Jensen Huang: To unveil a "world's first" new chip in March

On February 19th, according to foreign media Wccftech, NVIDIA CEO Jensen Huang previewed the upcoming GTC 2026 conference, clearly stating that a “world-first” new chip will be unveiled at the event, sparking widespread industry attention. As a leader in AI chips, NVIDIA’s major announcement is believed to further solidify its leading position in AI infrastructure.

It is reported that the keynote speech at GTC 2026 will be held on March 15th in San Jose, California, focusing on a new era of AI infrastructure competition. Huang Huang admitted that developing these new chips is highly challenging, “all technologies are approaching their limits,” but given NVIDIA’s past track record, the industry is full of expectations for this new product.

Currently, the specific model of the new product has not been disclosed, but it is widely speculated that it will come from two major chip series: one is a derivative of the Rubin series (such as the previously revealed Rubin CPX), which was showcased at CES 2026 and includes six new-designed chips now in full production; the other is the next-generation Feynman series chips, described as “revolutionary,” with NVIDIA exploring extensive integration centered on SRAM or using 3D stacking technology to combine LPUs, though details are still unconfirmed.

It is worth noting that NVIDIA is adjusting to quarterly changes in AI computing demands, shifting from Hopper and Blackwell series focused on model pre-training to Grace Blackwell Ultra and Vera Rubin series targeting inference scenarios. The new products are expected to specifically address latency and memory bandwidth bottlenecks. Huang Huang stated that broad collaboration and investment are key to NVIDIA maintaining its leadership. The company is strategically developing the entire AI industry chain, covering energy, semiconductors, data centers, and other fields to promote comprehensive AI industry growth.

(Source: Kechuangban Daily)