NVIDIA’s CEO is making waves in Taiwan, as reports suggest that his meetings with TSMC executives focus on securing a substantial share of 3nm production for their AI products. This strategic move highlights NVIDIA’s commitment to staying ahead in the AI technology race.
NVIDIA Eyes TSMC’s 3nm Production for Rubin AI Lineup
The AI industry is buzzing with excitement as NVIDIA gears up for the Blackwell Ultra ramp-up, resulting in a significant surge in chip demand. Sources indicate that Jensen’s visit to Taiwan includes a tour of TSMC’s Tainan fab, which is pivotal for 3nm mass production. The talks are reportedly centered on increasing capacity allocation for TSMC’s N3 process, crucial for the next-gen Vera Rubin lineup. This expansion plan aims to increase TSMC’s 3nm capacity from 100,000 to 160,000 wafers per month, with NVIDIA expected to receive a large portion. With anticipation of high demand for Rubin AI chips, NVIDIA is securing supply early, bolstering TSMC’s revenue for the upcoming quarters.
Anticipation Builds Around Rubin AI Release
The Rubin AI lineup is poised to be a groundbreaking release for NVIDIA, featuring comprehensive architecture improvements and a significant leap in computing power. Incorporating TSMC’s N3P process along with advanced HBM4 technology, the Rubin lineup is set to redefine standards. Reports highlight that NVIDIA has already secured key customers for Rubin, underscoring the high interest within the AI sector, even with two quarters to go before mass production.
For TSMC, the influx of demand from high-performance computing customers, particularly NVIDIA, plays a crucial role in its revenue stream. The company is swiftly adopting new processes and is planning to transition to A16 (1.6nm) technology soon. During his visits to Taiwan, Jensen Huang consistently underscores TSMC’s importance to NVIDIA’s strategic goals.