Huatai Securities says the AI industry will shift in 2026 from large‑model pretraining to AI Agent commercialization, pushing inference compute and storage demand into a faster growth phase. The report highlights three market implications: domestic compute closed loops will form faster as super‑node interconnect and storage upgrades amplify inference demand, with upcoming edge product cycles (foldable devices, AI glasses) supporting further uptake; higher AI power consumption will lift volumes a

Huatai Securities says the AI industry will shift in 2026 from large‑model pretraining to AI Agent commercialization, pushing inference compute and storage demand into a faster growth phase. The report highlights three market implications: domestic compute closed loops will form faster as super‑node interconnect and storage upgrades amplify inference demand, with upcoming edge product cycles (foldable devices, AI glasses) supporting further uptake; higher AI power consumption will lift volumes and prices for MLCCs, inductors, capacitors and power semiconductors, creating a price‑up cycle that dovetails with domestic substitution; and upstream manufacturing, equipment and component localization will accelerate while advanced packaging content and value per system rise.