The Market Just Remembered Cost of Capital Exists. Here's Who Survives It.
A quick note from Behind the Markets
Wall Street has a bad habit.
It treats cost of capital like it's background noise… right up until the day it becomes the main character.
This week you can feel that shift.
The market is finally asking the question it should've asked months ago: Who exactly is paying for all this "AI capex"?
1) AI is turning into a debt trade — and that changes who wins
You've heard the pitch: "AI is the future." True.
But how it's being financed is the part the Street buries in the footnotes.
Reuters reported that Nasdaq futures fell 2% Tuesday — with the Nasdaq 100 on pace to shed over $1 trillion in value — as investors worried about rising borrowing costs and debt-backed corporate spending on AI. SpaceX's new bond sale revived the fear that Big Tech is funding its buildout with borrowed money, and even Alphabet just raised a historic $84.75 billion to fund AI capex that now exceeds its operating cash flow.
That's not just a headline. It's the regime change.
When funding is cheap, everyone looks like a genius. When funding isn't cheap, only a few business models survive: companies with real free cash flow (not "adjusted"), suppliers that get paid regardless of end-demand (the toll booths), and firms that can pass through costs without demand collapsing.
Here's the contrarian angle. The AI "winners" might not be the companies building the smartest models. They might be the ones that get paid no matter which model wins — without having to borrow billions to build a data center.
There's a near-perfect example of that: the company that sells the software every chip designer needs, regardless of who comes out on top.
Company: Cadence Design Systems (SYM: CDNS)
Electronic design automation (EDA) — the software toll booth on virtually every advanced chip.
Cadence sits in a duopoly that makes the software used to design almost every advanced semiconductor on earth. It doesn't matter whether Nvidia, Broadcom, a hyperscaler's custom silicon, or a startup wins the AI chip race — they all design on Cadence (or Synopsys) tools. That's the definition of a toll booth "paid regardless of end-demand." The model is asset-light and recurring: roughly 19% revenue growth, a record $8 billion backlog, and strong free cash flow without needing to debt-fund data centers.
CDNS currently trades around $385, down from its 52-week high near $417. Now the honest risks, and they matter. It's expensive — roughly 50x forward earnings, so there's little margin for error if the AI narrative cools. It carries China exposure: U.S. EDA export controls to China are a live risk baked into guidance (and Beijing's new countermeasures only sharpen it). And it just funded a €2.7 billion acquisition partly with debt. In a broad chip selloff like Tuesday's, it falls with the group. This is a high-quality toll booth, not a bargain — treat the valuation with respect.
Bottom line: If AI capex is increasingly debt-financed, you don't just analyze products — you analyze funding. The market loves the story. The bond market decides the outcome.
2) The Fed's "rate hike" whispers are a tax on everything levered — especially small caps
The meme version of the market is "Fed cuts = stocks go up." Reality is messier.
Traders are now pricing a cumulative 50 basis points of Fed hikes by December — up from expectations of just one 25-basis-point move only two weeks ago — as they brace for a more hawkish stance under new Fed Chair Kevin Warsh.
Whether those hikes happen isn't the point. The point is the market is being forced to price the possibility again. That reprices leverage. It reprices duration. It reprices the entire "roll the debt, refinance later" economy.
And that's why small- and mid-caps get hit first. They're the part of the market that borrows at higher spreads, relies on local banks and private credit, and doesn't have infinite equity-multiple support.
The Street will pretend this is "rotation." It's not. It's a stress test. And the cleanest gauge of that stress test is the small-cap benchmark itself.
ETF: iShares Russell 2000 ETF (SYM: IWM)
The benchmark for U.S. small caps — and the clearest real-time read on refinancing stress
IWM here isn't a pound-the-table buy, but it's the pressure gauge to watch and respect. The Russell 2000 is stuffed with companies that need healthy local credit creation to survive — exactly what gets squeezed when the market sniffs sustained higher rates. Watch it relative to the S&P 500: when small caps lag badly while megacaps hold up, that's not healthy rotation, it's the financing bucket getting drained.
If you want small-cap exposure into this, the discipline is simple: go up in quality. Demand net cash or low leverage and real free cash flow — own the survivors, not the names one capital raise away from losing control. A frozen refinancing window doesn't care how cheap a stock looks.
Bottom line: If the market even sniffs sustained higher rates, small caps don't need a recession to underperform — they just need refinancing friction.
3) Oil "stabilized." Great. Now watch the second-order costs hit the real economy
This is where Wall Street's shallow thinking gets people wrecked.
Oil drops 3% one day and everyone declares victory. But the economy doesn't run on a one-day chart.
Reuters reported oil rebounded after a steep selloff — Brent back around $78, WTI near $74 — as investors awaited clearer signs that crude shipments through the Strait of Hormuz are truly normalizing, with the market skeptical given lingering U.S.–Iran distrust. (Notably, the U.S. Strategic Petroleum Reserve just fell to its lowest level since 1983 — the shock absorber is nearly empty.)
That's the setup for the real trade. Even if crude calms down, volatility creates lingering costs. War-risk insurance doesn't instantly reset. Shipping schedules don't magically normalize. And corporates don't roll back price increases they already pushed through.
So the "peace deal" trade can be a trap. Energy volatility can fade while Q2/Q3 earnings get clipped by higher logistics costs, fatter inventory buffers, and margin compression in low-pricing-power sectors. The damage shows up late.
Which is exactly why you want a business with the pricing power to pass those costs through — and the defensive profile to hold up when the AI trade is bleeding.
Company: Coca-Cola (SYM: KO)
Global beverage giant with extraordinary pricing power and a low-volatility, defensive profile
Coca-Cola is the textbook "pass-through" winner. When freight, packaging, and input costs rise, Coke raises prices — and people keep buying, because it's a $2 habit, not a discretionary splurge. That pricing power shows up in a roughly 33% EBITDA margin, and the stock's beta near 0.1 means it barely moves with the market's tech-driven swings. In a week where the Nasdaq is shedding a trillion dollars, that's a feature. Coca-Cola beat expectations and raised its 2026 outlook even amid the energy volatility and Middle East stress.
KO currently trades around $80, just below its all-time high of $84 set on June 11. The honest caveat: this is not a cheap, undiscovered name — it trades near full value, and in a roaring risk-on tape it'll lag the flashy growth names badly. You're buying durability and pricing power, not upside fireworks. But when oil's lagging taxes start clipping margins across the economy, the company that can simply raise its price is the one whose earnings hold.
Bottom line: Don't just watch oil. Watch the lagging taxes oil volatility imposes: freight, insurance, working capital, margins.
4) China's export controls are a warning shot at the supply chain — and rare earths are ground zero
Retail investors still treat "export controls" like a policy debate. Wrong. It's a pricing event.
China added 10 U.S. firms to its export-control list (restricting "dual-use" items) and banned government procurement from 46 U.S. firms, including subsidiaries of major defense companies — explicit retaliation for the Pentagon's blacklist expansion. The export list pointedly named U.S. rare-earth producers.
The investable takeaway isn't "WW3." It's that the world is re-learning that supply chains are weapons. Rare earths and magnets aren't sexy — but they sit inside missiles and drones, radar and optics, EVs and motors, and industrial automation. China processes roughly 90% of the world's supply. When Beijing restricts that, the value of non-Chinese production and processing capacity goes up.
Yesterday we flagged MP Materials, the U.S. flagship. Today, the complementary play is the largest producer outside China entirely.
Company: Lynas Rare Earths (SYM: LYSDY)
Largest rare-earth producer outside China; the only producer of separated heavy rare earths (dysprosium, terbium) outside China
Lynas owns the high-grade Mount Weld mine in Australia and processing facilities in Kalgoorlie and Malaysia — and as of late 2025, it's the only producer of separated heavy rare earths outside China. Those heavies (dysprosium, terbium) are the hardest for the West to source and the most exposed to Chinese export controls, which makes Lynas's capacity strategically scarce. The company is in talks with the U.S. Department of Defense on rare-earth price-floor support, and its quarterly revenue has been surging.
LYSDY currently trades around $12 (as a U.S. OTC ADR). Be clear-eyed about the risks here, because they're real: it's an over-the-counter ADR, so liquidity is thinner than a primary U.S. listing; it pays no dividend; it's a volatile commodity producer whose fortunes swing with NdPr prices; and like all rare-earth names, it can sell off hard when U.S.–China tensions ease. This is a speculative, position-size-it-carefully idea. But on the thesis that allied, non-China processing capacity is becoming strategically priceless, Lynas is the purest established play.
Bottom line: Export controls aren't politics — they're a cost shock. The winners are the firms with secure inputs and domestic processing capacity. The losers are the ones pretending 2019 supply chains still exist.
Before You Go
Wall Street is built to sell narratives.
Independent investors get paid by tracking the constraints.
Higher rates are a constraint. Energy volatility is a constraint. Supply chains are a constraint.
And constraints don't care about press conferences.
Found this helpful? Share it with others.
Written by Behind the Markets