SPECIAL REPORTBULLISH

AI Capex Cycle Splits Winners From Margin Casualties As Earnings Season Grinds On

GE Vernova upgraded on serial guidance raises and Woodward flagged on aerospace-power demand, while Super Micro's turbulent margins and a looming Cerebras print test the trade's quality tier

Executive takeaway

The latest earnings tranche sharpens a widening divide within the artificial-intelligence complex: electrical and power-infrastructure suppliers are compounding upward guidance revisions, while hardware assemblers continue to convert revenue growth into unreliable gross margin. GE Vernova drew a rating upgrade on rising guidance and a brighter 2028 outlook, Woodward was highlighted as a beneficiary of simultaneous aerospace and power booms, and Super Micro delivered a good quarter undermined by the same margin volatility that has dogged it for a year.

The AI trade is no longer monolithic. Capital is migrating decisively toward the electrification and thermal-management layer of the buildout, where pricing power is structural and backlogs are contractual. GE Vernova's upgrade rests on a pattern investors reward above almost anything else: guidance raised every quarter, with a 2028 setup that implies the power-demand cycle is longer than the current order book suggests. Woodward is being positioned similarly, straddling aerospace aftermarket strength and grid-adjacent turbine demand. Contrast that with Super Micro Computer, which again posted respectable topline expansion while gross margin behaved erratically, the recurring signature of a business that assembles rather than differentiates. The upcoming Cerebras report, and what chief executive Andrew Feldman signals on customer concentration and capacity, becomes a live referendum on whether challenger silicon can command durable pricing against incumbent accelerators. Meanwhile the perennial capital-allocation debate persists between AMD's momentum-and-balance-sheet profile and Alphabet's discounted cash-generation case. Our conclusion for institutional portfolios: prefer the bottleneck over the box. Power, turbines, cooling and precision components are where AI demand converts into defensible margin.
Wire sources cited

Produced automatically by the INDY NEWS Desk from the public wire sources cited above, and checked against them before publication. Market commentary, not investment advice.