Defensive Stocks To Balance AI Exposure This Fall

Have you noticed how quickly a portfolio can start looking like a single bet? One good run in data-center names, a few extra adds on weakness, and suddenly most of the risk sits in the same crowded trade. I have been there. The charts still look fine until the discount rate moves, and then those generous multiples stop feeling generous at all.

Why Defensive Stocks Matter When AI Multiples Stretch

That is the quiet problem with concentrated AI exposure. The buildout is real. The spending is visible. The earnings stories are still improving in plenty of cases. None of that cancels a simple piece of market math. When expected policy rates rise, the present value of distant cash flows falls. High-duration growth stocks feel that first. Lower-multiple businesses with stickier demand feel it later, if they feel it much at all.

A model book recently trimmed a data-center related name after a strong run and started two fresh positions instead. The buys were 240 shares of BNY, the old Bank of New York Mellon franchise, near $161, and 365 shares of Kimberly-Clark near $108. After those tickets, each name is about one percent of the portfolio. Small on purpose. The point is balance, not a dramatic all-clear on technology.

I like that sizing. One percent is enough to matter if the thesis works and small enough that a stumble does not wreck the year. In my experience, people talk about diversification as if it requires a dozen new ideas. Sometimes two boring additions do more work than another chase in the same crowded theme.

Higher interest rates raise the discount rate applied to future earnings and cash flows, which compresses price-to-earnings multiples. That is not a philosophy. It is arithmetic.

Markets are leaning toward a September policy hike in about two weeks. Whether that hike arrives exactly on schedule is almost beside the point. The expectation alone changes how investors price long-duration stories. If you only own the stories that need cheap money to look cheap, you are negotiating with the calendar.

The Rotation Is Partial, Not A Panic Exit

Nobody serious is arguing that artificial intelligence spending disappears next quarter. Capex plans from the large cloud platforms still point to a multi-year build. Power, cooling, networking, and specialized chips remain scarce in the good way. The smarter question is whether every dollar in a personal book should stay tied to that same scarcity.

Perhaps the most interesting aspect is how uneven the pain can be. A stock that already rerated to a premium multiple can drop on a modest change in the risk-free rate even if next year’s earnings estimate barely moves. A stock trading at a mid-teens multiple with a fat dividend has less air underneath it. That is the entire case for adding defensive stocks without abandoning growth altogether.

Think of it like packing for a trip when the forecast keeps flipping. You do not throw out the summer clothes. You add a jacket. The jacket here is fee-heavy financial infrastructure and branded household products. Not glamorous. Useful.

BNY: A Trust Bank That Does Not Live Off The Credit Cycle

First name up is BNY. It sat in a watchlist for weeks before the buy. That waiting period matters. Watching a name without owning it is how you learn whether the story is a slogan or an operating reality.

Compare the model to a conventional lender. A typical large bank gathers deposits and pushes those deposits into consumer and commercial loans. That model is tightly bound to credit quality, the shape of the yield curve, and net interest income. When loans sour or the curve misbehaves, the story gets noisy fast.

BNY is built differently. Its core job is servicing, safeguarding, and moving financial assets for institutions. Roughly seventy percent of revenue is fee based. Loans exist, but they are not the engine. Asset growth, meaning the market value of what clients park with the firm, is a primary driver of those fees. If markets grind higher over time, the meter keeps running.

The credit book is aimed at investment-grade clients who already use the non-credit services. That overlap is easy to miss and important. You are not underwriting a random borrower. You are extending limited credit to a relationship that already pays you to keep the pipes working.

• Fee-heavy revenue mix rather than loan-spread dependence
• Institutional custody and servicing as the center of gravity
• Credit used as a relationship tool, not the main profit engine
• Sensitivity to asset values more than to charge-off cycles

I have found that investors still lump every ticker with “bank” in the description into one mental folder. That habit is expensive. A trust platform and a consumer lender can trade like cousins in a panic and then diverge for years once people read the income statement.

What Changed After The 2022 Leadership Reset

Robin Vince took over as chief executive in 2022 and pushed a cultural reset that is now visible in the numbers. The firm leaned harder into selling tools and services across the investment industry and raised the technology spend that those clients actually pay for. Culture talk is usually cheap. Margin expansion is not.

From 2022 to 2025, revenue moved from $16.5 billion to $20.1 billion. Pre-tax margin climbed from 21.1 percent to 35.1 percent. Earnings per share rose from $2.88 to $7.40. That is not a rounding error. That is a company that used a few years of operational discipline to change its earnings power.

Steady top-line growth plus a much fatter margin is the combination you want in a defensive add. You are not buying a melting ice cube that merely looks cheap. You are buying a platform that already proved it can convert scale into profit without needing a heroic credit cycle.

Those figures will not repeat at the same slope forever. Margin gains eventually flatten. That is fine. The initiation case does not require another fourteen-point jump in pre-tax margin. It requires the firm to keep harvesting a more efficient base while asset pools and servicing volumes grind higher.

A $180 Target And Why The Multiple Is The Point

The new position carries a $180 price target. That target uses about 17.5 times consensus 2027 earnings of $10.25. Is that aggressive? Not really, if the earnings path holds and the business keeps looking less cyclical than a classic lender.

In a market that may apply a higher discount rate to long-duration growth, a mid-to-high teens multiple on a fee platform can look almost old-fashioned. Old-fashioned is useful when the alternative is paying twenty-five or thirty times for a story that needs rates to cooperate.

Could the stock stall if markets sell off hard and custody assets shrink? Yes. Fee income linked to asset values is not immune to a drawdown. The difference is the shape of the hit. You are not staring at a sudden spike in loan-loss provisions on a leveraged consumer book. You are staring at a volume and market-value dip that can reverse when risk appetite returns.

That is why this name can sit next to technology without pretending to replace it. Different drivers. Different failure modes. That is the whole idea of balance.

Kimberly-Clark: Household Brands With A Healthcare Overlay

The second ticket is Kimberly-Clark, the paper-products specialist behind Cottonelle, Huggies, and Kleenex. The name returned to a watchlist this week after living there once before. The first serious look came in November 2025, right after management agreed to buy consumer-health company Kenvue at an enterprise value near $49 billion.

The stock fell from about $120 to $102 in the days after the announcement. It still has not fully retraced that drop. Markets hate complexity, and this deal arrived with two loud objections. One was litigation risk around Tylenol. The other was a simpler gut reaction: why is a toilet-paper company buying a healthcare portfolio?

Those objections are not imaginary. They are also not the entire story. When the transaction closes, expected before year-end, the combined company picks up over-the-counter medicine, oral care, skin care, wound care, and more baby products. Scale changes the conversation even if the category mix looks unusual at first glance.

The combined entity is set to become the second-largest consumer health and wellness company by revenue, behind only the obvious household giant, with ten different billion-dollar brands in the mix.

I originally preferred the larger rival for quality. That position was later sold in search of better growth. Fair enough. Quality without growth can become a bond substitute with a logo. The Kimberly-Clark pitch is that the Kenvue combination can re-rate the multiple if the integration is even halfway competent.

Synergies, Scale, And A Catalyst With A Date

Management has pointed to $1.9 billion in cost synergies. That number is large enough to matter and small enough, relative to the combined cost base, that it does not require magic. Joint procurement on materials can lower cost of goods sold. Systems can be rationalized. Overlapping corporate functions can shrink. None of that is romantic. All of it shows up in margins if execution holds.

Revenue synergies are the extra layer. Shared shelves, shared baby-care adjacency, and a broader wellness basket give the sales force more to put in front of retailers. I would not underwrite heroic cross-selling in year one. I would underwrite the option that a bigger catalog makes the company harder to sideline in buyer meetings.

1. Close the deal before year-end and remove the binary headline risk.
2. Start capturing procurement savings on jointly sourced inputs.
3. Collapse duplicate systems and central functions without starving brands.
4. Let the larger brand set support a modest multiple re-rating.

A dated catalyst is underrated. Plenty of “cheap” staples stay cheap because nothing on the calendar forces a rethink. A closing date does that work for you. Investors who refused to look during the announcement hangover often come back when the pro forma statements turn from theory into filings.

Valuation, Yield, And The Gap Versus The Category Leader

Shares recently traded near fourteen times 2026 earnings estimates with a dividend yield around 4.75 percent. Those are the traits you want when growth multiples look tired. Income is not a consolation prize here. It is part of the return math while you wait for the integration story to clarify.

The initiation target is $120, or about sixteen times consensus 2027 earnings of $7.46. That is still almost a four-turn discount to the larger consumer staple that trades near twenty times 2027 estimates. Four turns is a lot of skepticism already baked in. Some of that skepticism is earned. Some of it is inertia.

If Kenvue brings faster growth and better mix, the market can pay a higher multiple without anyone needing a new religion about paper products. That re-rating is the upside. The yield is the ballast. Together they make a defensive add that is not just “less bad” than a falling semiconductor name.

Simple hold case:
  Current multiple: ~14x 2026 earnings
  Target multiple: 16x 2027 earnings
  Dividend while you wait: ~4.75%
  Deal close: expected this year

Will every synergy dollar land on schedule? Of course not. Integrations slip. Cultures clash. Retailers squeeze. That is why the position is one percent, not a conviction-max overweight. You are paying for optionality with a cash yield attached, not betting the farm on a perfect merger model.

How Rate Expectations Change The Whole Screen

Let us stay with rates for a minute, because this is the hinge. A higher path for policy rates does two jobs at once. It lifts the hurdle rate used in every discounted cash-flow sketch. It also changes the relative appeal of cash and short paper versus long-duration equity.

Growth stocks can still win in that world. They just need earnings revisions to outrun the multiple squeeze. That is a tougher race than the one they ran when money was easy and every long-term story got the benefit of the doubt. Staples and fee platforms do not need that benefit. Their cash arrives sooner.

I keep a simple habit. When the market starts talking about a hike that is only two weeks away, I scan the book for names whose valuation depends on 2030 looking perfect. If too many names fail that test, I look for businesses whose customers still buy the product when mortgages get expensive. Tissues, diapers, custody fees. Unfashionable demand is still demand.

Does that mean rates are the only variable? No. A hard landing would pressure almost everything, including BNY asset values and retailer orders for paper goods. Defensive is not invincible. It is less fragile to a change in the discount rate. That distinction gets lost when people treat “defensive” as a synonym for “cannot fall.”

Position Sizing When You Already Own Too Much Of One Theme

One percent each sounds timid until you remember how portfolios actually break. They rarely break because a one-percent idea went to zero. They break because a thirty-percent theme all moved together. Adding ballast is less about finding the next winner and more about reducing the chance that one narrative owns your year.

After selling the data-center name, the book had room. Using that room on two unrelated cash-flow types is cleaner than recycling the proceeds into a cousin of the stock you just sold. I have watched that recycling habit in real time. It feels disciplined because you “took profits.” It is often just staying in the same crowded room with a different chair.

• Keep AI winners if the fundamental case is intact.
• Trim names where the multiple already prepaid years of good news.
• Add businesses with earlier cash flows and simpler demand.
• Cap new ideas at a size that cannot dominate the book.

There is also a behavioral bonus. A small staple or trust-bank line item gives you something to do besides stare at semiconductor candles. That sounds silly until you have lived through a week when every growth name ticks down together and your only instinct is to average down. A separate sleeve interrupts that reflex.

Risks That Would Break The Thesis

Honesty first. BNY still feels market levels in its fee base. A long, ugly equity and bond drawdown would slow asset growth and pressure servicing volumes. Technology investment has to keep paying for itself. If clients decide the new tools are optional, the margin story stalls.

Kimberly-Clark carries deal risk until closing and integration risk after it. Litigation headlines can linger even when the strategic logic is sound. Consumer volumes can wobble if private-label gains share during a squeeze on household budgets. A 4.75 percent yield is attractive until a cut enters the conversation, and cuts enter the conversation when integrations go sideways.

Currency, input costs, and retailer negotiations never vanish in staples. Paper and pulp move. Freight moves. A strong dollar can pinch international results. None of that is new. All of it can still knock a quarter off course and test the patience of anyone who bought only for the dividend.

The combined risk that worries me more than any single headline is correlation sneaking back in. If a risk-off tape hits custody assets and consumer staples in the same week, the “balance” sleeve will not look balanced. That is possible. It is also why these are starter weights rather than heroic overweight calls.

What To Watch After The Tickets Are In

For BNY, the live checklist is simple. Revenue growth without giving back the margin gains. Fee mix staying dominant. Credit remaining a satellite, not a surprise centerpiece. Management commentary on technology adoption by institutional clients. If those items hold, the 17.5 times framework on 2027 earnings can stay reasonable even if the multiple does not expand in a hurry.

For Kimberly-Clark, watch the close, the first synergy scorecard, and whether the brand set is treated as a portfolio or a science project. Ten billion-dollar brands only help if they are funded and merchandised like winners. Listen for tone on litigation and on private-label competition. The multiple gap versus the category leader should narrow if growth and margins improve together. If they do not, the stock can sit at fourteen times and pay you to wait. That is an acceptable consolation. It is not the bull case.

Policy meetings matter for both, though in different ways. A hike that lands as expected may already be in the price. A hike paired with hawkish language could pressure long-duration growth again and make these two names look smarter by comparison. A surprise pivot the other way would likely revive the AI complex and leave the defensive sleeve looking early. Early is not the same as wrong.

A Practical Way To Think About “Cheap” In This Tape

Cheap is a slippery word. A stock can be cheap versus its history and expensive versus the cash you can earn in a money-market fund. It can be cheap versus a peer and expensive versus the growth you are actually getting. The useful version of cheap right now is cheap versus duration risk.

BNY at a mid-teens multiple on improving earnings power is a duration argument as much as a bank argument. Kimberly-Clark at fourteen times with a high-four yield is a duration argument dressed as a consumer argument. Both collect cash earlier than a story that needs 2029 capex to justify today’s price.

I would rather own a little of that earlier cash than pretend I can time the exact week when AI multiples stop compressing. Timing that week is a hobby. Building a book that survives the wait is a job.

Putting The Two Names Side By Side

They do not hedge each other in a textbook sense. One is financial infrastructure. One is branded consumables plus an incoming health portfolio. The link is temperament. Both are less fashionable than the AI complex. Both can keep producing earnings if the buildout narrative gets noisy. Both were available at weights that do not force you to become a full-time analyst of either industry overnight.

Notice what is missing from that table. There is no claim that either stock is the best idea on the board. There is no claim that AI spending is finished. There is only a claim that a book with too much of one duration profile needed a second duration profile. That claim is modest. Modest claims age better than slogans.

A Note On Process, Patience, And Avoiding Hero Trades

Good process is often boring to describe, which is why it gets skipped. Watchlist first. Read the mix of fees versus spreads. Read the deal model twice. Wait for a sale that frees cash so you are not stretching to fund the new idea. Size the new idea so that being early is survivable. Then let the quarter-by-quarter evidence argue with you.

Hero trades feel better in the moment. They also create the exact concentration problem this whole exercise is trying to fix. If you need a stock to double for the year to work, you do not own a portfolio. You own a dare. One-percent sleeves are the opposite of a dare.

I will admit a bias here. I would rather explain a dull add that cushions a drawdown than explain why every name in the book needed the same macro outcome. Dull is underrated. Dull collects dividends. Dull still participates if the AI complex resumes leadership, because you never sold the whole growth sleeve to begin with.

The Bottom Line Without The Cheerleading

Two new lines do not turn a growth book into a fortress. They tilt it. BNY offers fee income tied to institutional asset infrastructure and a margin story that already showed up in reported results. Kimberly-Clark offers branded demand, a sizable yield, and a closing catalyst that can force a second look at a stock still marked down from its deal announcement.

If rates rise as the market fears, those traits should matter more, not less. If rates do not rise, you still own businesses that can compound without a perfect multiple expansion. That is the unglamorous version of balance, and it is usually the version that survives contact with the calendar.

Keep the AI winners you still believe in. Stop pretending every spare dollar has to live in the same neighborhood. And if a one-percent position in a trust bank or a paper-and-health combination feels too quiet, ask what your book will feel like on a week when discount rates move and the crowded trade all answers at once.