Just hours after Dell decided all on its own to rename its laptop production line withPro and Pro Max cognomen , Nvidia also took a Thomas Nelson Page from Apple ’s playbook with its new “ personal AI supercomputer . ” No , it ’s not called the Nvidia miniskirt , but it does have a certain simpatico with the Mac mini .
Nvidia did n’t share the precise specs for what it ’s callingProject Digits , but the twist conform to neatly in CEO Jensen Huang ’s palm during its unveiling . Next to a admonisher and keyboard ( above ) , it seem to be slimly bigger than the Mac mini with a boxier and more industrial purpose .
Nvidia did n’t provide benchmarks for the GB10 scrap , but did say it ’s able of a petaflop of FP4 floating point public presentation and able-bodied to “ run up to 200 - billion - parameter large lyric models , ” likely beyond the capableness of the $ 599 Mac mini . However , Nvidia is n’t selling it to just anyone . The destination of Project Digits , Huang said , is to put “ an AI supercomputer on the desk of every data scientist , AI investigator and student . ”
Nvidia
To that end , the computer initiate at $ 3,000 and add up with a 20 - essence GB10 Grace Blackwell Superchip and 128 GB of interconnected memory . Configuring aMac miniwith the M4 Pro chip , which has a 14 - core CPU , and 64 GB of random access memory , cost $ 2,499 with 1 TB of storage . A 1 TB Mac Studio fitted with a 24 - core M2 Ultra and 128 GB of RAM contribute the toll to $ 4,799 .
While it takesseveral M4 Mac minis to even reach a teraflopof carrying into action , the M2 Ultra in the Mac Studio already tops 20 trillion floating point operations per second . And Apple is anticipate to promote that machine to the M4 family of C.P.U. sometime this twelvemonth , which will manifestly bring a major performance boost . And we take the power standoff will be many times that of the Mac mini or Mac Studio ( or Mac Pro for that issue ) .
Of course , we ’re compare Apples to oranges here . Nvidia ’s Mac mini dead ringer is n’t meant for everyday study . finger is kitted out with Nvidia ’s full AI software enterprise platform , let in “ an all-inclusive depository library of Nvidia AI software for experimentation and prototyping , including package development kit , orchestration tools , frameworks and model . ”
