Fisher’s AI-Infrastructure Crowding Trade Is Real—But It’s the “Grid-to-Aircraft-to-Network” Link That Should Matter for Investors

Fisher Asset Management’s Q2 2026 13F adds to a “real-economy infrastructure” theme that sits downstream of semiconductors: it couples compute growth with grid upgrades, aircraft powerplants, and enterprise networking. That matters because AI capex doesn’t just buy chips—it forces utilities, defense/aerospace supply chains, and network equipment to expand, retool, and keep operating.

Fisher’s largest “infrastructure” bet in this theme centers on GE Vernova. Its business mix directly touches the electrical system buildout needed for AI data-center loads: power equipment, grid modernization, and electrification services. The linkage is structural: more AI compute drives more electricity demand, and that demand must be delivered through transmission/distribution capacity and conversion infrastructure.

The other “real economy” leg in the Fisher read-through is GE Aerospace. AI drives demand not only for data centers, but also for the industrial throughput that moves people, parts, and defense readiness. In aerospace, that expresses through engine production, maintenance/replacement cycles, and defense/propulsion demand—cycles that can tighten supply when broader capex is rising.

Fisher’s disclosed new position in Cisco Systems fits the “infrastructure” framing at the traffic layer: compute output only matters if networks can move it reliably and securely. AI data centers increase east-west and north-south traffic, drive network assurance demand, and expand security/segmentation needs. In other words, networking converts AI compute into services.

GE Aerospace and Cisco Systems both show strong recent profitability in the data tool snapshots, while GE Vernova shows large scale revenue with positive net income even as it operates in a capital-intensive grid cycle. That blend—profitability plus infrastructure sensitivity—is exactly what makes “real-economy AI” trades durable when AI capex volatility hits.

Short-term, “crowded infrastructure AI” trades can reprice on guidance beats, order visibility, and data-center capex sentiment. Longer-term, the thesis depends on whether grid and mobility replacement cycles keep compounding—i.e., infrastructure demand persists beyond a single AI hype window.