Nvidia, the global leader in artificial intelligence chips, has requested that TSMC, the world’s largest chip manufacturer, dramatically increase the production of its advanced Blackwell processors. According to reports, the request involves a 50% surge in dedicated 3-nanometer wafer output, pushing capacity from an estimated 100k–110k wafers toward 160k wafers per month. This move underscores the relentless demand for AI technology and signals Nvidia’s willingness to invest massive capital to meet the needs of its largest clients.

Strong Demand: The Engine Driving the Increase

Nvidia CEO Jensen Huang addressed the issue during a TSMC event, emphasizing that demand for the Blackwell platform is “very strong” and strengthening “month by month.” This demand is not confined solely to the Graphics Processing Units (GPUs); rather, it extends across the entire system architecture, including networking equipment, switches, and associated components related to the Blackwell platform. To support this phenomenal growth, Nvidia noted that all three of its HBM (High Bandwidth Memory) suppliers have already scaled up their capacity, indicating a broader supply chain mobilization.

Implications for TSMC and the Chip Sector

Nvidia’s request necessitates that TSMC leverage its Southern Taiwan Science Park facility, dedicating approximately 35,000 extra wafers monthly to Nvidia’s products. For TSMC, this reinforces its position as the dominant player in 3-nanometer fabrication technology and serves as a highly positive signal regarding its future revenues. For the wider semiconductor sector, the surge in orders is a vital barometer for the continuation of the AI “boom.” It indicates that demand for the most advanced products is nowhere near saturation, which in turn supports the high valuations of companies focused on this domain.

Supply Chain Challenges and Forward Look

Such a dramatic increase in throughput places heavy pressure on TSMC’s entire supply chain and its partners. The need to ensure a steady supply of raw materials, wafers, and memory components becomes a significant logistical and operational challenge. Will TSMC and other suppliers be able to meet these demanding targets without major delays? Other competing manufacturers may face longer wait times for 3-nanometer production, potentially impacting their own launch schedules. Ultimately, Nvidia’s hefty orders solidify the perception that artificial intelligence is not merely a trend but an intensifying and continuous wave of technological investment that will dictate the pace of the chip market for years to come.