Jensen Huang Calls for Massive TSMC Expansion to Support NVIDIA AI Growth
NVIDIA CEO Jensen Huang hosted a high level banquet in Taiwan which included the major supply chain partners who brought their top executives from TSMC and Foxconn and MediaTek and ASUS. Huang required increased manufacturing capacity because NVIDIA needed to produce more AI chips at a faster rate.
Huang appeared to confirm that TSMC needed to increase its production rate so that it could satisfy the growing demand for both wafers and CoWoS (Chip on Wafer on Substrate) packaging. He expected TSMC needed to expand its production capacity more than 100% during the upcoming ten years. Huang stated that this expansion represents human history's most extensive infrastructure investment because it requires dedicated capacity doubling to fulfill NVIDIA's needs.
NVIDIA has completed its transition to manufacturing operations for all of its upcoming next generation architectures. Huang confirmed the following milestones
- Blackwell The product now enters full production to satisfy current data center requirements.
- Vera Rubin The new platform which contains six advanced chips has entered production phase.
Industry experts predict that cloud service providers will develop custom AI server computing ASICs which will achieve better performance than NVIDIA’s specialized chips. Huang dismissed these claims as "nonsense," because several obstacles exist which would prevent competitors from entering the market.
- R&D Talent High end research and development personnel required to outperform NVIDIA are exceptionally rare.
- Economic Scale NVIDIA’s massive production scale provides a competitive advantage that is difficult for individual companies to match.
- Infrastructure Scope Huang noted that NVIDIA does not simply build chips; it builds the entire end to end AI infrastructure.
Huang showed how NVIDIA holds a special status within the industry because of its exclusive technological advantage. He stated that NVIDIA maintains the exclusive right to work with all AI related stakeholders through its partnerships with Google and the global open source community. His broad integration enables NVIDIA to maintain its market position against companies which create their own standalone hardware products.
