Luke lango wrote this awhile ago and was spot on , I will be reading his stuff in the future 

In 1945, an engineer named Percy Spencer was tinkering with a radar magnetron, a device built to spot enemy aircraft, when the chocolate bar in his pocket turned to mush. Most people would have cursed the dry-cleaning bill. Spencer saw a fortune. Within a few years, that same wartime component was humming on kitchen counters as the microwave oven.

Here’s the thing about markets: the biggest money often gets made when a technology built for one war gets drafted into another. The hardware was already there. It just needed a new battlefield. I’m watching that exact pattern play out across the AI infrastructure buildout right now, and it’s minting winners faster than most investors can update their watchlists.

The catch? For every honest reinvention story that runs, there’s a hype trade wearing the same costume. Telling them apart is the whole game. This week on Being Exponential, we walked through five stocks that show you both sides of that coin. Four I really like. One I wouldn’t touch with your money, let alone mine.

There’s Marvell Technology Inc. (MRVL), which we see carving a genuine path toward a $1 trillion valuation as custom silicon and connectivity become the twin bottlenecks of the AI buildout. There’s Dell Technologies Inc. (DELL), still filed under boring PC maker, whose AI server revenue grew 757% year over year on $24.4 billion in fresh orders and a $51.3 billion backlog. There’s Fluence Energy Inc. (FLNC), a left-for-dead battery story now growing 48% as its storage technology finds new life inside power-starved data centers. And there’s Redcat Holdings Inc. (RCAT), a tiny drone maker with revenue climbing 274% as Washington warms to the dronification of modern warfare.

The fifth name, a fund trading as VCX, is the one to leave alone, even though it remains the only public doorway to one of the most coveted private companies in the world. The full podcast explains why the math refuses to work, and the single price that would change his answer.

Watch the latest episode of Being Exponential With Luke Lango below:

Let’s start with Marvell stock…

Nvidia Corp. (NVDA) CEO Jensen Huang name-checked it on stage, and the stock jumped 20% to 30% in a day. Marvell sits in two of the most important bottlenecks in all of AI: custom silicon and connectivity. Nvidia’s GPUs own training. For inference, the day-to-day running of these models, custom chips win on economics.

Marvell and Broadcom Inc. (AVGO) are the two big dogs building that silicon alongside the hyperscalers. And all those AI clusters have to talk to each other in real time, which is where Marvell’s connectivity gear comes in.

If they hold, the company’s 42% revenue growth can hold too. Push that forward and you get $50 billion in revenue, $25 billion in earnings, and at a 40 times multiple, a $1 trillion valuation, the fourth chip stock to join that club next to Nvidia, Broadcom, and Micron Technology Inc. (MU). I love it long term, but it’s overbought now, so I’d wait for a pullback toward $200.

Then there’s Fluence Energy stock, the clearest reinvention story of the bunch. This was a $40 stock that collapsed into the single digits when its grid-storage dream stalled out. Now it’s roaring back, because the batteries it built for the grid are exactly what power-starved data centers are desperate for.

We’ve seen this movie before. Bloom Energy Corp. (BE) went from a mid-teens stock to north of $300 by pointing its fuel cells at data centers. The money everyone wrote off as wasted spend on electric vehicles and renewables is suddenly the answer to the AI power gap.

Fluence’s data center pipeline jumped 30% sequentially, it’s signed supply agreements with two major hyperscalers, and it’s guiding toward an estimated 48% revenue growth this year while trading at just 1.4 times revenue. I like it from $30 to $40 near term.

Speaking of reinvention, look at Dell stock. Everybody still files it under “boring PC company.” The golden goose is the server business. When you build an AI data center, the Nvidia chips have to go into something, and increasingly they go into Dell’s full-rack solutions.

Last quarter, Dell’s AI-optimized server revenue grew 757% year over year. Total revenue climbed roughly 88%. The company booked $24.4 billion in AI orders, exited with a $51.3 billion backlog, and raised its full-year AI server target toward $60 billion. Some of that came from rival Super Micro Computer Inc. (SMCI), which was beating Nvidia in 2023 and 2024 before accounting problems and federal investigations made it untouchable for many hyperscalers. Those orders found a sticky new home at Dell. The stock’s run hot, so I’d expect a retreat toward the $300 area. Dell is becoming the premier AI server play in the market.

Now the one to avoid. There’s a Fundrise fund trading as VCX that hands ordinary investors a public doorway to private AI darlings, with roughly 20% in Anthropic plus stakes in OpenAI, SpaceX, Anduril, Databricks, and Ramp. It’s the only public wrapper on Anthropic, and I get the appeal.

You can buy OpenAI and SpaceX exposure elsewhere. Anthropic, you cannot. Here’s the problem, and it’s just arithmetic. VCX trades at roughly 10 times its net asset value. Run the optimistic case where Anthropic, OpenAI, and SpaceX all balloon into $10 trillion companies, and the fund’s underlying value still wouldn’t clear $100 a share. It’s been trading around double that, propped up by a liquidity-starved hype trade.

The day Anthropic IPOs, that premium evaporates, and I could see VCX falling toward $50 or $60. That’s where it gets interesting. Not today. And never buy options on this thing. It has swung from $20 to $500 and back to the low hundreds in months. That’s how you lose your shirt.

Finally, Redcat Holdings Inc., a high-torque pure play on the dronification of modern warfare. We’ve watched it in Ukraine and the Middle East, where cheap, consumable drones now dominate the fighting. Redcat has graduated from a niche hopeful to a legitimate U.S. defense supplier, with a flagship Black Widow drone tied to a U.S. Army reconnaissance program.

And Washington is reportedly weighing direct stakes in drone makers. That’s a convergence, and I love a convergence: geopolitical, technological, and company-specific tailwinds all arriving at once. Revenue is expected to surge 274% to $152 million this year.

The enterprise value sits around $2 billion, so you’re paying roughly 10 times next year’s sales, but defense primes like Lockheed Martin and Northrop Grumman command rich multiples. If Redcat executes, this could grow into a $20 billion to $40 billion company. The execution risk is enormous.