Nvidia CEO Jensen Huang announced the company’s innovations at the Las Vegas technology trade show. Nvidia wants to revolutionize the industry with AI technology.
Tech company Nvidia and its founder Jensen Huang want to bring to market a small AI supercomputer powered by a new computer chip.
Nvidia shrinks AI supercomputer to desktop format – Gallery
Nvidia CEO Jensen Huang announced the company’s innovations at the Las Vegas technology trade show. Nvidia wants to revolutionize the industry with AI technology.
Tech company Nvidia and its founder Jensen Huang want to bring to market a small AI supercomputer powered by a new computer chip.
Semiconductor company Nvidia wants to bring an AI supercomputer to the desktop. The company’s boss, Jensen Huang, unveiled a small device called Project Digits on Monday at the start of the CES technology trade show in Las Vegas.
According to Nvidia, millions of developers, data scientists, and students will be able to use it to work on software with artificial intelligence. The device will start at $3,000 and will be available from multiple manufacturers starting in May.
Inside the computer is a previously secret Nvidia chip called GB10. Using Digits, it should be possible to work on complex AI models, for example.
Nvidia’s chips have become a key technology for the growing AI industry in recent years. The company’s systems are used worldwide for training applications using artificial intelligence. Powerful companies like Google and Facebook group Meta have entire data centers filled with them, and AI startups like OpenAI, the company that invented ChatGPT, also rely on them. Desktop-based technology could further strengthen the group’s dominance in the industry.
Focus on humanoid robots and self-driving cars
Nvidia is also increasingly focused on robotics and self-driving car technology. Self-driving cars are a reality and are likely to become a multitrillion-dollar business, Huang said. Among other things, Nvidia wants to offer a computer called Thor that processes information from various vehicle sensors. Huang emphasized that this chip system can also be used, for example, in humanoid robots.
Nvidia has also developed Cosmos software, which generates realistic environments for training artificial intelligence. The goal is to “teach AI to understand the real world,” Huang said. The AI model behind Cosmos was trained on 20 million hours of video recordings.
