NVIDIA’s close ties with Taiwan’s chip powerhouse were on full display as CEO Jensen Huang appeared at TSMC’s Sports Day in Taiwan, telling the crowd that without TSMC, NVIDIA would not be where it is today. The remark highlights just how central the foundry’s manufacturing and advanced packaging capabilities have become to NVIDIA’s rise in AI and data center computing.
This marks Huang’s fourth trip to Taiwan this year, and each visit has reinforced the same message: TSMC is a critical pillar in NVIDIA’s growth. The latest stop also carries a clear business objective. With demand for NVIDIA’s Blackwell-based chips and rack-scale AI systems surging, the company is seeking additional wafer allocations to keep pace with customer orders. NVIDIA is also expected to be a major customer for TSMC’s 3nm process, reportedly targeting around 30% of total output, a signal that the AI wave is only accelerating.
Strip away the sound bite and the logic is straightforward. TSMC sits at the center of NVIDIA’s supply chain, from cutting-edge logic to advanced packaging services like CoWoS (Chip-on-Wafer-on-Substrate) that are vital for high-performance AI accelerators. While NVIDIA once tended to wait before jumping on the latest nodes, the company is now racing to secure the most advanced technologies, including next-generation offerings such as TSMC’s A16, cementing the foundry’s importance in its roadmap.
The stakes are high. Top-tier AI capacity at advanced nodes is the most sought-after resource in the semiconductor industry, and competitors across cloud, hyperscale, and infrastructure markets are vying for the same production slots. By keeping its relationship with TSMC tight and proactive, NVIDIA is positioning itself to sustain leadership in GPUs, AI accelerators, and full-stack data center systems built on the Blackwell architecture.
Expect the partnership to deepen as NVIDIA scales output and pushes into more advanced nodes and packaging. If the past year is any indication, Huang’s frequent visits to Taiwan are less about ceremony and more about ensuring the silicon pipeline stays full—because for NVIDIA’s AI ambitions, every wafer counts.
