How Are the AI-Immune HALO Stocks Doing in 2026?

The article argues that HALO stocks are holding up unevenly in 2026: ExxonMobil is up 24%, FedEx nearly 30%, Coca-Cola 13%, Deere 26%, and Caterpillar 59%, while McDonald's is down about 7% year to date and 16% since late February. It frames AI as a tailwind for energy and equipment providers, but a headwind for some software and financial data firms, with rising consumer weakness still pressuring discretionary names. Overall, the piece is thematic stock-picking commentary rather than a single catalyst, with moderate sector implications.

Analysis

The market is starting to price a more important second-order effect than simple “AI disruption”: physical bottlenecks. Names tied to data-center power, logistics throughput, and earthmoving equipment have a cleaner earnings path because AI capex still has to be embodied in concrete, steel, fuel, and grid upgrades before it becomes software revenue. That makes the current leadership in CAT and, to a lesser extent, FDX more durable than the multiple reset in software-adjacent data providers, where AI can compress pricing power before it shows up in revenue.

The more interesting split is within the HALO bucket itself. Asset-heavy consumer franchises look like defensive holdings only until household budgets tighten; then low obsolescence does not matter if transaction volume rolls over. That makes MCD the weakest quality-growth/defensive hybrid here, while KO is better insulated because it has more pricing latitude, a lower ticket size, and less direct sensitivity to discretionary trade-down.

The biggest near-term risk is that the “AI build-out” trade becomes overcrowded and rotates from capex beneficiaries into earnings realizers. CAT is the cleanest momentum expression, but it is also the most vulnerable to a 3-6 month pause in hyperscaler spend or a backlog digestion phase. FDX has a more subtle setup: if AI raises route optimization and labor productivity, margins can expand without top-line acceleration, giving it a less reflexive but potentially steadier rerating over the next 2-4 quarters.

Consensus is probably underestimating how much of this is a factor-duration trade masquerading as a theme trade. The market is paying up for companies with visible, near-term cash flow tied to AI physical infrastructure, while punishing businesses whose AI exposure is ambiguous or where end-demand is weakening. That suggests the next leg is less about buying “AI winners” and more about shorting the businesses where AI accelerates commoditization before investors have fully marked down terminal margins.

Key Decisions for Investors

• Go long CAT on a 1-3 month horizon on dips; the setup favors continued multiple support as long as hyperscaler capex stays firm. Risk/reward is attractive, but trim aggressively if order commentary suggests backlog normalization faster than expected.
• Use FDX as a lower-beta AI productivity beneficiary over 2-4 quarters; long only if you want earnings leverage from automation rather than pure capex beta. The trade works best if margins expand without needing a strong macro rebound.
• Short MCD against KO as a consumer-demand relative-value pair for the next 1-2 quarters. The thesis is that low ticket size does not fully shield against weak household spend, while KO has better pricing and mix resilience.
• Avoid chasing IBM and TRI/MCO downside after the first AI selloff; wait for a second leg lower or evidence of guidance compression. The cleaner entry is on any rebound that fails to recover prior support, because the market is still repricing long-duration intangible franchises.
• If you want a thematic hedge, pair long CAT with short a basket of AI-exposed information intermediaries or software names. That expresses the view that physical picks-and-shovels outperform digital intermediaries as AI monetization shifts from model excitement to infrastructure spend.