers in space — which rewards a different set of infrastructure stocks tied to memory, CPUs, satellite imagery, and quantum computing.”
}
},
{
“@type”: “Question”,
“name”: “Is Micron (MU) still a buy after joining the trillion-dollar club?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Micron’s memory cycle is being extended by hyperscaler capex climbing into 2030, with a pathway to roughly $200 billion in EBITDA over four to five years and a potential $2 trillion valuation on a conservative 10x multiple. The stock is overbought near term, so Luke recommends waiting for a pullback into the $700 to $800 zone before buying. Existing holders should stay in their seat.”
}
},
{
“@type”: “Question”,
“name”: “What makes Qualcomm (QCOM) a sleeper AI stock?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Qualcomm has evolved beyond its smartphone and edge AI roots into a data-center inferencing story with the launch of its AI 200 and AI 250 platforms, landing ByteDance as a major customer for TikTok workloads. It trades at only 19 times forward earnings while peers sit at 30 to 50 times, and the chart recently reclaimed the $230 to $240 breakout zone, making it buyable at current levels.”
}
},
{
“@type”: “Question”,
“name”: “How does Planet Labs (PL) fit into the AI infrastructure trade?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Planet Labs runs a 50% to 60% gross margin earth observation business growing more than 30% per year, with additional upside from Google’s Project Suncatcher — a plan to launch AI data centers into orbit in 2027 with Planet Labs already at the table. Luke sees a pathway to $300 million to $500 million in EBITDA within five years and a market cap of $30 billion to $50 billion, with the best entry point on a dip toward $40.”
}
},
{
“@type”: “Question”,
“name”: “Why does Arm Holdings (ARM) matter for the inferencing era?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “The AI buildout is shifting from training to inferencing, and inferencing requires a CPU-based orchestration layer. Arm just launched its new AGI CPU with every hyperscaler as a customer, and revenue is set to compound at more than 30% per year into 2030 with EBITDA margins scaling from 45% to 50%. The stock trades at 116 times forward EBITDA, which Luke considers defensible for a high-growth name at the epicenter of the inferencing shift, though he’d wait for a pullback to $240 to $250 to add.”
}
}
]
}