As Nvidia is commencing its annual software developer conference, GTS, from 17 to 21 March, the AI giant company’s CEO, Jensen Huang, is expected to share details about the company’s latest AI chips’ family, Rubin, in his keynote address.
Last year, Huang announced at a trade show in Taipei that the company is working on next-generation AI chips that will be rolled out in the market by 2026. While elaborating about Rubin, he said that the family of chips will include new graphics (GPU), central processors (CPU), and networking chips. The new graphic chips that will be used to power AI applications will be comprised of next-generation high-bandwidth memory along the lines of Samsung (005930.KS), and Hynix (000660.KS)
At the trade show, he also expressed that the company has paced up its plan to release a new family of chips every year. However, Nvidia’s latest project, Blackwell, is facing a design flaw, thus lagging behind in manufacturing and impacting the AI industry at a broader level.
Nvidia has also been challenged by DeepSeek by announcing its potential to develop a chatbot with less computing power, thereby requiring fewer Nvidia chips. The giant AI chip manufacturing company adamantly responded that the newer AI models will be more dependent on Nvidia chips because they are currently the fastest ‘token’ generating chips.
Amid this testing time of Nvidia, the company is expected to share more details about its upcoming family of chips, Rubin, to satisfy its customers and reaffirm its position in the AI industry. Nvidia is the leading global AI chip manufacturing company, controlling 80 percent of the global AI chip market. The company manufactures chip systems for automobiles, robotics, and computers. Its AI-based Compute and Networking business is the largest revenue generation section.
