Manufacturing Geopolitical Risk Turns Every Plant Decision Into a Timing Bet

TL;DR: Manufacturing geopolitical risk has turned plant siting into a timing decision as much as a location decision. Companies that commit capital early can lock in tariff relief or subsidy windows before they close, but 2025 and 2026 show that both commitment and delay carry a real, quantifiable cost.

Key takeaways:

• Volkswagen shelved plans for a second US Audi plant in January 2026 after CEO Oliver Blume said US tariffs cost the company roughly 2.5 billion dollars in the first nine months of 2025 alone.
• GM committed 4 billion dollars in June 2025 to expand gas-vehicle production at three US plants, explicitly to avoid tariff costs, while keeping electric versions of the same vehicles in Mexico.
• Battery maker AESC paused construction on its South Carolina plant in June 2025 after already spending roughly 1 billion of a planned 1.6 billion dollars, citing policy and market uncertainty.
• The US government converted 8.9 billion dollars of Intel's promised CHIPS Act cash grants into a government equity stake in August 2025, a structure change that did not apply to every recipient.
• TSMC raised its total Arizona investment to 265 billion dollars in July 2026, tied to a US-Taiwan trade deal that cut US tariffs on Taiwanese goods from 20 to 15 percent.

In the same week of January 2026, Volkswagen shelved a planned second US Audi plant and GM was seven months into a 4 billion dollar bet on expanding three other US plants. Both companies were reading the same tariff and geopolitical risk environment. They reached opposite conclusions about whether to commit capital now or hold it back, and both decisions were defensible.

Commit or Wait: The Numbers From 2025

The pattern is not one company hesitating while its competitors move. It runs in both directions across the same period.

• Volkswagen paused rather than committed. CEO Oliver Blume said Trump-era tariffs cost the company roughly 2.5 billion dollars over the first nine months of 2025, and Volkswagen said in January 2026 it may cancel a planned second US Audi plant without tariff relief, despite already owning land near its Tennessee plant.[^1]
• AESC paused after sinking real capital in. The battery maker halted construction on its Florence, South Carolina plant in June 2025, after spending roughly 1 billion of a planned 1.6 billion dollars, saying explicitly that the pause was "due to policy and market uncertainty."[^2]
• BYD is holding a decision open rather than closing it. The company has delayed finalizing a Mexico plant, with a vice president saying the plan "remains in place" pending clarity on proposed Mexican tariffs of up to 50 percent on vehicles from countries without a Mexican trade agreement.[^3]
• The pullback runs wider than any single company. Clean-energy manufacturing project cancellations and downsizing totaled 22 billion dollars in the first half of 2025 alone, affecting 35 projects and an estimated 16,500 jobs, according to the independent tracker E2.[^4]

The Real-Options Case for Waiting

None of this is new in kind, even if the current wave of uncertainty is unusually fast-moving. A 1994 paper by Bruce Kogut and Nalin Kulatilaka modeled a multinational manufacturing network explicitly as a portfolio of options: the value of a plant is not just the output it produces, but the flexibility it buys to shift production somewhere else if conditions change.[^5] A separate 1996 paper by Pietra Rivoli and Eugene Salorio applied the same logic directly to the entry-timing decision, treating "commit now" and "wait" as two different bets on how policy uncertainty will resolve.[^6]

GM's June 2025 commitment is the clean current example of buying the option rather than betting the whole position. The company put 4 billion dollars into shifting gas-vehicle production for the Chevrolet Equinox, Blazer, and full-size trucks to three US plants, explicitly to avoid tariff costs, while leaving the electric versions of the same vehicles in Mexico.[^7] That is not full commitment and it is not full flexibility. It is GM paying for tariff protection on the part of its lineup it judged most exposed, while keeping the rest of its footprint free to move again if the policy shifts back.

What the record shows: companies making the largest, fastest capital commitments in 2025 and 2026 tied them explicitly to a specific, named policy trigger, not to a general read of the business climate. What we infer: the size of the commitment tends to track the size of the policy window being locked in, not the size of the company. What we recommend: a plant decision framed as "commit or wait" is usually the wrong framing; the real question is which portion of the investment to commit now against a specific trigger, and which portion to leave flexible.

TSMC's 265 billion dollar Arizona commitment, raised in July 2026, is the same logic at a different scale. The increase followed a January 2026 US-Taiwan trade framework in which Taiwan pledged at least 250 billion dollars in direct US investment in exchange for the US cutting reciprocal tariffs on Taiwanese goods from 20 to 15 percent.[^8] TSMC did not commit that capital because Arizona became a better manufacturing location in six months. It committed because a specific tariff number was on the table and would not stay there indefinitely.

Building the Trigger List Before the Deadline

The part of this picture that gets less attention is what happens after a commitment is made: a signed subsidy deal is not a guaranteed one. In August 2025, the US government converted 8.9 billion dollars of Intel's remaining CHIPS Act funding, money originally structured as a cash grant, into a roughly 10 percent government equity stake instead.¹ Samsung's Taylor, Texas fab has slipped past its original 2024 target with 2026 now cited as the newest completion date, even though its 6.6 billion dollar CHIPS subsidy was finalized in December 2024.² TSMC's own CHIPS grant, finalized the same month as Samsung's, has not been renegotiated the same way, which is the useful detail: the uncertainty is deal-specific, not a blanket reversal of every subsidy commitment made under the prior administration.³

That distinction is the practical takeaway for a company sitting on a footprint decision today. Waiting for perfect clarity is not a real option; clarity has not been the normal condition these three years. What is available instead is a specific list: which subsidy programs have already been renegotiated for at least one recipient, which tariff numbers are tied to a dated trade framework rather than a standing rule, and which competitors have moved first and on what trigger. Volkswagen's land near its Tennessee plant is still sitting there. So is BYD's stated intent to build in Mexico. Neither company gave up the option. Both are waiting for a specific number to move, not for the environment to feel calmer.

A footprint decision made without naming the trigger it is waiting on, or the trigger it is committing against, is not actually a decision. It is a default, and 2025 showed that a default can cost a company a plant, a market, or, in Volkswagen's case, 2.5 billion dollars before the plant was ever built.

Fortius Intel note: Name the specific number before committing or waiting: the tariff rate, the subsidy deadline, the contract clause. GM, TSMC, and AESC all made their decisions against a named trigger. The companies that got caught flat-footed in 2025 were the ones treating "uncertainty" as a single condition rather than a specific date on a specific policy.

This analysis draws on Volkswagen, GM, and AESC corporate statements as reported by CBT News, Detroit News, Axios, and MarketBeat, US government CHIPS Act actions as reported by Manufacturing Dive and Tom's Hardware, TSMC's own investment announcements, and the academic real-options literature on multinational manufacturing networks, cross-checked across at least two independent sources for the Volkswagen, GM, and Intel figures.

---

Footnotes

Footnotes

1. "US government takes roughly 10 percent stake in Intel, converting 8.9 billion dollars in CHIPS Act funding," Manufacturing Dive, 23 August 2025. https://www.manufacturingdive.com/news/us-government-10-percent-stake-intel-chips-funding-8-9-billion/758518/. Retrieved 24 August 2026. Source for: the August 2025 conversion of Intel's remaining CHIPS Act funding from a cash grant into a government equity stake. ↩

2. "Samsung delays 44 billion dollar Texas chip fab, sources say completion halted because there are no customers," Tom's Hardware, 2025 and 2026 reporting. https://www.tomshardware.com/tech-industry/semiconductors/samsung-delays-usd44-billion-texas-chip-fab-sources-say-completion-halted-because-there-are-no-customers. Retrieved 24 August 2026. Source for: repeated delays to Samsung's Taylor, Texas fab despite its 6.6 billion dollar CHIPS Act subsidy being finalized in December 2024. ↩

3. TSMC Corporate Press Release Archive and National Institute of Standards and Technology CHIPS Program Office. https://pr.tsmc.com/english/news/3122 and https://www.nist.gov/chips/tsmc-arizona-phoenix. Retrieved 24 August 2026. Source for: the finalization of TSMC's original Arizona CHIPS Act award in November 2024 and its continued, unrenegotiated status as TSMC's own investment expanded. ↩