Supply Chain Bottlenecks

TSMC's CoWoS packaging remains the critical bottleneck for high-end AI accelerators. Despite significant capacity expansion through 2024, the shift to next-generation architectures like NVIDIA's Blackwell (utilizing CoWoS-L) and AMD's Instinct MI325X maintains a supply-demand gap. Availability is governed by packaging throughput rather than front-end wafer fabrication.

Leading-edge manufacturing capacity (3nm and the upcoming 2nm) remains highly concentrated. TSMC holds over 90% of the advanced foundry market share. The bottleneck is exacerbated by the simultaneous demand for AI accelerators and mobile SoCs, while 2nm capacity is already being pre-booked for 2025-2026 production.

US export controls and prioritization policies restrict high-end AI GPU shipments. Affects NVIDIA, AMD, and downstream AI infrastructure globally. China restrictions and retaliatory import bans compound supply uncertainty.

The 300mm wafer market has largely stabilized following 2023-2024 inventory adjustments. Supply remains concentrated among five major players, led by Shin-Etsu and SUMCO. While structural capacity for high-end epitaxial wafers used in AI and advanced logic remains tight, general availability has improved. Long-term agreements (LTAs) ensure stability, though the industry remains sensitive to long lead times (2-3 years) for future greenfield capacity expansions. Severity: 2/5.

Glass substrates are emerging as replacements for organic substrates in advanced packaging. Required for next-gen chiplets and high-density interconnects. Production capacity is extremely limited with Intel, Samsung, and substrate makers racing to scale.

Standard DRAM production is concentrated in Samsung, SK Hynix, and Micron. While less acute than HBM constraints, DRAM supply remains cyclical with capacity additions requiring 18-24 month lead times. Server DRAM demand continues growing with AI workloads.