Jim Cramer's top 10 things to watch in the stock market Thursday

TSMC reported a 77% jump in Q2 operating profit and plans an extra $100B Arizona foundry investment to meet soaring AI demand, but NY-listed shares are still down ~4% premarket amid broader chip weakness. In deals, Uber agreed to buy Delivery Hero for $14.8B (premium after a higher bid) and Eli Lilly is paying $2.8B upfront for late-stage psychedelic drugmaker AtaiBley. Elsewhere, UnitedHealth is surging on a strong quarter and raised full-year profit guidance, GE Aerospace beat on sales/profits/FCF and lifted full-year targets, and J.B. Hunt shares are up 7% on a strong Q2 plus price-target hikes.

Analysis

The key signal is not the chip print itself but the market’s refusal to pay up for it. That usually happens when positioning is crowded and investors start discounting the next leg of returns: capex intensity, not demand, becomes the valuation debate. In that setup, TSM is better read as a supplier-side confirmation for AI infrastructure than as a near-term equity catalyst; the cleaner beneficiaries are the tool and equipment chain, while high-beta fabs and legacy semis can keep underperforming if AI monetization does not broaden fast enough.

By contrast, UNH and GE look like true earnings-quality names with visible operating leverage, which should attract capital if the market keeps rotating out of momentum tech. UNH’s reset is more durable than a one-quarter beat because it implies pricing discipline and a lower probability of future estimate cuts; GE’s upside is more cyclical, tied to aftermarket and engine throughput, but the cash-flow visibility supports a multi-quarter re-rating. Goldman sits in the same bucket structurally: if M&A and financing activity stay hot, revenue is less rate-sensitive than the Street models suggest.

The contrarian miss is that weak semis and strong old-economy growth are telling the same story: the market is broadening, not merely wobbling. That favors relative-value rotation over outright index beta. The risk to that thesis is a faster-than-expected AI spending reacceleration from NVDA or a macro shock that hits cyclicals first; otherwise, the next 1-3 months should reward idiosyncratic earnings strength over crowded theme exposure.

Key Decisions for Investors

• Buy UNH on strength or on any post-gap consolidation; target a 1-3 month re-rating as guidance revisions get reflected, with thesis invalidation if medical cost trends re-accelerate in the next earnings update.
• Add GE on pullbacks for a 3-6 month hold; the free-cash-flow upgrade matters more than the one-day move, and the main risk is engine/service disruption or supply-chain slippage.
• Overweight GS versus broad financials for the next 6-12 months; M&A and financing activity should support estimate upside, with invalidation if deal flow or capital-markets activity rolls over sharply.
• Use any semiconductor bounce to short INTC or AMD against TSM strength; this is a relative-value trade on execution dispersion, not a market-direction bet, and should work over the next 2-8 weeks unless NVDA resets the entire AI complex higher.
• Do not chase TSM after the gap lower; wait for a washout or a revision from the sell side. If TSM regains its premarket selloff and holds, that would confirm the market is turning the print into an AI-capex positive rather than a margin-dilution story.