Industry Analysis
The AI compute boom is triggering a deep restructuring of the semiconductor stack: sub-3nm nodes reliant on EUV lithography have cemented TSMC (Taiwan, China) and Samsung’s duopoly in advanced foundry, while HBM memory and CoWoS packaging emerge as critical bottlenecks for data movement—favoring Micron and SK Hynix. Geopolitical policies, notably the U.S. CHIPS Act and export controls, have materially increased supply chain costs and forced ‘friend-shoring,’ extending equipment lead times by 12–18 months. In response to NVIDIA’s dominance in AI training, Intel and AMD are aggressively pushing Gaudi and MI300X chips, betting on cost efficiency and open ecosystems. Over the next 18 months, ETF inflows will concentrate on infrastructure enablers, but outsized returns will accrue to niche leaders in advanced packaging, high-speed interconnects, and domain-specific IP—not broad-based ‘chip plays.’
This page displays AI-generated summaries and metadata for research purposes. Original content belongs to the respective publishers.