Nvidia CEO Jensen Huang recently announced that the company’s next-generation chips are in “full production” and will deliver five times the AI computing power of previous models. During his keynote at CES in Las Vegas, he highlighted the Vera Rubin platform, comprising six chips designed to significantly enhance AI applications like chatbots. These chips can generate “tokens,” the basic units of AI, with a tenfold efficiency increase, thanks to a new proprietary data method. While Nvidia currently leads the AI model training market, it faces rising competition from companies such as AMD and customers like Google, which is developing its own AI chips. Huang shared insights on improved storage technology, called “context memory storage,” which enhances chatbot responses. He also introduced new networking switches featuring co-packaged optics, crucial for connecting numerous machines. Nvidia’s acquisition of startup Groq aims to bolster its product lineup without affecting its core business.
Source link
Share
Read more