Real-world scenario: A Japanese maker of chip-packaging substrates — sitting between wafer fabs and OSAT shops — needs a new Asian plant. Finalists: Taiwan (next to the world’s largest fabs, the most complete semiconductor cluster on earth) or Vietnam (lower cost, near shifting OSAT/assembly shops). The board asks the question no TCO spreadsheet can answer: “If the supply chain breaks, which plant keeps us alive?”
This is a risk-management and business-continuity (BCM) problem, not a cost problem. In the semiconductor supply chain — where one bad substrate lot can idle a packaging line worth millions of dollars a day — the right question is not “where is cheaper” but “where is total risk lower, and when risk materialises, where do we recover faster”. This article presents a four-group risk framework for scoring both locations, with Vietnam’s legal framework updated to 18 October 2026. The Taiwan section is framework-level; local counsel should confirm details.
Four risk groups to score
Group 1 — Natural-disaster and physical-infrastructure risk. Taiwan sits on the Pacific Ring of Fire — its fabs have been built to strict seismic standards for decades, but quake-induced production disruption is real and has happened. Vietnam has fewer earthquakes but faces typhoons, central-region flooding and extreme heat affecting cleanroom cooling. Score not “which place has fewer disasters” but: what your plant is designed to withstand, and what insurance covers. Business-interruption (BI) insurance for a semiconductor plant carries materially different premiums and exclusions between the two locations — take insurance quotes as an independent “risk meter”.
Group 2 — Geopolitical and export-control risk. The semiconductor chain sits at the centre of US–China technology competition. Taiwan is at the epicentre of that tension: US and allied export controls on semiconductor equipment, materials and technology can change scope, directly affecting a Taiwan plant’s ability to import machinery and export products. Vietnam holds a more neutral position in the global supply chain, but is not immune: goods made in Vietnam and exported to export-control jurisdictions must still comply with the technology-origin country’s rules (e.g. direct-product rules for US technology). Every semiconductor plant needs an export-compliance programme — partner screening, control-code classification, licence management — wherever it sits.
Group 3 — Concentration risk (single points of failure). Placing the plant next to the world’s largest fabs in Taiwan gives unbeatable logistics and engineering-loop speed — but creates a geographic single point of failure: one event hitting Taiwan strikes your suppliers and customers simultaneously. Placing it in Vietnam diversifies geography but creates a different single point of failure: dependence on a thin bench of specialised material suppliers (see the OSAT ecosystem analysis). The current playbook of large groups is “China + 1” evolving into “Taiwan + 1” — keep the foot in the main cluster, open a second foot where geographic risk diversifies.
Group 4 — Operational and legal risk. This covers: stability of incentive policy (today’s incentive reviewed tomorrow), permitting speed, contract enforcement, and IP protection. Vietnam has materially improved its investment framework (Investment Law 143/2025/QH15), but courts and IP enforcement remain areas technology investors assess cautiously. Taiwan is generally regarded by investors as having the more mature legal system and IP protection — with correspondingly higher compliance and operating costs.
Vietnam legal framework for a semiconductor supply-chain plant
Investment: IRC under Investment Law 143/2025/QH15 (1 March 2026) and Decree 96/2026/ND-CP; enterprise registration under Decree 168/2025/ND-CP. Semiconductor material/component manufacturing is high-tech — check the specific incentive lists; and model the global minimum tax (Resolution 107/2023/QH15 from fiscal year 2024; Decree 236/2025/ND-CP effective 15 October 2025) for in-scope groups.
Tax: 20% CIT under Law 67/2025/QH15 (1 October 2025); incentives by location/sector. Dividends to foreign organisations: domestic law imposes 0% withholding at source — important when structuring cash flows back to the Japanese parent (combined with the Vietnam–Japan double-tax treaty for overall optimisation).
Land: IZ sub-lease under Land Law 31/2024/QH15 (1 August 2024) via Decree 35/2022/ND-CP. Semiconductor material plants need large footprints and compliant chemical warehousing — verify the IZ’s fire-protection and environmental capacity before signing.
Customs and exports: Customs Law 54/2014/QH13. A semiconductor-specific note: certain materials, equipment and technologies may fall under export controls of the technology-origin country — build partner, end-use and end-user screening into day-one processes, not when the destination country’s customs starts asking.
Digital industry: Digital Technology Industry Law 71/2025/QH15 (1 January 2026) — review applicability for semiconductor material projects.
Risk scoring matrix (template)
| Risk group | Taiwan | Vietnam |
|---|---|---|
| Natural disaster | Earthquakes (high resilience standards); BI premiums reflect the risk | Typhoons/floods/heat; cleanroom design must include cooling redundancy |
| Geopolitics & export control | Epicentre of tech competition; directly hit when control scope changes | More neutral position; still bound by technology-origin export controls |
| Geographic concentration | Single point of failure: fabs + suppliers + customers on one island | Geographic diversification; single point of failure shifts to specialised material chains |
| Operations & law | Mature legal/IP system; higher cost | Sharply improved investment framework; IP enforcement needs cautious assessment |
How to use the matrix: do not add scores mechanically. Identify “which risk your company cannot bear” (risk appetite) — e.g. if customer contracts penalise delivery disruption heavily, Groups 1 and 3 carry the highest weight; if the product uses tightly-controlled US technology, Group 2 decides.
Business continuity: required wherever you choose
Dual-sourcing for critical materials: every critical material needs at least two customer-approved sources — one in the main cluster, one at the diversified location.
Strategic inventory: for long-lead-time, geopolitically risky materials, hold 8–12 weeks of safety stock instead of pure just-in-time. Inventory cost is the cheapest “insurance premium”.
Scenarios and drills: at least 2 scenarios per year (logistics disruption, wide-area power loss, export-control change) with drills and minutes — every fabless customer asks for them.
Semiconductor supply-chain companies evaluating sites in Vietnam and needing a review of the investment–tax–customs framework plus an export-control compliance programme can speak with our team.
Frequently asked questions
1. Does “Taiwan + 1” mean leaving Taiwan?
No. It means keeping the foot in the main cluster (engineering loops, complete supply chain) while opening a second foot where geographic risk diversifies. The two feet complement, not replace, each other.
2. Must a Vietnam plant comply with US export controls?
If the product, equipment or technology is of US origin (or exceeds the controlled US-technology content threshold), US export-control rules can apply extraterritorially — even with the plant in Vietnam. That is why every semiconductor plant needs an export-compliance programme regardless of location.
3. What should we watch in BI insurance for a semiconductor plant?
Read the exclusions closely (which disasters are not covered), the waiting period, and the loss-calculation basis (gross profit vs revenue). The premium gap between two locations is the market pricing risk — use it as an independent input.
4. Which site-selection risk is most often missed for semiconductors?
Water and power: semiconductor material plants consume ultrapure water and stable power at scale. An IZ with “land available” does not necessarily have “power/water capacity available” — verify real technical infrastructure, not just the rent.
5. Are Vietnam’s semiconductor investment incentives stable?
Incentives attach to commitments (capital, technology, headcount) and can be reviewed. Model the narrowed-incentive scenario too — and read the maintenance conditions in the IRC, not just the advertised incentive level.
6. Should a material plant sit near fabs or near OSAT shops?
It depends on your chain position: the substrate maker in our scenario sits between fabs and OSAT — near fabs speeds the engineering loop, near OSAT enables just-in-time delivery. If forced to choose one, choose proximity to the customer who pays you directly.
This article has been quality-reviewed by the FLAW AI Legal Council