Nvidia’s $3,000 ‘Personal AI Supercomputer’ Will Let You Ditch the Data Center

Nvidia already sells batches of computer chips to all large companies creating their own artificial intelligence models. But now, at a time when public interest in open source At the same time that DIY AI is soaring in popularity, the company announced that later this year it will also begin offering a “personal AI supercomputer” starting at $3,000 that anyone can use in their home or business. office.

Nvidia’s new desktop computer, called Digits, will go on sale in May and is the size of a small book. It contains an Nvidia “superchip” called the GB10 Grace Blackwell, optimized to speed up the computation needed to train and run AI models, and features 128 gigabytes of unified memory and up to 4 terabytes of NVMe storage to handle particularly large AI programs.

Jensen Huang, founder and CEO of Nvidia, announced the new system, as well as several other artificial intelligence offerings, during his keynote address today at the conference KESannual computer industry conference held in Las Vegas. (You can see all the most important announcements on WIRED CES Live Blog.)

“Putting an AI supercomputer on the desks of every scientist, AI researcher and student gives them the opportunity to participate in and shape the age of AI,” Huang said in a statement released ahead of his keynote.

Nvidia says the Digits machine, which stands for “deep learning GPU intelligence training system,” will be able to run one large language model with up to 200 billion parameters, a rough measure of the model’s complexity and size. To do this today, you’ll need to rent space from a cloud provider like AWS or Microsoft, or build your own system with multiple chips dedicated to running the AI. If two Digits machines connected using a proprietary high-speed connection, Nvidia says they could run the most capable an open source Meta version of the Llama model is available, which has 405 billion parameters.

The numbers make it easier for hobbyists and researchers to experiment with models that get closer to basic capabilities GPT-4 from OpenAI or Gemini by Google 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 can handle.

Nvidia was one of the biggest beneficiaries Artificial Intelligence Boom. Its stock price has skyrocketed over the past few years as tech companies have demanded huge volumes of the cutting-edge hardware chips it produces, a critical ingredient for developing advanced artificial intelligence. The company has proven its ability to create hardware and software optimized for artificial intelligence, and its product roadmap has become an important signal of where the industry is expected to go next.

When released, Digits will be the most powerful consumer computing hardware Nvidia offers. The company already sells a range of artificial intelligence development chipsets known as Jetson, which start at about $250. They can run smaller AI models and be used as a mini-desktop or mounted on a robot to test various AI programs.