South Korea committed $3 billion in semiconductor investment incentives this week while India accelerated chip manufacturing buildouts and TSMC disclosed its Arizona footprint will reach $265 billion across twelve facilities. The announcements arrived within seventy-two hours of each other. The pattern is geographic hedging at sovereign scale.
The South Korean package targets battery and biotech alongside chips, but the semiconductor allocation represents the hard core—direct subsidies for domestic fab capacity in a country already home to Samsung's advanced nodes. India's expansion, less publicized in dollar terms, focuses on assembly and test capacity rather than leading-edge fabrication. TSMC's Arizona disclosure updates a commitment first announced at $40 billion, then raised to $100 billion in recent capex guidance, now formalized at $265 billion through build-out completion. The Arizona site will manufacture 3-nanometer and 2-nanometer process chips, the same nodes currently reserved for Apple and Nvidia orders out of Taiwan.
This is not competition. It is redundancy by design. South Korea hedges against Taiwan risk. India builds the back-end that used to live in Malaysia and Vietnam. TSMC builds a second Taiwan inside the United States, subsidized by CHIPS Act funds that now look modest against the total outlay. The capital intensity alone—TSMC's $265 billion represents more than the market cap of Intel—signals that leading-edge semiconductor production has become a sovereign asset class, priced and funded like defense infrastructure rather than consumer electronics supply chain. Family offices with exposure to applied materials manufacturers, lithography tooling, or rare-earth processing should note the margin implications: when three governments independently fund overlapping capacity, equipment suppliers capture the spend without the demand risk.
The timing matters. TSMC's capex expansion comes as the company already runs fabrication lines at capacity, unable to satisfy existing orders for AI accelerators and mobile processors. Building twelve Arizona facilities does not solve the 2025 bottleneck; it solves the 2030 scenario where US-China trade restrictions, Taiwan Strait tensions, or simple logistics fragility make single-source foundry exposure unacceptable to Western OEMs. South Korea's move, structured as incentives rather than direct state investment, allows Samsung to expand without appearing to chase TSMC's Arizona model. India's play remains the labor-cost arbitrage on assembly, but with the added benefit of non-China sourcing for firms facing Western export controls.
Allocators should track three follow-on events. First, whether TSMC's $100 billion annual capex guidance—already disclosed in recent earnings—gets revised upward again in the January 2026 report, signaling Arizona cost overruns or accelerated timelines. Second, Samsung's formal response to South Korea's $3 billion incentive, expected before year-end, which will clarify whether the funds flow to advanced-node expansion or legacy-process domestic production. Third, India's Ministry of Electronics and IT typically publishes facility approval lists in quarterly cycles; the next update in October will show which foreign partners committed capital beyond assembly into packaging and testing.
TSMC's Arizona build now costs more than the company's entire market cap a decade ago. South Korea and India are not competing with that. They are making sure no single earthquake, blockade, or policy shift can take the whole board offline.