Almost Everyone Chased AI. We Picked Caterpillar and Exxon — Here’s How That’s Going.

In early January, when almost everyone was crowding into pure AI and mega‑cap growth, we published a concentrated 5‑stock portfolio for early 2026 under stockresearch.ai: JPMorgan, Amazon, Caterpillar, UnitedHealth, and ExxonMobil. It wasn’t a backtest or a hypothetical basket, it was a live, high‑conviction view of where we saw the best risk‑adjusted upside over the next 6–12 months.

A few weeks later, the portfolio looks exactly like real investing always does: two big winners, one steady anchor, one high‑quality name in drawdown, and one thesis that’s under review. That mix is the point. If you’re pushing genuine conviction instead of hugging the index, you will never be 5‑for‑5. The real question is how you respond when the tape disagrees with you.

The Wins: Caterpillar and ExxonMobil

Caterpillar: We Bet on an Industrial Re‑Rating

*Price on January 15, 2026

When we put Caterpillar (CAT) into the January report, the pushback was predictable: “Why buy a cyclical industrial near all‑time highs?” The answer was that CAT stopped being a plain‑vanilla cyclical a while ago.

Our thesis:

CAT is morphing into an industrial tech platform, not just a seller of iron. Its “Cat AI Assistant” and autonomous solutions turn machines into data‑rich, software‑enhanced assets.
The profit mix is quietly tilting toward services and uptime, which are structurally higher‑margin and less cyclical than equipment sales.
Financial quality (high returns on equity, strong cash generation, buybacks) supports a higher multiple than the old 12–15x industrial range.

We argued that the market was still partially pricing CAT as a classic industrial while the business was being repriced as “industrial tech.” Since then, the stock has moved sharply higher and already pushed through our original target. That tells us two things: the market is starting to see what we saw, and our next job is not victory‑lapping, but deciding whether to take profits, raise our target, or demand more proof from earnings before committing fresh capital.

ExxonMobil: The Ballast That Paid Us to Wait

*Price on January 15, 2026

ExxonMobil (XOM) was never meant to be the hero of the portfolio. It was the ballast:

A base of advantaged, low‑cost barrels in places like Guyana and the Permian.
A clear roadmap for earnings and cash flow growth into the end of the decade.
A healthy dividend plus buybacks, translating into a high single‑digit shareholder yield.

The idea was simple: in a market obsessed with AI and high‑multiple growth, we wanted something that could quietly return capital and hedge inflation, commodity, and geopolitical risks. That’s pretty much what XOM has done. The stock moved above our initial target while continuing to pay out income.

Together, CAT and XOM validate two different parts of our framework: spotting re‑rating opportunities where the business mix has quietly changed, and using boring, cash‑rich names as a stabilizer around the growth bets.

Subscribe now

The Anchor and the Opportunity: JPM and AMZN

Before we get to the pain trade, it’s worth noting that we didn’t ignore AI or growth entirely — Amazon sits at the center of our AI‑infrastructure thesis, and JPMorgan is our way of anchoring the portfolio in a high‑quality financial compounder.

JPMorgan: The Fortress Bank Behaving Like a Fortress

JPMorgan (JPM) was the most straightforward inclusion. The basic case:

A “fortress” balance sheet and scale advantages that are almost impossible to replicate.
Best‑in‑class returns on capital compared with major US peers.
Multiple earnings engines: net interest income, investment banking, trading, and now the Apple Card portfolio.

Since January, JPM has largely done what we expected a core financial holding to do: continue to print solid earnings on a reasonable multiple. It hasn’t been the fireworks name of the portfolio, but it has behaved like a reliable anchor, which is exactly the role it was meant to play.

Amazon: Good Business, Tough Entry — and a Better Setup Now

Amazon (AMZN) is where nuance matters.

In our January report, we didn’t pitch “people will keep shopping online.” The focus was on:

AI infrastructure via Project Rainier and Amazon’s custom silicon stack.
The combination of AWS, Trainium, and key AI customers as a vertically integrated way to monetize the AI wave.
High‑margin advertising and ongoing robotics deployment in fulfillment as structural drivers of margin expansion.

Those pillars haven’t disappeared just because the stock is lower. The reality is more boring and more honest: we were early on price, not obviously wrong on the business.

From today’s level, that actually improves the risk/reward:

The long‑term thesis is still multi‑year in nature.
The market has taken some heat out of the AI/hype multiple.
The same cash‑flow drivers we liked in January are available at a better entry.

Amazon is the textbook case of “good thesis, impatient market.” We’re not pretending this one has “worked” yet — but we still like the setup from here more than on the day we published.

The Pain Box: UnitedHealth Under Review

UnitedHealth (UNH) is the hard one.

In January, we framed UNH as a fallen angel:

A high‑quality healthcare and services giant that had been hit by a spike in medical costs and a messy year on the policy front.
A path to normalization as utilization patterns stabilized and management refocused on profitability.
An opportunity to buy a structurally strong franchise at what looked like distressed pricing for a blue‑chip name.

The intent was not crazy. Where we’ve been challenged is on timing and depth of the damage. Since the report, guidance and regulatory noise have extended the pain more than our base case assumed. What looked like “we’re near the bottom of the margin shock” is increasingly looking like “we’re still in the middle of the clean‑up.”

So we’re not going to spin it: right now, UNH has graduated from “mispriced quality” to higher‑risk turnaround. That doesn’t automatically mean “sell,” but it does mean we are explicit about what we need to see:

Clear evidence that medical cost ratios are stabilizing in a sustainable way.
Better visibility on reimbursement and policy for the next 1–2 years, not just the next quarter.
A risk/reward profile that compensates for the policy and earnings volatility involved.

Until then, UNH sits in the “thesis under review” bucket, not the “pound‑the‑table” bucket.

What This Says About Our Process

If all you saw were the CAT and XOM charts, you could write a cute “told you so” post. If you only looked at AMZN and UNH, you could write an apology tour.

The reality is more interesting, and more useful:

Two names where the thesis has already been rewarded by the market.
One anchor that’s behaving exactly as a high‑quality core holding should.
One high‑quality compounder where we were early on price, but still like the business.
One contrarian where the facts have shifted against us, and we’re openly reassessing.

That’s what a real, concentrated portfolio looks like. The January report wasn’t meant to be a flawless prediction; it was meant to be a transparent snapshot of our best ideas and the frameworks behind them.

If you want to see the full logic — the macro backdrop, detailed write‑ups for all five names, risk sections, and the portfolio construction rules we used — you can grab the complete “Top 5 US Stocks for Early 2026” report from stockresearch.ai. It’s a free sample while we’re developing our platform.

Stockresearch Free Reports

And in future posts, we’ll keep doing this in public: not just picking names, but showing how we update the story when the market answers back.

Disclaimer: This report is for informational purposes only and does not constitute financial advice. All data is sourced from publicly available information as of February 2026. Investors should conduct their own due diligence and consider their risk tolerance before making investment decisions. Stock market is highly volatile and may result in total loss of capital.