Announcements of new semiconductor plants have arrived steadily since the pandemic shortages, backed by public money in the United States, Europe, Japan and India. Construction, however, is slower than politics.
Leading-edge logic remains overwhelmingly concentrated in a handful of facilities. Packaging and testing, the less glamorous end of the chain, is even more concentrated and receives a fraction of the subsidy attention.
Cost is the constraint that policy has not solved. A fab built for strategic reasons still has to compete commercially once it opens, and several projects have quietly slipped their timelines.
Diversification is real but partial. On current trajectories the industry will be less fragile by the end of the decade, not immune.
