Taiwan plans to raise billions of dollars to develop industrial parks in the US that would build a base for a supply chain for Nvidia and other AI hardware makers, Bloomberg reported on 16 October 2026. “The initiative is starting with a complex in Texas that would cost as much as $4 billion to build,” Bloomberg wrote, citing Taiwan’s Minister of Economic Affairs Kung Ming-hsin.
The facilities would help smaller suppliers that are essential to the operations of TSMC, Nvidia’s main production partner, and of AI server assemblers including Foxconn to set up shop in the US, Kung said, according to Bloomberg.
Business Pill · SOVEREIGN AI
A one-minute explainer of sovereign AI: who controls the model, the data and the infrastructure underneath. It teaches the general idea only and says nothing about any organisation in this story.
The key insight: As we read it, Taiwan is offering to move the smaller layer of the AI hardware chain, not only its champions: Kung said the parks would help small and medium-sized suppliers that TSMC and Foxconn depend on, with loan guarantees and help with US officials, according to Bloomberg and Economic Daily News.
Update, 16 Oct 2026: Taiwan to tighten AI-chip export controls
Bloomberg also reported on 16 October that Kung told Bloomberg News Taiwan is preparing to tighten export controls on AI chips to align further with US restrictions, adding AI servers and cutting-edge semiconductors to its list of controlled items (Bloomberg, via The Straits Times). Once the measures take effect, he said, shipping restricted AI hardware to China will become a criminal offence under Taiwan’s Foreign Trade Act. “Going forward, some of our measures on controlled items will correspond to American ones because we have made quite some commitments in the Agreement on Reciprocal Trade,” Kung said in an interview on 12 October. He declined to say when the measures will be implemented.
Bloomberg notes that Taiwan does not currently treat unauthorised AI-chip exports to China as a crime. As we read it, the US parks and the export rules come from the same minister and the same trade deal: more of Nvidia’s supply chain moves into the US, and diverting its hardware to China becomes a crime in Taiwan.
Update, 16 Oct 2026: China proposes export controls on 23 AI technologies
The same day, China moved on its own side of the AI export question. In Announcement 2026 No. 48, published at 16:00 Beijing time on 16 October, the Ministry of Commerce and the Ministry of Science and Technology propose adding 40 technologies to China’s Catalogue of Technologies Prohibited or Restricted from Export: 23 in AI (covering large models, agents, computing clusters and information processing), 5 in biomedicine and 12 in photovoltaics. Technology already in the public domain through open-source releases or published academic papers is excluded. Public comments run from 16 October to 15 November. Until the revised catalogue takes effect, exporters must register on the ministry’s platform before exporting any listed technology, including exports that have started but are not yet complete. Registration covers transferring ownership abroad, licensing Chinese technology for use overseas, and a change of the owner or controller to a foreign entity.
The thresholds sit in an attachment to the notice. According to Bloomberg, the draft would restrict the weights, checkpoints, source code and architecture of models with at least two trillion parameters, such as Moonshot AI’s Kimi K3 or Alibaba’s Qwen3.8-Max, or ranking among the world’s top 10 on measures including performance, usage and safety (Bloomberg, via The Edge). Bloomberg reports that deploying a covered model overseas would count as a technology export, while deployments on overseas infrastructure owned by the Chinese provider could qualify for automatic licensing. Reuters reports the commerce ministry said access through APIs, websites, apps and software-as-a-service platforms is excluded, and that most AI technology would remain freely exportable (Reuters, via Devdiscourse).
As we read it, Taiwan is tightening the hardware route to China while Beijing puts its own frontier model weights under licence, and the API carve-out keeps Chinese model services sellable abroad.
Update, 17 Oct 2026: China’s State Council on ‘AI Plus’
A day after the export notice, Premier Li Qiang chaired a State Council executive meeting on 16 October that heard a report on the progress of the “AI Plus” initiative, according to Xinhua. The meeting called for work on key technological challenges, algorithmic innovation, “the coordinated development of computing capacity and electricity supply”, and more high-quality data.
It also called for “strengthening risk assessment to ensure AI safety”, with “differentiated measures tailored to different industries”. Xinhua’s account names no new measures, budgets or deadlines for AI, and does not mention the export-control draft.
Kung spoke to Bloomberg in an interview on 12 October, according to Economic Daily News, which published a Chinese-language account of the interview. In that account, Kung said each park would cost about $2 billion to $4 billion to build, and that the government plans more than one in the US.
The government is working with the Taiwan Electrical and Electronic Manufacturers’ Association, which would be responsible for raising the funds needed for development, according to the same account.
Kung said the government would support suppliers that want to move into the parks, including with credit guarantees for loans, and would help them deal with US local and federal officials on needs such as visas and land, Economic Daily News reported.
Each Taiwan-backed industrial park in the US would cost about $2 billion to $4 billion to build, with the first, in Texas, costing as much as $4 billion, according to Taiwan’s Minister of Economic Affairs Kung Ming-hsin in an interview with Bloomberg published on 16 October 2026.
Who Pays, and How
Bloomberg’s report describes the plan as Taiwan raising billions of dollars. In Economic Daily News’ account, the money for development is to be raised by the manufacturers’ association, with the government offering loan guarantees and help with officials, rather than a direct budget figure.
When Taiwan reached a trade agreement with the US this year, it committed to help provide $250 billion in credit guarantees to support further investment in the US semiconductor supply chain, according to the same report.
By our arithmetic, a $4 billion first park is 1.6% of that $250 billion guarantee commitment. That is our calculation; neither report links the two figures.
Why Smaller Suppliers
TSMC and Foxconn can raise funds for and manage their own US expansion, but small and medium-sized suppliers may not be able to, according to the Economic Daily News account. Taiwan hopes the new parks can help them, it says.
Kung said the scale of AI demand offers a large opportunity to move manufacturing back to the US. The business is profitable enough to help cover the high initial cost of building US plants, and costs fall as operations scale, he said, according to the same account.
He also said Taiwanese suppliers taking AI orders face heavy pressure on delivery times, so they are expanding capacity at home and abroad at the same time, Economic Daily News reported.
Texas, TSMC and the Land
Bloomberg, citing people familiar with the matter, reported that Taiwan has already secured land for a park in Texas, according to Economic Daily News. The Ministry of Economic Affairs did not confirm the details and said it would explain at an appropriate time, the report says.
TSMC said on Thursday that it is also evaluating a new US site in Texas, as we reported from its earnings call. Asked whether TSMC had chosen Texas, Kung said the company is evaluating many locations, according to the Economic Daily News account. TSMC declined to comment, it says.
Foxconn and Wistron have recently expanded in Texas, making the state a popular destination for Taiwanese companies, the report says. Kung also joked that Elon Musk is welcome to use the new supply chain Taiwanese companies build in the US, but that Taiwan would not transfer all of its technology to him, according to the same account.
Where Nvidia Sits
The plan comes in the same week that Nvidia chief executive Jensen Huang said the Nvidia gear in each gigawatt of AI data centre has gone from about $18 billion to over $40 billion, as we reported, and Nvidia set out plans to invest in five UK startups.
The Structural Read
As we read it, the money is structured as support rather than a budget line: in Economic Daily News’ account of the interview, the manufacturers’ association raises the development funds and the government offers loan guarantees, while Bloomberg describes Taiwan as planning to raise billions of dollars.
As we read it, the parks address the part of the chain that cannot finance itself: TSMC and Foxconn can fund and manage their own US expansion, but smaller suppliers may not be able to, according to the same account.
As we read it, the Texas choice follows where Taiwan’s large companies are already moving: Foxconn and Wistron have recently expanded in Texas, and TSMC said on Thursday it is evaluating a new US site there, while Kung said TSMC is evaluating many locations.
Bloomberg, 16 October 2026
“The initiative is starting with a complex in Texas that would cost as much as $4 billion to build”
Three Implications
SMALLER TAIWANESE SUPPLIERS Kung said the government would support suppliers that move into the parks, including with loan credit guarantees and help with US officials on visas and land, according to Economic Daily News.
TSMC AND FOXCONN The parks are meant to help the smaller suppliers essential to TSMC, Nvidia’s main production partner, and to AI server assemblers including Foxconn set up in the US, Kung said, according to Bloomberg.
TAIWAN’S MINISTRY The Ministry of Economic Affairs did not confirm the reported Texas land and said it would explain at an appropriate time, according to Economic Daily News.
The framework starts from the claim that “The digital economy rests on a physical foundation that most people in that economy never see.”
As we read it, Taiwan’s parks are an attempt to move part of that foundation, the smaller suppliers behind TSMC and Foxconn, into the US with government guarantees. The framework predates the report and is about materials and processing; it does not discuss Taiwan’s plan.
Taiwan committed to help provide $250 billion in credit guarantees for US semiconductor supply-chain investment under its trade agreement with the US this year; the first Texas park would cost as much as $4 billion, according to Bloomberg and Economic Daily News, 16 October 2026. Shown for scale; the reports do not link the two figures.
What Is Not Established
In what we read, there is no Ministry of Economic Affairs release giving a site, a size, a timetable or the companies that would move in. The $2 billion to $4 billion range and the up-to-$4 billion Texas figure come from Kung’s interview, and Bloomberg describes the Texas cost as “as much as $4 billion.”
The land in Texas is reported by Bloomberg citing people familiar with the matter and was not confirmed by the ministry. Neither report says how much the manufacturers’ association would raise or how large the loan guarantees for these parks would be. We read Bloomberg’s first two paragraphs on bloomberg.com; the rest of the interview we read in Economic Daily News’ Chinese-language account, and the English wording here is our translation.
Business Engineer Framework
AI’s Geopolitical Chokepoint
A Business Engineer framework on the physical foundation under the AI economy, and who controls it.
Taiwan plans industrial parks in the US for the supply chain of Nvidia and other AI hardware makers, starting with a Texas complex that would cost as much as $4 billion, Minister of Economic Affairs Kung Ming-hsin told Bloomberg.
In Kung’s account, each park would cost about $2 billion to $4 billion, there would be more than one, and the government’s support would come as loan guarantees and help with US officials for the smaller suppliers that TSMC and Foxconn depend on.
A note on sourcing. We read the headline, byline and first two paragraphs of Bloomberg’s 16 October 2026 report on bloomberg.com; the rest is paywalled. The remaining details of Kung Ming-hsin’s interview come from Economic Daily News’ Chinese-language account of it, read in full; the English wording is our translation. The land in Texas is reported by Bloomberg citing people familiar with the matter and was not confirmed by Taiwan’s Ministry of Economic Affairs. The 1.6% is our arithmetic. The TSMC, Nvidia and UK references link to our own earlier reporting.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
![Value Chain Analysis vs Supply Chain: Key Differences & When to Use Each [2026]](https://i0.wp.com/fourweekmba.com/wp-content/plugins/contextual-related-posts/default.png?resize=150%2C150&ssl=1 "Value Chain Analysis vs Supply Chain: Key Differences & When to Use Each [2026]")
