Five years ago, Nvidia was spending somewhere between $10 billion and $15 billion annually in Taiwan. Now that number is roughly $150 billion. Jensen Huang dropped that figure at Computex 2026, and if you needed a single data point to understand why Taiwan matters more than ever to the global economy, that’s the one.
The Nvidia CEO called Taiwan the “epicenter of the AI revolution,” which sounds like keynote hyperbole until you consider the math. TSMC, Nvidia’s primary foundry partner, produces approximately 90% of the world’s most advanced semiconductor nodes used for AI accelerators. That’s not a market share. That’s a near-monopoly on the silicon that powers every major AI model, data center expansion, and GPU-hungry workload on the planet.
The numbers behind the Nvidia-TSMC alliance
Nvidia’s $150 billion annual investment in Taiwan flows overwhelmingly through TSMC, which manufactures the company’s cutting-edge AI GPUs using advanced packaging technologies like Chip-on-Wafer-on-Substrate, known as CoWoS. That packaging tech is critical for stacking chiplets together in ways that boost performance for AI workloads.
AMD announced its own commitment of over $10 billion to support Taiwan’s AI sector around the same timeframe. When the two biggest GPU companies on Earth are both racing to deepen their manufacturing ties to a single island, it tells you something about where the leverage sits in the global chip supply chain.
