On Monday, Nvidia kicked off its annual GPU Technology Conference (GTC) with a keynote address from CEO Jensen Huang. Investors have been keenly interested in updates, especially regarding the company’s latest GPU offering, codenamed “Blackwell” or the B100 AI chip. Wedbush analyst Dan Ives confirms that Wall Street will be eagerly eyeing the details and pricing of the Blackwell B100 AI chip and its architecture. This new offering marks a pioneering milestone in the AI industry, and all eyes are on what is revealed in the conference.

The B100, the company’s next-generation GPU, will likely be manufactured by Taiwan Semiconductor (TSM). At November’s SC2023 Super Computing conference, Nvidia suggested that its new AI chip would deliver more than twice the AI performance of its Hopper GPUs. Analysts agree that this is a big deal. Joseph Moore at Morgan Stanley sees the B100 as leading Nvidia into a new product cycle. And Wedbush analysts predict substantial performance gains. It is rumored that the B100 will be based on TSMC’s CoWoS-L packaging technology (Chip on Wafer on Substrate), advanced tech that helps develop large processors.

Who Are the Other Players Involved? 

With a 90% market share of advanced chip development, Taiwan Semiconductor is the world’s leading contract AI chip manufacturer. With the exploding interest in AI, chip makers are seeing an enormous interest in chip technology. This has led to greater demand for its products. The company is considering expanding its packaging capacity for advanced chips into Japan. This would be the first time Taiwan Semiconductor’s CoWoS technology was brought outside the country. This decision comes at a critical time, and it’s a strategic move, given the geographical risk inherent in an industry that is exclusively Taiwan-based.

Subscribe to our SMS alerts for up-to-the-minute stock alerts