Pune, India – November 10, 2025- Nvidia’s Blackwell chips are seeing unprecedented demand as artificial intelligence adoption accelerates worldwide, CEO Jensen Huang said at a Taiwan Semiconductor Manufacturing Company (TSMC) event on Saturday. The surge highlights the growing importance of advanced semiconductors for AI infrastructure, including data centers, research labs, and autonomous systems.
Huang stressed that Blackwell is more than a standard graphics processor. “We develop GPUs, but our work also includes CPUs, networking systems, and switches. Blackwell combines multiple chip types into one platform,” he said, underlining the complexity and versatility of Nvidia’s latest technology.
This event marked Huang’s fourth visit to Taiwan this year, reflecting the company’s deep reliance on TSMC for wafer production. Although he would not provide specific numbers, Huang said Nvidia is asking for significant wafer allocations to keep up with the very strong demand. “TSMC is doing a great job supporting us,” he mentioned, underlining the fact that all Nvidia successes depend on close partnership relations with the world’s largest contract maker of chips.
TSMC CEO C.C. Wei confirmed that Nvidia's wafer needs are increasing, but said he couldn't go into details because of confidentiality reasons. The deal comes as Nvidia celebrates a milestone: it became the first company in October to ever reach a $5 trillion market valuation. Wei even referred to Huang as the “five-trillion-dollar man,” underscoring the company’s meteoric rise in the technology sector.
Despite strong growth, Huang warned that supply chain challenges remain, especially regarding memory. “Business is expanding rapidly, and shortages of various components are possible,” he said. However, he expressed confidence in Nvidia’s memory partners SK Hynix, Samsung, and Micron for scaling production to meet Blackwell’s needs. Moreover, “These are top-tier memory producers who have significantly increased capacity to support us,” he added.
Nvidia has already received advanced memory samples from all three suppliers. These components are crucial in unlocking the full performance of Blackwell chips. Asked about a possible increase in memory module prices, Huang declined to speculate: “It is up to them to decide how to manage their businesses.”
The memory market itself is highly active. SK Hynix said its chip supply for the next year is fully booked and will invest heavily, expecting a long “super cycle” driven by AI. Samsung confirmed that negotiations are underway to supply next-generation high-bandwidth memory chips, known as HBM4, to Nvidia.
On the geopolitical front, Huang made it clear that Nvidia would not sell Blackwell chips to China. Moreover, restrictions from the U.S. government are still in place, effectively prohibiting advanced chip sales to China that could enable military or AI development. Additionally, “There are no active discussions,” he said, further affirming Nvidia’s compliance with U.S. export rules.
Organizations are rapidly adopting Nvidia's Blackwell chips, driving the global race to build next-generation AI infrastructure. Analysts note Nvidia must secure wafer and memory supplies to stay ahead as competitors boost AI hardware investments.
Huang concluded with optimism about the future. He said Nvidia is entering extraordinary growth and will scale production to meet rising AI-driven demand. Blackwell chips will be at the core of autonomous systems, AI research, and enterprise data centers worldwide.
With rising demand, strategic partnerships, and technological innovation, Nvidia is shaping the future of AI hardware. Industries adopting AI will rely on Nvidia’s Blackwell platform to ensure speed, reliability, and efficiency.
