Nvidia’s $3,000 “Personal AI Supercomputer” will let you ditch the data center

Nvidia already sells large quantities of computer chips to every major builder it owns artificial intelligence Models. But now, at a moment when it has become public interest Open source With AI on the rise, the company announced that it will also begin offering an “AI-powered personal supercomputer” later this year, starting at $3,000, that anyone can use in their home or office.

Nvidia’s new desktop device, called Digits, will go on sale in May and is about the size of a small book. It has a “super chip” from Nvidia called the GB10 Grace Blackwell, which is optimized to speed up the calculations needed to train and run AI models, and comes equipped with 128GB of unified memory and up to 4TB of NVMe storage to handle particularly large AI programs.

Jensen Huang, founder and CEO of Nvidia, announced the new system, along with several other AI offerings, during a keynote speech today at CESan annual computer industry conference held in Las Vegas. (You can see all the major announcements at WIRED CES Live Blog.)

“Putting an AI supercomputer on the desks of every data scientist, AI researcher and student empowers them to participate and shape the AI ​​era,” Huang said in a statement issued before his keynote.

Nvidia says the Digits device, which stands for “GPU Intelligence Training System for Deep Learning,” will be able to run a single large language model containing up to 200 billion parameters, a rough measure of the model’s complexity and size. To do that today, you’d need to rent space from a cloud provider like AWS or Microsoft, or build a custom system with a set of chips designed to run AI. If two Digits devices are connected using a proprietary high-speed interconnect link, Nvidia says they will be able to run… The most capable The version is available from Meta’s open source Llama model, which contains 405 billion parameters.

The numbers will make it easier for hobbyists and researchers to experiment with models that come close to basic capabilities OpenAI GPT-4 or Google Gemini In their offices or basements. But the best versions of those proprietary models, housed inside giant data centers owned by Microsoft and Google, are likely larger and more powerful than anything Digits can handle.

Nvidia has been one of the biggest beneficiaries of Artificial intelligence boom. Its stock prices have risen dramatically over the past few years as technology companies demand huge amounts of the advanced hardware chips it produces, a crucial component for the development of advanced artificial intelligence. The company has proven its prowess in optimizing hardware and software for AI, and its product roadmap has become an important signal of where the industry is expected to go next.

When it is released, the Digits will be the most powerful consumer computing hardware Nvidia has to offer. It already sells a set of chips specifically for AI development known as Jetson that start at around $250. These devices can run smaller AI models and can be used as a small desktop computer or installed on a robot to test different AI programs.