The AI Boom’s Next Phase Is the Hangover - 7/1

So Many Narratives, So Little Margin of Safety. Here's What Actually Decides the Next Move.    

A quick note from Behind the Markets

Wall Street is addicted to one trick. Take a real risk… and turn it into a comforting narrative.

"AI is unstoppable." "Oil is calming down." "The dollar is fine."

This morning, it's better to be focused on what actually moves portfolios: capital flows, funding costs, and the stuff that shows up in earnings with a lag — right after the headlines fade.

1) The BIS just threw cold water on the AI party — and that's bullish for select names

Here's how bubbles actually blow up. Not in one dramatic crash. They blow up when the smartest people in the room finally say: "Okay… but where are the returns?"

That's the signal the market got when the Bank for International Settlements — the central bankers' central bank — used its annual report to warn about over-investment risk and the "hangover" when AI spending crests. Worth noting: U.S. chipmaker stocks staged a record 75% rally in Q2, and the five biggest hyperscalers' AI capex is approaching $1 trillion this year. 

The key line for independent investors: the AI boom can be real… and still create terrible stock returns for the "wrong" owners at the "wrong" price. The regime shift is from growth at any cost to ROI with a stopwatch.

What changes in that regime: CFOs stop green-lighting "nice to have" projects. Buyers consolidate vendors. Enterprises demand measurable productivity. And the market stops paying 30–50x sales for companies that need the next funding window.

So the BIS warning is a shot across the bow — but it's also a roadmap. If you're hunting underfollowed opportunities, think second-order: compliance and governance layers, vertical software with hard ROI (claims, audit, fraud, security), and infrastructure "toll booths" that get paid regardless of which model wins. One name sits squarely in that third bucket — and the market has handed it to you on sale.

Company: Verisk Analytics (SYM: VRSK)
Insurance data and analytics toll booth; recurring subscriptions for risk, fraud, and claims decisions

Verisk is the data backbone the insurance industry runs on — underwriting, fraud detection, claims estimating, catastrophe modeling. Insurers can't price or pay a claim without this kind of data, which makes it a textbook toll booth: recurring, subscription-based, with 56% adjusted EBITDA margins and 7% subscription growth. It's asset-light and doesn't need a funding window to survive — the opposite of the cash-burning AI names the BIS is warning about. And here's the contrarian kicker: VRSK trades around $179, down from a 52-week high near $315 — a brutal de-rating that's handed you a wide-moat compounder at a discount.

Be clear-eyed about why it's cheap: free cash flow fell 17% year-over-year, debt has risen, auto-insurance demand is soft, near-term AI investments are pressuring margins, and it was just dropped from the Nasdaq-100. Consensus is a cautious "Hold," with targets in the $220–260 range. This is "buy the de-rated toll booth and be patient," not a quick pop. But on the regime shift the whole issue is about, Verisk is exactly the kind of recurring-revenue, pricing-power business that wins when buyers start demanding ROI.

Bottom line: AI isn't going away. But the easy multiple expansion phase is. The winners will be the businesses with pricing power, recurring revenue, and the ability to survive when buyers start asking hard questions.

2) Oil's "peace dip" is what complacency looks like — and it's never free

Markets love to act like geopolitics is a toggle switch. War on. War off. Prices normalize. In the real world, supply chains don't heal on command.

Oil fell as investors focused on potential U.S.–Iran talks in Doha, with Brent back near $73 and WTI near $71 — roughly pre-war levels, and a Q2 drop that ranks as crude's worst quarter since 2020. But the Strait of Hormuz is still mined, hundreds of vessels remain stranded, and strategists warn the snap-back won't be clean.

Here's the contrarian angle: lower oil after a geopolitical flare-up isn't "disinflation." It's a market pricing relief — and relief trades overshoot. Even if oil holds down here, the second-order effects are still in motion: insurers keep charging war-risk premiums, shippers build buffers, corporates hedge late, and governments convince themselves the energy problem is solved.

The setup isn't "trade crude futures." It's the spread between companies that can pass costs through (pricing power) and companies that eat costs because their customers are tapped out. That's how you get weird earnings prints later — oil looks calm, then a quarter later, margin pressure hits transport, chemicals, packaging, and industrials nobody modeled.

So you want the purest pricing-power machine you can find — ideally one already proving it can pass through brutal input-cost inflation.

Company: Hershey (SYM: HSY)
Dominant U.S. confectioner and snacker; enormous pricing power and a nearly all-domestic revenue base

Hershey is a live demonstration of pricing power. It's spent two years passing through record cocoa-cost inflation — raising prices, shrinking bars, leaning on brand loyalty — and still beat Q1 2026 estimates by ~15% (EPS $2.35), with North America salty snacks up 26%. It guides to 30–35% EPS growth this year as cocoa costs are expected to ease. Roughly 90% of its revenue is domestic, so it's the rare staple that's both a pricing-power story and insulated from the dollar (its beta is near zero — it barely moves with the market).

HSY trades around $181, down from a 52-week high near $239, yielding ~3.1% with a 56-year dividend-increase streak. The honest caveats: at ~33x earnings it isn't cheap for a slow grower, cocoa-cost volatility cuts both ways, confection volumes have been soft, and several analysts have trimmed targets on market-share worries. You're buying durable pricing power, not a bargain. But when oil's lagging taxes start squeezing margins across the economy, the company that can simply raise its price — and has proven it — is the one whose earnings hold.

Bottom line: The market is treating oil like a headline. You should treat it like a lagging profit-and-loss weapon.

3) The "strong dollar" problem is coming back — and it'll show up as 'surprise' guidance cuts

Retail investors watch stocks. Pros watch the dollar. Because the dollar is the quiet tax on multinational revenue (translation), emerging-market demand, commodity-linked growth stories, and anything funded in dollars.

And the warnings are stacking up. HSBC flagged an "explosive" U.S. dollar rally as one of the biggest potential "pain trades" of the second half — driven by sticky inflation and a more hawkish Fed — and noted hedge funds are already positioned the most long-dollar in 16 months.

Wall Street always notices currency at the worst time. First they tell you "FX is a headwind." Then the quarter ends. Then guidance gets cut. Then they pretend it was unforeseeable. If the dollar strengthens into earnings season, you'll see a separation: domestic revenue plus pricing power quietly wins, while global story stocks miss and blame "FX."

The trade isn't to chase the biggest multinationals on a weak day. It's to own U.S.-centric cash generators that don't have to translate shrinking euros, yen, and yuan back into dollars.

Company: Cintas (SYM: CTAS)
Uniform rental and facility services leader; ~entirely North American revenue, recurring contracts, pricing power

Cintas is about as dollar-proof as a large-cap gets — it rents uniforms, mats, and first-aid/safety services to over a million North American businesses on sticky, recurring contracts, with essentially no overseas translation risk. It's a wide-moat compounder with 51% gross margins, strong free cash flow, and 45 consecutive years of dividend growth, and its pending UniFirst acquisition would push its North American share toward ~50% with an estimated $375M in synergies. When global stress is rising and the dollar is firm, that domestic, recurring, price-it-up cash engine is exactly what the "FX headwind" crowd wishes it had.

CTAS trades around $170, down roughly 22% over the past year — a meaningful de-rating for a company of this quality. The honest caveats: even after the pullback it's not cheap (~36x earnings, ~1% yield), the UniFirst deal is under FTC review (timing and remedies are uncertain), and a sharp economic slowdown would soften employment-linked uniform demand. This is a quality-on-sale idea, not a deep-value bargain. But on the strong-dollar thesis, Cintas is a quintessential domestic winner.

Bottom line: A stronger dollar is not trivia. It's a catalyst for "surprise" earnings disappointment — and a tailwind for domestically focused names with real pricing power.

4) Biotech reminder: one data point doesn't make a franchise — but it does reveal the market's psychology

Biotech is where Wall Street does its worst work. They hype the upside, ignore the risks, then act shocked when the risk shows up.

Abivax shares surged ~31% after fresh data on its lead drug, rebounding from a ~44% plunge earlier this month tied to cancer cases seen at the highest dose in the same trial program. That round trip — down 44%, then up 31% — on a single clinical-stage company with no revenue is the whole lesson. (To be clear: Abivax is the cautionary example here, not a recommendation — a single-asset, pre-revenue name that can swing violently on one data point.)

Here's what matters if you don't want to be exit liquidity. Biotech trades on narrative until it trades on finance — when cash gets tight, "good charts" with weak balance sheets get punished. Volatility isn't the enemy; dilution is. And M&A is the great sorter: deals spotlight who's real and who's a perpetual fundraiser.

So the best filter isn't "next catalyst." It's cash runway, a realistic path to approval, strategic fit for buyers, and a model that doesn't require perfect timing. The cleanest way to own that discipline is to skip the lottery ticket and buy the profitable franchise that doesn't need the capital markets at all.

Company: Vertex Pharmaceuticals (SYM: VRTX)
Profitable, cash-rich biotech; a dominant cystic-fibrosis franchise plus a diversifying pipeline

Vertex is the anti-Abivax. It's solidly profitable — roughly 38% operating margins, about $4.2 billion in annual operating cash flow, and over $11 billion in short-term assets — so it funds its R&D and dealmaking from its own cash, never forced to sell stock at the bottom. Its cystic-fibrosis franchise (Trikafta/Alyftrek) throws off durable cash, and it's diversifying into new revenue drivers: the non-opioid pain drug Journavx, the Casgevy gene therapy, and a promising kidney-disease pipeline. Management guides to roughly $13 billion in 2026 revenue.

VRTX trades around $445, about 12% off its 52-week high, with a Buy-leaning consensus near $550. The honest risks: it's still heavily dependent on the CF franchise, pipeline setbacks happen (it recently discontinued one mRNA CF program), and any single trial readout can move the stock. But the entire point of this section is that capital markets decide who survives to the payoff — and Vertex is a real business that already has. It's the survivor, not the fundraiser.

Bottom line: Biotech headlines move fast. Capital markets move slower — and that's what decides who survives to the payoff.

Before You Go

Ask yourself this: If the market is so confident about AI, energy, and a "soft landing"… why do we need so many narratives to keep everyone calm?

That's usually your tell. When investors need constant reassurance, it means positioning is crowded. And crowded trades don't need bad news to break. They just need less good news.

If you want the retail edge, keep doing the opposite of what Wall Street trains you to do: Ignore the noise. Track the funding. Buy cash flow when it's unloved. And let the Street chase the headlines.