Nvidia Will Pour $150 Billion a Year Into Taiwan, the “Epicentre” of the AI Revolution — and Build a Taipei HQ

Nvidia chief executive Jensen Huang said the company will spend around $150 billion a year in Taiwan, framing the island as the indispensable hub of the global AI build-out. Speaking on May 27 at an all-employee celebration in Taipei, Huang told staff that “Taiwan is the epicentre of the AI transformation,” adding: “This is where the chips come, product packaging comes, this is where the systems are made, this is where AI supercomputers are produced.”

The headline figure marks a roughly tenfold jump in a handful of years. “Four years, five years back, Nvidia was investing about $10, $15 billion dollars a year in Taiwan,” Huang said. “Now we’re investing $100, going to $150 billion dollars in Taiwan each year.” Huang did not say how long Nvidia expects to sustain spending at that level, but the scale underscores how dependent the company — now worth around $5 trillion — has become on Taiwan’s manufacturing base as demand for AI accelerators keeps outrunning supply.

Anchoring the commitment is a new Taiwan headquarters. The campus is set to break ground this year and become operational by 2030, eventually housing roughly 4,000 employees, according to reporting from Reuters and CNBC. The location is no accident: it places Nvidia closer to TSMC, the world’s largest contract chipmaker and the company that fabricates the advanced GPUs at the heart of every Nvidia AI system. Taiwanese chip stocks climbed on the news.

The announcement also lands squarely in the middle of one of the most sensitive fault lines in technology policy. Taiwan produces the bulk of the world’s most advanced semiconductors, and Washington has spent years pressing to diversify that supply chain amid concerns about cross-strait tensions with Beijing. Nvidia deepening — rather than hedging — its physical footprint on the island is a vote of confidence in Taiwan’s centrality, even as governments elsewhere subsidize domestic fabs to reduce their reliance on it.

For Huang, the message to employees was as much about identity as economics: Nvidia’s fortunes and Taiwan’s are now tightly bound. The pledge follows a string of mega-scale infrastructure commitments across the industry — from multibillion-dollar cloud and compute deals to nation-level data-center plans — that together signal a phase shift. The constraint on AI is no longer ideas or models; it is the chips, packaging and power needed to run them, and the companies racing to lock up that capacity years in advance.