Nvidia’s US$150 Billion Taiwan Spending Plan Shows Where the AI Supply Chain Still Runs
Nvidia Corporation (NASDAQ: NVDA) is planning to spend around US$150 billion a year in Taiwan, according to comments from chief executive Jensen Huang at a Taipei event on 27 May 2026. Huang said Nvidia’s annual Taiwan spending had risen from about US$10 billion to US$15 billion several years ago to around US$100 billion now, with the figure moving toward US$150 billion each year. Reuters also noted that Huang did not give a timeframe for how many years Nvidia expects that spending level to continue.
The number matters because Taiwan sits at the centre of Nvidia’s artificial intelligence supply chain. Taiwan Semiconductor Manufacturing Co (TWSE: 2330; NYSE: TSM) makes many of the advanced chips used in AI systems, while Foxconn, Wistron and Quanta Computer are key manufacturing partners for AI servers and related infrastructure.
For investors watching Nvidia, the announcement is less about a one-off capital spend and more about supply-chain depth. The AI boom is no longer just a software story. It depends on advanced chips, packaging, server assembly, power availability and manufacturing partners that can scale quickly.
What Nvidia actually said in Taiwan
Huang made the comments at a launch celebration for Nvidia’s planned Taiwan headquarters, which Reuters reported is expected to break ground this year and become operational in 2030. The new site is expected to employ around 4,000 people.
He called Taiwan the “epicentre” of the AI revolution, pointing to the island’s role in chips, packaging and systems manufacturing. That framing is useful because it shows how Nvidia sees the region: not just as a supplier base, but as a full AI infrastructure ecosystem.
That distinction matters. Nvidia designs the chips, but it depends on partners to manufacture, package and assemble the systems that hyperscalers and enterprises buy. As demand for AI infrastructure keeps rising, the ability to secure capacity from those partners becomes a major strategic issue.
How Nvidia’s Taiwan plan compares with AMD
The timing also puts Nvidia’s comments next to a recent move from Advanced Micro Devices, Inc. (NASDAQ: AMD). On 21 May 2026, AMD said it would invest more than US$10 billion in Taiwan’s AI supply-chain ecosystem, including partnerships in advanced packaging, substrates and server manufacturing. The plan is tied partly to AMD’s upcoming Helios rack-scale AI platform, expected to begin deployment in the second half of 2026.
The comparison is useful, but it should not be overstated. AMD’s figure is a multi-year investment plan. Nvidia’s figure, as described by Huang, is an annual spending run-rate moving toward US$150 billion. That makes the two numbers different in both scale and structure.
| Company | Taiwan exposure described | Scale | Main focus |
|---|---|---|---|
| Advanced Micro Devices, Inc. (NASDAQ: AMD) | Multi-year AI ecosystem investment | More than US$10 billion | Advanced packaging, substrates, server manufacturing, Helios AI platform |
| Nvidia Corporation (NASDAQ: NVDA) | Annual Taiwan spending plan | Around US$150 billion a year | Chip manufacturing, packaging, AI servers, infrastructure partners |
The better takeaway is not that Nvidia has “upstaged” AMD. It is that both companies are pointing to the same reality: Taiwan remains central to the next phase of AI hardware buildout.
Why Taiwan remains so important to Nvidia
Taiwan’s position starts with TSMC. TrendForce data cited by Taipei Times showed TSMC held 69.9% of the global foundry market in 2025, up from 64.4% in 2024. Samsung was a distant second at 7.2%.
That may sound technical, but the investor point is simple. Advanced AI chips are only useful if they can be produced, packaged and assembled at scale. Nvidia’s Taiwan spending shows how much of that physical work still runs through a relatively concentrated network of suppliers.
Nvidia’s own results explain why the supply chain is under pressure. The company reported US$81.6 billion in revenue for the first quarter of fiscal 2027, up 85% from a year earlier. Data Center revenue reached US$75.2 billion, up 92% year-on-year, while GAAP gross margin was 74.9%. Nvidia guided for second-quarter fiscal 2027 revenue of around US$91.0 billion, plus or minus 2%.
Those numbers make the Taiwan spending figure easier to understand. Nvidia is scaling at a pace where manufacturing access, packaging capacity and server production are not background issues. They are central to whether the company can keep meeting demand.
The bull case and bear case
The bull case is that Nvidia’s spending footprint could strengthen its relationships with the suppliers that matter most. If AI infrastructure demand remains strong, deeper ties with TSMC, Foxconn, Wistron and Quanta may help Nvidia protect access to manufacturing and assembly capacity. That would support the view that Nvidia’s competitive edge is not just in chip design, but in its ability to turn designs into full systems at global scale.
The bear case is concentration. Taiwan’s importance to Nvidia also means Taiwan is a key point of exposure. Any disruption involving advanced foundry capacity, packaging, server assembly, energy supply or geopolitics could matter for Nvidia’s delivery timelines. Power demand is another issue to watch. Reuters reported that TSMC sees energy efficiency becoming a bigger design priority as AI workloads drive higher electricity demand and supply challenges.
There is also a valuation angle, but it needs careful wording. Nvidia’s high gross margins and rapid revenue growth show the strength of current demand. They do not guarantee that margins stay near current levels or that the market will continue to value the company the same way. The more investors expect from Nvidia, the more sensitive the stock can become to any sign of slower growth, tighter supply or rising costs.
What to watch next
The first thing to watch is whether Nvidia gives more detail on what falls inside the US$150 billion figure. A broad annual spending estimate is powerful, but investors will want to know how much relates to wafer supply, packaging, server assembly, local infrastructure and workforce expansion.
The second is capacity. Advanced packaging remains one of the most important bottlenecks in AI hardware, and both Nvidia and AMD are pushing deeper into Taiwan’s ecosystem for that reason.
The third is Nvidia’s next earnings update. The company’s guidance for around US$91.0 billion in second-quarter fiscal 2027 revenue keeps expectations high. If Nvidia continues to meet or beat that level, the Taiwan supply chain story will look like a sign of scale. If growth slows or margins come under pressure, the same spending figure could raise sharper questions about cost, concentration and execution.
