Stocks making the biggest moves midday: Meta, General Mills, Datadog, Sandisk & more

Meta surged 11% as it builds a new cloud business and will sell excess AI computing capacity to outside customers. Across earnings, General Mills shares rose 6% on Q4 adjusted EPS of $0.95 and revenue of $4.61B (both above consensus) and announced $3B cumulative cost savings through FY2030; Progress Software climbed 18% on a Q2 beat and stronger Q3 guidance, while Nike gained 4% despite a 12% YoY China sales decline. In deal news, Getty/ Shutterstock dropped after Getty called off the merger over U.K. regulator demands (Shutterstock -30%, Getty -6%), and Alcoa fell 9% after agreeing to acquire $4.1B of South32 assets. Offsetting the caution, Datadog rose 2% on an acquisition of AI startup Adaptive ML, and Salesforce (+5%) / ServiceNow (+6%) rose after a Guggenheim upgrade backed by AI-linked valuation support.

Analysis

The cleanest signal is that AI is shifting from pure capex absorption to monetization and financing. A hyperscaler monetizing spare compute is a better margin story than another round of spend, but it also implies the market may be over-assigning durability to the current AI capex burst; the incremental revenue is likely lumpy and lower-margin than investors want to assume. The second-order winners are power, cooling, and infrastructure financiers, while the loser is the crowded “AI infra beta” trade if utilization data does not confirm.

For software, the upgrades to CRM and NOW matter more for multiple support than near-term fundamentals. The consensus is still pricing AI as a displacement risk; the counter is that large workflow vendors can embed agents and defend seat-level economics, which should cap downside over the next 3-6 months if renewals hold. DDOG’s acquisition is a small but relevant sign that agent deployment remains an add-on feature war, not a winner-take-all platform reset.

The broken media deal is a bigger strategic negative for the standalone asset than the market is likely modeling: it removes exit optionality and forces a slower self-help path. AA looks capital-intensive rather than accretive at first glance; the asset mix may improve cycle exposure later, but near-term balance-sheet use can compress FCF flexibility. Memory names are probably experiencing position cleanup after an extended run; that short only works if pricing or utilization rolls over in the next 1-2 quarters, not on momentum alone.

Key Decisions for Investors

• Long META / short SNDK on any SNDK bounce over the next 3-5 sessions; 1-3 month relative-value trade on AI monetization vs crowded AI supply-chain beta. Falsifier: memory ASPs keep firming and META’s cloud disclosure proves immaterial.
• Buy 2-4 month call spreads in NOW on a pullback of 3-5%; CRM is the lower-beta alternative if you want less valuation risk. Thesis breaks if renewal rates or seat expansion weaken as AI features roll out.
• Short SSTK vs long GETY for 1-2 months; the deal break likely leaves SSTK with the bigger premium unwind and more event-driven disappointment. Cover on any renewed bid chatter or regulatory workaround.
• Prefer BN over BE for AI power exposure; BN is the cleaner way to own financing/capital deployment into data-center power buildout. Treat BE as a trading name only if backlog converts into contracted projects, not just partnership headlines.