Stocks making the biggest moves midday: SK Hynix, Meta, WD-40, Delta, Vodafone & more

Midday movers were broadly mixed: Meta jumped ~6% on a Reuters report that Facebook is looking to put an AI chip into production in September, with expected compute cost reductions. SK Hynix surged ~17% after its ADR priced at $149 to raise $26.5B for expansion, while Marvell/Intel fell ~3% on SK Hynix’s Nasdaq debut. On the calls, Wells Fargo lifted Seagate to overweight and raised its price target to $1,100, citing a path to >$50/share in earnings and potential capital returns (buybacks/dividends), while WD-40 rallied ~10% after Q3 adjusted EPS of $2.33 beat expectations of $1.56 and guidance was raised; Delta slipped >2% despite a Q2 beat as management expects jet-fuel pricing power to persist. Circle rose ~6% after U.S. OCC approval to launch a crypto-focused bank, and Vodafone jumped ~13% after a new ~16% stake by Xavier Niel worth about $6B.

Analysis

META is the cleanest fundamental winner because the market is finally pricing that in-house silicon can bend the AI cost curve, which matters more than the chip itself. If the September production timeline sticks, the next leg is not just margin upside but a higher ceiling on ad load, video, and AI product experimentation because free cash flow stops looking hostage to third-party accelerator pricing. The risk is execution: any yield/tape-out slip would turn this into a one-day squeeze rather than a multi-quarter earnings revision.

The harder read is the semiconductor complex. A large new public financing event for an AI supply-chain player signals more capex, which is good for equipment demand but can be bad for pricing discipline if it ultimately adds memory supply into an already crowded investment cycle. That makes MRVL and INTC vulnerable as “AI beneficiary” multiples compress when investors realize more of the stack is being internalized or commoditized. STX is the cleaner way to express AI storage demand, but only if enterprise/data-center spending outruns the broader supply build; otherwise buybacks can cushion downside, not re-rate the stock.

CRCL’s approval is a balance-sheet story, not just a product story: a bank charter should lower funding friction and improve credibility with institutional counterparties, but it also increases regulatory operating drag. NFLX’s move reads as defensive: bundling and live channels suggest management is trying to arrest engagement softness, which is usually a sign the market should demand more proof on ARPU and churn before re-rating. DAL and LAMR are more duration-sensitive; if fuel and rates keep easing, their near-term pressure should fade, but any re-acceleration in crude or yields would quickly reverse the setup.

Key Decisions for Investors

• Long META into any weakness; 1-3 month thesis is margin expansion from lower AI compute costs. Falsify if the company misses the September production window or guides capex back up.
• Pair trade: long META / short MRVL or INTC for 1-3 months as a proxy for in-house silicon displacing merchant AI hardware demand. Cover if MRVL order commentary inflects up or INTC shows real foundry traction.
• Small long STX on pullbacks for 6-12 months as the cleaner AI-storage beneficiary with capital-return support. Stop if data-center capex slows or enterprise storage pricing rolls over.
• Avoid chasing NFLX; use rallies to fade unless management proves bundling lifts engagement within the next two quarters. A sustained move above recent highs without churn improvement would be a short setup.
• Treat CRCL as a watch item rather than a full-size long until deposit growth and fee income data confirm the charter benefit; the approval lowers cost of capital, but operating friction could offset it.