Josh Brown says the stock market is ‘broadening out.' Here’s how to play it

Josh Brown (Ritholtz) argues investors should “broaden out” from hyperscalers/“Magnificent 7” into non-hyperscaler stocks seeing AI-driven upside earnings surprises. The call followed Alphabet raising its capex forecast to $195B–$205B, which pressured hyperscalers Thursday, while Travelers (+1% to +2%) and Chubb rose similarly and industrials like GE Vernova (+5%) and Caterpillar (~+1%) gained. Net impact is a sector/positioning shift rather than a single-company fundamental change.

Analysis

This reads more like a factor migration than a clean AI fundamental shift: the market is punishing the companies funding the buildout while rewarding names that can monetize AI through labor savings or operating leverage. That favors TRV, CB and arguably BRO because insurers can convert better underwriting and claims triage into near-term margin lift without needing top-line acceleration; the key is that expense ratio improvement can show up before anyone can prove incremental AI revenue.

GEV is the cleaner industrial expression because data-center power demand and grid capex create a direct order bridge from AI spend to equipment backlog. CAT is a second-order winner, but more diffuse: it participates through power generation, backup systems and infrastructure intensity rather than a single AI-specific lever, so its upside is likely lower-beta and slower to compound.

The risk is that this becomes an overcrowded ‘non-megacap AI’ narrative with weak follow-through. If hyperscaler free-cash-flow margins hold and capex simply reflects demand visibility, the rotation can reverse quickly; the thesis only works if AI spend increasingly shows up as earnings revisions in the beneficiaries rather than just lower multiples for GOOGL/META/MSFT/AMZN. Over 1-3 months, watch for commentary on expense leverage and order growth; over 6-18 months, the winner is whichever cohort can turn AI into structural ROIC expansion, not just higher investment intensity. Falsifiers: no upgrade cycle in TRV/CB/BRO EPS, or any moderation in hyperscaler capex guidance.

Key Decisions for Investors

• Put on a 1-3 month pair: long TRV/CB basket vs short META/MSFT basket. Base case is continued breadth rotation; risk/reward improves if upcoming earnings show even modest underwriting/expense-ratio upside. Cut if hyperscaler margins stabilize and capex guidance stops rising.
• Buy GEV on pullbacks as the highest-beta AI power proxy. Best entry is after 3-5% weakness, with a 2-3 month target tied to backlog/order revisions; thesis breaks if data-center demand commentary softens or utilities defer spend.
• Use CAT as a lower-volatility complement, but size it smaller than GEV. Prefer it only if you want exposure to infrastructure spillover without paying full multiple for AI purity; exit if industrial order growth decelerates or the rotation narrows back to software/mega-cap tech.
• Add BRO as a cleaner, less headline-sensitive way to express AI-driven insurance productivity. Favor it over direct balance-sheet insurers if you want upside from process automation with lower catastrophe/underwriting noise.