Nvidia launches an AI supercomputer for end users
Nvidia just announced a Modern $ 3,000 AI supercomputer during the CES 2025 tonic . Currently , it ’s calledProject Digits . This will be available for purchase starting in May 2025 , targeting developer , investigator , and even students . So what precisely is this AI supercomputer , and how does it differ from a normal one ? Here ’s everything you require to eff about Nvidia ’s Project Digits .
What Is Project Digits?
Project Digits is a Mac - mini - like gimmick designed to consort AI models . It is power by Nvidia’sGB10 Grace Blackwell Superchip . That ’s a bit of a mouthful , but it combines a powerful Blackwell GPU with fifth - generation Tensor Cores and a gamy - performance Grace CPU featuring 20 Arm - ground burden . Interestingly , you could run it using your paries socket power issue which was antecedently not potential .
In simpler terms , it comes with a super crisp specifically plan for AI calculations and tasks that give birth apetaflop of AI performanceat FP4 precision ( make the calculations faster by clear approximation ) .
Also Read :
Users can leverage this twist for AI task like prototyping , fine - tuning AI modeling , and even running an intact prominent AI model directly on the gimmick . It supports AI manakin with up to200 billion parameter , and if you involve more , you could link two units together using Nvidia ConnectX to handle 405 - billion - argument simulation . For context , with this setup , you’re able to play Meta ’s largestLlama 3.1 405Bmodel locally .
In addition , it includes128 GB of unified memoryand up to4 TB of NVMe memory , enabling it to manage monumental datasets and complex models required for AI .
Coming to software , the twist run on theNVIDIA DGX OS , which is built on top of open - author OS Ubuntu Linux . developer can use democratic AI model like PyTorch , Python , and Jupyter notebooks . Nvidia ’s AI software program library , let in tools like NeMo for model mulct - tuning and RAPIDS for data science workflow , is also supported . Once good example are built locally , they can be deploy to Nvidia DGX Cloud or datum centers without compatibility issues .
Pricing and Availability
Project Digits will begin at$3,000and is set to set in motion inMay 2025through Nvidia and its partners . For AI researchers , data scientists , developers and scholar , who are look to try out with AI or scale up their project , this could be a game - transfer tool to bridge over the gap between local development and large - scale deployment .
