Listen to the audio version of this article (generated by AI).

Back in 2015, a certain kind of stock looked irresistible…

Tilly’s (TLYS) was trading at a single-digit P/E. So was Abercrombie & Fitch (ANF). American Eagle (AEO). The Buckle (BKE).

A whole shelf of mall-based apparel names had been bludgeoned by Amazon (AMZN), and bargain hunters were lining up to scoop them out of the discount bin.

The pitch wrote itself. Real revenue. Real stores. Real dividends. And a multiple that left almost nothing to the downside.

And guess what? For a few of them, it worked out. Abercrombie (ANF) had its day in the sun a decade later.

But for most of that cohort, the cheap multiple did not mark a bottom. It marked the slow grinding middle of an existential transition the income statement had not caught up to yet. The customers were already gone. The leases were the last to know.

Now, what I’ve been calling SaaSmageddon is not a forecast anymore. It’s happening in real time. Platforms like Claude, ChatGPT, and Gemini are absorbing the workflows that used to live inside dedicated SaaS applications, and ServiceNow (NOW), Salesforce (CRM), Adobe (ADBE), Intuit (INTU).. the once-untouchable enterprise stalwarts… have all been touched. Hard. Now they’re bouncing, and the headlines are framing it as a recovery.

I’m not so sure.

This week’s episode of Being Exponential walks through five stocks that draw the line between the AI Boom’s beneficiaries and its casualties. And we’ll get into why the most dangerous bargains in 2026 might be the ones that look cheapest on a screener:

The Picks-and-Shovels Side of the Boom

Lumen Technologies (LUMN) is the first name on my list, and it captures the picks-and-shovels logic cleanly. On paper, Lumen is an old telecom. In practice, it’s becoming a fiber backbone for the AI infrastructure buildout. The company has roughly $13 billion in signed hyperscaler contracts, with Microsoft (MSFT) among them, and is laying down a major new data center connectivity pipeline in the Northwest, on a route I believe runs from Seattle toward Denver.

The historical headwind here has been balance sheet risk… a heavy debt load with looming maturities. That overhang is clearing. Lumen has been paying down debt and pushing maturities out to 2029, which gives the revenue ramp from those new contracts time to materialize before any refinancing pressure returns.

Technically, the stock just put in a clean bounce off its 200-day moving average. I think it takes out $12 soon. And from there, it goes meaningfully higher.

CoreWeave (CRWV) is the larger and more aggressive bet on the same trend. The NeoCloud thesis — that compute capacity will remain structurally undersupplied for years — has gained fresh momentum, and CoreWeave is the most levered name in the category.

It’s also the biggest.

Its OpenAI exposure, recently treated as a headwind during OpenAI’s perceived market-share losses to Anthropic, looks more like a tailwind again now that ChatGPT 5.5 has put some pep back in OpenAI’s step and the IPO calendar is coming into focus. CoreWeave is riskier than most AI infrastructure plays. It also has more torque.

Redwire (RDW) is my favorite of the five. The pitch is straightforward…

Of the four space stocks I’ve championed in this franchise, three of them — AST SpaceMobile (ASTS), Rocket Lab (RKLB), and Planet Labs (PL) — have gone parabolic, with our recommendations on all three up more than 1,000%. Redwire (RDW) has not. Yet.

The reason, in my view, is timing. Redwire’s specialty is outer-space solar panels… the Rosa solar arrays that power the International Space Station.

Demand for outer-space solar has been steady but not explosive. That changes the moment Elon Musk takes the $75 billion he expects to raise in the SpaceX IPO and routes a meaningful chunk of it into orbital compute. Data centers in space need outer-space solar. Outer-space solar means Redwire. I see a path to $50-plus over the next 6 to 12 months as the SpaceX IPO catalyst arrives.

The Other Side of the Line

ServiceNow (NOW) is the cautionary tale. After getting destroyed during the spring’s AI-disruption sell-off, ServiceNow has led the bounce-back in enterprise software. Headlines have noted insider buying, including reports that Donald Trump took a position. None of it changes my read.

My call? This is the dead-cat bounce, not the recovery. AI-native platforms are going to absorb 80% to 90% of what NOW does. The remaining 10% to 20% will face heavy pricing pressure, which means lower revenues, lower margins, and lower profits all compounding into a multiple that has nowhere good to go. The single-digit P/E thesis is the same one that lured investors into Tilly’s a decade ago. And most of those investors did not end up where they thought they would.

POET Technologies (POET) is the speculative wild card, and I’m passing on it. The optics buildout is real, and POET could absolutely run if it lands a hyperscaler order. But it hasn’t landed one yet. Its proof of concept is a small win with Lumalens — a name no one is going to recognize — and it recently lost a Marvell (MRVL) opportunity after a credibility-damaging CFO disclosure. The stock oscillates wildly between $10 and $20 with no fundamental anchor. In the small-cap optics bucket, I prefer Applied Optoelectronics (AAOI), which has the hyperscaler orders POET is still trying to win.

The Bottom Line

The through-line across all five names is the same question, asked five different ways. Which side of the AI Boom does this business sit on?

LUMN, CRWV, and RDW are infrastructure plays… picks-and-shovels exposure that benefits as the buildout accelerates. NOW is on the disrupted side. POET is the speculative ticket that needs proof of concept before it earns a seat. The framework is more important than any single ticker, because the same logic that made the 2010s retail trade a value trap is the logic now running through software.

The cheap multiple is the lure. The structural disruption is the trap. Knowing the difference is the whole job in the back half of 2026.

Tap in to this week’s episode of Being Exponential for the full breakdown — including the technical setup on Lumen, the SpaceX IPO timing on Redwire, and why I think ServiceNow’s bounce ends the same way the 2010s retail bounces did. Also, be sure to subscribe to Being Exponential on X (formerly Twitter) for more exclusive content.