Listen to the audio version of this article (generated by AI).
Picture two businesses at closing time.
At one, a clerk replaces yesterday’s sale sign with a deeper discount. Yet, its customers have mostly gone elsewhere, leaving behind full shelves.
At the other, the business’ phone keeps ringing long after the lights go out. Customers want deliveries sooner, and the owner has orders to fill, machines to install, and too few hours in the day.
On a bad afternoon in the stock market, both businesses can lose value.
But you would not want to buy them for the same reason.
That’s at the heart of our latest episode of Being Exponential with Luke Lango. Because beneath the familiar spectacle of falling indices and rising Treasury yields, companies face very different problems: Some are fighting for customers while others are racing to supply them.
Consider Qualcomm (QCOM). Its new Amazon (AMZN) partnership gives its AI ambitions a concrete foothold in custom data-center chips. Broadcom (AVGO) projects AI semiconductor revenue of $115 billion in fiscal 2027, followed by $230 billion in fiscal 2028. Marvell Technology (MRVL) raises its revenue outlook for both years.
Those forecasts still require the companies to execute well. But they describe businesses preparing to deliver substantially more… even when their share prices suggest investors are preparing for less.
Then there is Lululemon Athletica (LULU), where second-quarter comparable sales fall 9%. A cheaper stock does little to resolve the question of how the company can win back shoppers who have gone elsewhere?
And Howmet Aerospace (HWM) presents the more intriguing puzzle. When customers pursue their own turbine-component manufacturing, are they threatening an established supplier… or revealing just how desperately the market needs its products?
The opportunity begins when you understand which problem you are buying.
Watch the full episode below for my five-stock breakdown, plus the earnings trends, valuations, and chart signals that separate a promising rebound from another markdown on yesterday’s merchandise:
Custom Chips Give the AI Boom More Ways to Grow
For much of the AI boom, the investing conversation has revolved around Nvidia Corp. (NVDA).
That makes sense, but the next phase gives investors more companies to consider because the biggest AI customers want more control over their computing costs.
Enter custom silicon: chips designed around particular workloads.
A broadly programmable processor offers flexibility. A custom chip trades some flexibility for the potential to perform selected tasks more efficiently. When you process enormous volumes of AI requests, improvements in power consumption and cost can become meaningful competitive advantages.
That does not mean Nvidia stops winning. It means a growing market can support several approaches.
And those approaches require chip designers, networking technology, and specialized engineering.
That is where the opportunity becomes much bigger.
Qualcomm Gets a More Immediate Growth Catalyst
Qualcomm Inc. (QCOM) already has an appealing long-term AI story: Snapdragon processors powering connected devices, with opportunities across smart glasses, computers, vehicles, and robotics.
The challenge is timing. A compelling product category does not automatically produce a large revenue stream next quarter.
Now Qualcomm has another growth engine.
Its new collaboration with Amazon.com Inc. (AMZN) covers multiple generations of custom silicon for AI inference, along with optical connectivity for data centers. Inference is what happens when a trained model responds to a request – answering your question or generating computer code.
For Qualcomm, the significance extends beyond adding a recognizable customer. Amazon’s participation validates its push into the infrastructure where AI spending is already happening.
The investment case now has two timelines: data-center demand and the longer rollout of AI across physical devices.
I recommend Qualcomm in the episode because that combination creates a stronger growth setup. The next evidence to watch is how quickly the partnership translates into revenue and profit.
Broadcom and Marvell Supply the Expansion
Broadcom Inc. (AVGO) and Marvell Technology Inc. (MRVL) offer another way to invest in custom silicon.
They provide technology and engineering that customers need to build specialized chips and move data through increasingly complex computing systems.
Broadcom projects AI semiconductor revenue of about $115 billion in fiscal 2027 and $230 billion in fiscal 2028. Those are forecasts, but they illustrate the scale of the expansion management anticipates.
Marvell also raised its outlook to approximately $12 billion in total revenue for fiscal 2027 and $18 billion for fiscal 2028, up from $11.5 billion and $16.5 billion, respectively.
A stock can decline while its expected earnings improve.
That combination deserves attention because you may be paying less for a stronger business outlook. It does not guarantee a rebound; you still need to understand why sellers are leaving.
In the episode, I recommend Broadcom and Marvell and compare their valuations, earnings estimates, and rebound patterns. Marvell carries a higher forward earnings multiple in that comparison, making the durability of its growth particularly important.
I also would not build the entire thesis around which AI developer has the best model this month. These companies are still early in a long race. Leadership changes.
Owning suppliers with exposure to several competitors gives you a way to participate without identifying the eventual winner.
Howmet: A Competitive Threat Requires a Timeline
That brings us back to turbine components and Howmet Aerospace Inc. (HWM).
Customers that secure their own casting capacity could eventually reduce their dependence on outside suppliers. But a plan to manufacture complex components does not instantly become qualified, reliable production at scale.
Investors need to separate the announcement from its financial consequences. How much capacity gets built? When does it arrive? How much business is actually exposed?
My interpretation is that efforts to secure casting capacity also underscore how valuable that capacity has become. Strong demand can support existing suppliers while new competitors work toward production.
Still, I am not recommending that you buy every downtick. In the episode, I explain why I want evidence that Howmet is finding support before buying a rebound.
A strong business thesis and a disciplined entry can coexist.
Lululemon Shows Why a Falling Price Is Not Enough
Lululemon Athletica Inc. (LULU) presents a different problem.
Second-quarter revenue fell 4%, while comparable sales declined 9%, or 10% excluding currency effects. Comparable sales measure performance across established stores and e-commerce, making them a useful check on underlying demand.
People still buy workout clothes, but my concern is where those purchases go.
Smaller brands and the creators promoting them can redirect spending away from established names. A customer can enjoy Lululemon’s shorts while an investor rejects its stock.
That is my position here. I recommend avoiding Lululemon until there is evidence of a turnaround.
A lower share price alone is not that evidence.
For investors, the task is to become a buyer of improving businesses at attractive prices. That takes more than spotting a stock far below its high.
Most importantly, you must follow the companies supplying what the next phase of AI requires.
Qualcomm, Broadcom, and Marvell illustrate the opportunity inside data centers. But what happens when that intelligence moves into robots, autonomous vehicles, and machines that perform physical work?
That is where my Vertical AI Masterplan begins.
At last night’s event, Louis Navellier, Eric Fry, and myself connected four layers of that opportunity – Data, Computing power, Connectivity, and Robotics – and explained why we are targeting the specialized suppliers behind Musk’s ambitions.
I also introduced a brand-new report, Elon’s Chosen Ones, which details supplier recommendations, tickers, and buy-up-to prices, alongside a blueprint showing where those companies fit.
If you missed last night’s presentation, make this your next stop. The event window is closing.
Go now to review our thesis and the special invitation to access his research before the event closes. Give yourself time to understand the recommendations and decide whether they belong in your portfolio.
Click below to access my Vertical AI Event and offer while they are still available.
Access my Vertical AI Event Now.
5 Stocks Our Experts Predict Could Double In the Next Year
By submitting your email, you'll also get a free pivot & flow membership. A free daily market overview. You can unsubscribe at any time.
