In the recent Nvidia’s Graphics Technology Conference, Nvidia CEO presented the world’s largest GPU. His talk covered a lot of really interesting topics — if you’re a graphical engineer, data scientist, or autonomous vehicle enthusiast. Jensen Huang also introduced brand-new Nvidia Quadro GPU.

Let me walk you through what he spoke about.

Nvidia Quadro GV100

The first product shown off was Quadro GV100, a new high-end Graphic Driver.


1. Lightning-Fast.


1. Not For Gaming.

It was designed for power professional work-stations.


1. Nvidia’s Next-gen Volta architecture.

2. 32-GB Memory.

It will cost about $7000, this price range is quite unexpected.


Likewise uncovered at the GTC keynote was one of Nvidia’s most intense — and most costly — bits of equipment: The Nvidia DGX-2, a GPU supercomputer that tips the scales at around 350 pounds. The “smaller than normal” supercomputer highlights 16 singular Tesla V100 GPUs, each with 32GB of framework memory. These cards all cooperate through Nvidia’s NV Switch innovation, which means this monstrous machine is fit for crunching some genuine numbers. Intended for profound learning applications which require huge measures of preparing power, the DGX-2 is amazing, yet again not for the normal client. Unless you have an extra $400,000 laying around.