DGX Spark
Product
Running 20B to 120B+ parameter models locally on a desktop; chaining multiple units to run 400B models
"We were one of the first people, along with Elon and a few others, to receive the DGX Spark as well, which sits on your desk and provides 128 GB of unified memory to run that kind of 20 to 120B model range. You can absolutely stack them on your desk almost ..."
More source quotes
Running local AI models for unlimited free inference
"The DGX Sparks a really great plug-and-play device. You just plug it into the wall and it works. It has 128 gigabytes of unified memory which basically means you can run any model that goes up to 128 gigs."