Nvidia already sells tons of computer chips to every major company that creates its own artificial intelligence models. But now, with public interest in open source software and do-it-yourself AI, the company has announced that it will begin offering a “personal AI supercomputer” later this year at a price starting at $3,000, with which everyone will be able to apply on their own. home or office.
Nvidia’s fresh computing machine, called Digits, goes on sale in May and is the size of a compact book. It features an Nvidia “superchip” called the GB10 Grace Blackwell, optimized to accelerate the computations needed to train and run AI models, and comes with 128GB of unified memory and up to 4TB of NVMe storage to run particularly huge AI programs.
Jensen Huang, founder and CEO of Nvidia, announced the fresh system along with several other AI offerings during a keynote today at CES, the annual conference for the computing industry held in Las Vegas (you can check out all the top announcements on the WIRED live blog CES).
“Putting an AI supercomputer on the desk of every data scientist, AI researcher and student enables them to engage in and shape the era of artificial intelligence,” Huang said in a statement released before his speech.
Nvidia says the Digits machine, or “GPU deep learning system,” will be able to run one huge language model with up to 200 billion parameters, a gritty measure of the model’s complexity and size. To do this today, you’d have to rent space from a cloud provider like AWS or Microsoft, or build a custom system with a few chips dedicated to AI. Nvidia claims that if two Digits machines are connected via a proprietary high-speed interconnect, they will be able to run the most proficient version of Meta’s open-source Llama model available, which has 405 billion parameters.
The numbers will make it easier for hobbyists and researchers to experiment with models close to the core capabilities of OpenAI’s GPT-4 or Google’s Gemini in their offices or basements. But the best versions of these proprietary models, housed in giant data centers owned by Microsoft and Google, are likely larger and more powerful than anything Digits could handle.
Nvidia is one of the biggest beneficiaries of the artificial intelligence boom. The company’s share price has skyrocketed over the past few years as technology companies clamored to buy huge quantities of the advanced hardware chips it produces, key ingredients in the development of cutting-edge artificial intelligence. The company has proven adept at optimizing hardware and software for artificial intelligence, and its product roadmap has become an significant signal of where the industry should be heading.
Upon launch, Digits will be the most powerful consumer computing hardware in Nvidia’s lineup. It already sells a range of artificial intelligence development chipsets known as Jetson, with prices starting at around $250. Smaller AI models can be run on them and used like mini desktop computers, or installed on a robot to test various AI programs.