Kenanga Research said the artificial intelligence investment cycle is entering a second phase in 2026, with demand broadening beyond graphics processing units (GPUs) to inference, AI agents, and commercial applications.
The demand, it added, creates opportunities across optical connectivity, power, advanced packaging and semiconductor equipment, the report said.
The research house, citing views from Huatai Securities, said the initial AI investment cycle from 2023 to 2025 was dominated by large-model training and benefited GPU and high-bandwidth memory suppliers.
The next phase is expected to be driven more by capacity expansion and volume growth as AI adoption spreads to a broader base of users and applications.
Investors should focus on bottlenecks across compute, storage, optical, power and equipment, while tracking orders, customer qualification, equipment installation, utilisation and delivery lead times, Kenanga said.
“AI exposure” alone is increasingly insufficient to justify premium valuations, as investors shift their focus towards execution and earnings conversion, Kenanga said.
Huatai’s preferred principle is that capacity expansion is a stronger indicator of structural demand than price increases.
Optical connectivity could become increasingly important as AI clusters require greater bandwidth and lower latency to connect thousands of accelerators, potentially boosting demand for optical modules and silicon photonics.
Power is also emerging as a constraint, with electricity availability potentially determining where AI data centres are built and how quickly they can expand.
Advanced packaging is another structural opportunity as AI processors become larger and require tighter integration between logic, memory and interconnects, the research house said.
For Malaysia, Kenanga said semiconductor equipment, advanced packaging, outsourced semiconductor assembly and test, photonics, data-centre infrastructure, and power-related supply chains could benefit from AI-driven capacity expansion.
Kenanga maintained a neutral stance on Malaysia’s technology sector, favouring companies with visible volume growth, improving utilisation and earnings conversion rather than AI exposure alone.
The research house highlighted optical component maker NATGATE, citing global demand for optical transceivers that continues to exceed supply.
Its planned production start in November 2026 and subsequent ramp-up are key milestones, with higher volumes and utilisation potentially supporting earnings and margins in 2027.