Jensen Huang stood before roughly 1,000 Nvidia employees at the Beitou-Shilin Technology Park in northern Taipei on Wednesday and announced that the company now spends $100 billion per year in Taiwan and is on a direct path to $150 billion — a 10-fold increase from the $10 to $15 billion Nvidia committed to the island just four or five years ago.

The setting was the official launch celebration for Nvidia Constellation, the company’s first overseas headquarters. Under a 50-year lease signed with the Taipei city government in February 2026 — extendable by another 20 years — Nvidia secured plots T17 and T18 of the Beitou-Shilin Technology Park for a campus investment exceeding NT$40 billion (approximately $1.27 billion USD). Construction is scheduled to begin in June or July 2026, with full operations targeted by 2030. According to plans submitted to the city government, the campus is expected to house roughly 4,000 direct Nvidia employees and generate more than 10,000 jobs in total once operational. Taipei Mayor Chiang Wan-an, who attended the celebration and presented Huang with a key to the city, called the agreement “a major milestone for both Taipei and Taiwan.”

“Taiwan is the epicenter of the AI revolution,” Huang told the crowd. “This is where the chips come, packaging comes. This is where the systems are made. This is where AI supercomputers were created.”

Why Nvidia Is Planting Its Flag — and What Taiwan Controls

Nvidia’s deepening commitment is inseparable from the geography of advanced semiconductor manufacturing. Taiwan’s TSMC commands roughly 72% of the global pure-play foundry market, according to Counterpoint Research — a concentration that has only grown more strategically significant as hyperscalers and sovereign AI programs compete for allocation.

TSMC is the sole manufacturer capable of producing Nvidia’s most advanced processors at scale. During his Taiwan visit, Huang met with TSMC Chairman C.C. Wei to discuss production commitments for Nvidia’s next-generation Vera Rubin platform, the chip architecture Huang has described as “the largest product launch, probably in the history of Taiwan.” Beyond TSMC, Nvidia’s Taiwan AI chip supply chain includes server assemblers Foxconn, Wistron, and Quanta Computer, along with advanced packaging firms that provide the chip interconnect technology making high-performance AI accelerators possible.

By constructing a permanent headquarters campus inside this ecosystem, Nvidia shortens engineering decision loops, deepens partner collaboration, and reinforces its position at a supply chain chokepoint no rival has yet been able to replicate at comparable scale elsewhere. Huang is scheduled to deliver the opening keynote at Computex Taipei 2026 on June 1, appearing alongside Marvell Technology CEO Matt Murphy to discuss next-generation AI infrastructure.

$150 Billion Conviction as China Revenue Vanishes

The scale of Nvidia’s financial commitment is especially striking given what is missing from its revenue picture. The company’s latest quarterly guidance of approximately $91 billion for Q2 fiscal 2027 — its own record — assumes zero data center compute revenue from mainland China, after U.S. export restrictions eliminated Nvidia’s ability to ship Hopper-class chips to the country. Revenue from China’s data center segment dropped from $4.6 billion in the same quarter a year earlier to nothing.

Taiwan filled much of the gap. Revenue from the island surged more than 50% year-over-year in the most recent quarter, even as the China business collapsed. Taiwan’s Taiex stock index reflected Wednesday’s announcement immediately, climbing 1.7% to a record close — with semiconductor and supply-chain names leading the advance. South Korea’s SK Hynix and U.S. memory maker Micron both crossed the $1 trillion market-capitalization threshold on the same day, underscoring the breadth of the AI hardware rally.

Huang did not specify how many years Nvidia intends to sustain the $150 billion annual pace. But a 50-year land lease and a purpose-built headquarters campus are not hedged bets: they are structural commitments that signal the company views the Taiwan relationship as a permanent strategic anchor, not a production-cycle convenience.

Physical AI and the Manufacturing Thesis

Huang used the Taipei stage to articulate a broader argument about the next phase of AI’s expansion. He described the convergence of AI with physical hardware — what he called “physical AI” — as a force that will transform manufacturing itself. Taiwan’s ecosystem, he argued, makes it uniquely positioned to lead that transformation: the island combines foundry capacity, advanced chip packaging, systems integration, and AI server assembly in a geographic footprint no other region currently matches.

The remarks land as Nvidia navigates a complicated regulatory environment on multiple fronts. Earlier in May, Taiwanese prosecutors executed Taiwan’s first-ever criminal enforcement action against illegal AI hardware exports, raiding 12 locations and detaining three individuals — including Supermicro co-founder Wally Liaw — over allegations of using forged documents to ship Nvidia-equipped servers to China. Huang, who arrived in Taiwan three days after the raid, commented publicly that Supermicro needs to “enhance and improve their regulation compliance.”

Jensen Huang Taiwan Roots and a Billion-Dollar Homecoming

Huang added a personal dimension to Wednesday’s event. Born in Tainan in 1963, he emigrated to the United States at age nine and has long held something close to rockstar status among Taiwanese tech workers and investors. His parents, wife, daughter, and son were present for the Constellation celebration. When Taipei Mayor Chiang Wan-an presented him with a key to the city and a calligraphy scroll, the gesture carried a weight that went beyond diplomacy: Nvidia is committing to Taiwan not just with capital but with a permanent address.

Industry-Wide Convergence on One Exposed Island

Nvidia is not alone in concentrating resources on Taiwan. Advanced Micro Devices announced on May 21 that it would invest more than $10 billion in Taiwan’s AI sector — its largest-ever commitment to the island — to expand advanced packaging capacity for its upcoming Helios AI server platform and deepen partnerships with firms including ASE Technology Holding, Powertech Technology, Sanmina, and Inventec.

The cumulative effect of these commitments — from both the dominant and the principal challenger in AI chip supply — further cements Taiwan’s structural role in the global technology stack. Yet that same concentration is what financial analysts and supply-chain researchers have identified as the industry’s most consequential unresolved risk. TSMC’s geographic diversification into Arizona, Japan, and Germany is real but not yet at commercial scale; none of those facilities are expected to match Taiwan’s leading-edge output capacity before 2028 at the earliest. Meanwhile, Taiwan imports 97% of its energy, and any sustained disruption to energy supply routes would raise chip costs and slow AI infrastructure deployment globally.

When the Constellation campus opens in 2030, it will be the most visible symbol yet of a supply chain that now has no credible alternative geography — and of the geopolitical concentration risk that comes with it. Investors, supply-chain professionals, and policymakers weighing Taiwan’s role in global AI infrastructure now have a 10-fold spending surge, a 50-year lease, and a permanent headquarters to measure that risk against.

Frequently Asked Questions

Why is Nvidia investing $150 billion a year in Taiwan?

Taiwan controls roughly 72% of the global pure-play foundry market, according to Counterpoint Research, and is home to TSMC — the sole manufacturer capable of producing Nvidia’s most advanced AI chips at scale. The island also hosts the packaging facilities, server assemblers, and supply-chain partners Nvidia needs to build complete AI systems. The $150 billion annual figure reflects what Nvidia already spends across this entire ecosystem, with current spending at $100 billion and rising — not a new standalone capital outlay.

What is the Nvidia Constellation headquarters campus?

Constellation is Nvidia’s first overseas headquarters, located in the Beitou-Shilin Technology Park in northern Taipei. Nvidia signed a 50-year land lease with the Taipei city government in February 2026 for a campus investment exceeding NT$40 billion (approximately $1.27 billion USD). The project is expected to break ground in June or July 2026 and open by 2030, housing roughly 4,000 direct Nvidia employees and generating more than 10,000 total jobs according to plans submitted to the city.

How does Taiwan’s semiconductor dominance affect the global AI industry?

Because Taiwan manufactures the overwhelming majority of the world’s most advanced chips — and because TSMC’s geographic diversification into Arizona, Japan, and Germany is not expected to reach leading-edge scale before 2028 — any supply disruption on the island would halt the production of AI accelerators globally. That concentration is the central supply-chain risk for every company that depends on AI hardware, from hyperscalers to sovereign AI programs, and it deepens each time a company like Nvidia or AMD adds a permanent headquarters or multi-billion-dollar commitment to the island.

How much Nvidia revenue comes from Taiwan vs. China?

Nvidia guided Q2 fiscal 2027 revenue at approximately $91 billion — a record — while assuming zero data center compute revenue from mainland China, where U.S. export restrictions have eliminated Hopper-class chip sales. By contrast, Taiwan revenue surged more than 50% year-over-year in the most recent quarter. Nvidia’s Q1 fiscal 2027 revenue reached a record $81.62 billion, with data center revenue up 92% year-over-year at $75.2 billion.