Can China repeat its electric vehicle leapfrog with semiconductors? - The World Economic Forum

The EV leap was systems-integration innovation; semiconductors are a full-stack vertical chain where photolithography physics, EDA algorithms, and material chemistry are interlocked—break one link and the entire chain fails. SMIC pushing 14nm via DUV multi-patterning is essentially 3-4 exposures simulating one EUV pass; yield loss and cost inflation are governed by physics, not engineering will. In EDA, Empyrean handles analog ICs competently, but digital front-end and P&R still trail Synopsys and Cadence by a generation. Verdict: 28nm-and-above mature nodes will be largely self-sufficient by 2025; sub-7nm remains bottlenecked by equipment and materials within a 24-month window. The long-tail of the Oct 2022 US export controls is now biting—compliance scrutiny on ASML spares, Shin-Etsu silicon wafers, and ArF photoresists will inflate industry procurement costs by 15-25%. TSMC's Arizona and Kumamoto expansions are geographic redundancy plays to offset China's mature-node capacity flood—China's 28-90nm share will exceed 30% of global output by 2025, triggering a price war. Core judgment: China will not leapfrog in leading-edge nodes, but will build structural dominance in mature nodes plus SiC/GaN specialty processes. The real variable is the industry paradigm shift from 'efficiency-optimal' to 'security-redundant'—once that shift hardens, the 'good-enough autonomy' route gains institutional headroom.