Here's a rapid-fire update on our stock portfolio, including the 5 names to buy

CNBC’s Investing Club focused on trimming parabolic AI-linked moves while citing an opportunity set in select names: Corning, GE Vernova, Eaton, and (outside AI) FedEx Freight and Johnson & Johnson. The discussion highlights AI memory/data-center cost pressure (so hyperscalers can’t “afford not to spend,” but profitability worries remain fluid) and urges patience/optionality in larger platforms like Alphabet and Microsoft. On the domestic tape, they trimmed Home Depot on rate sensitivity amid hopes for a Fed easing cycle, while Starbucks trimming was considered after a rally; meanwhile, J&J’s Q2 was described as imperfect due to a $150M Abiomed revenue miss but with the broader pharma business strong.

Analysis

The clearest tradeable message is a rotation from software capex promises to the industrial bottlenecks that monetize the buildout. That favors ETN, GEV, LIN and, selectively, NVDA over MSFT, CRM and AMZN, because the market is beginning to pay for power delivery, cooling, and alternative compute while it discounts software names that still need to prove AI ROI. Over the next 1-3 months, this should be expressed as a relative-value trade: the winners can keep rerating only if hyperscaler capex stays intact, but the upside in the laggards is capped until monetization evidence arrives.

The trims in HD and SBUX read more like balance-sheet/valuation discipline than a hard downgrade. Both are still fighting the same macro headwind: lower rates would help, but that catalyst is likely months away, so the opportunity cost of owning them is high versus higher-beta beneficiaries like GS and WFC, where even modest market activity or underwriting pickup can expand earnings faster than multiples compress. FDX/FDXF is the quieter cyclical setup: if freight data keep firming into the next quarter, the stocks can re-rate before consensus fully believes the recovery.

Contrarian takeaway: the market may be over-focusing on whether AI spending is "too big" and underappreciating that infrastructure scarcity can extend the trade well beyond one earnings season. What would falsify that is a real capex slowdown from GOOGL/MSFT/AMZN or a roll-over in power/electrical order books; absent that, dips in the supply-chain winners should be bought rather than chased. If software monetization starts to inflect, then MSFT and CRM become the catch-up longs, but that is still a later-stage catalyst.

Key Decisions for Investors

• Long ETN / GEV basket vs short MSFT for 1-3 months: express the capex-to-infrastructure rotation with better near-term earnings torque; cut if hyperscaler guidance shows capex moderation.
• Buy NVDA on 3-5% pullbacks over the next 2-6 weeks: the stock is still the operating leverage proxy to AI spend, but only at a better entry point after recent derisking; invalidate on weaker data-center demand commentary.
• Pair long GS / WFC vs short HD / SBUX for 1-3 months: favors names with clearer earnings leverage to activity over rate-sensitive consumer laggards; reassess if 10Y yields break materially lower and housing data inflect.
• Add FDX or FDXF on confirmation of freight improvement over the next quarter: use J.B. Hunt-style freight strength as the catalyst; stop if July/August volume data roll over.