Tech

NVIDIA’s $16,000 RTX PRO 6000 is built for professional workloads

The workstation GPU’s 96GB of error-correcting memory and professional driver support carry a steep premium over the GeForce RTX 5090.

Editorial persona
Owen Mercer
Markets and Finance Editor
Published
Draft
Source: Engadget · View original source
Black NVIDIA RTX PRO 6000 graphics card floating above a dark, ridged surface.
Markets

NVIDIA’s RTX PRO 6000 Blackwell Workstation Edition is aimed at AI developers, engineers and 3D professionals, rather than typical gaming PC buyers. The company lists it in the US for $16,000, up from $8,565 at launch, according to Engadget.

The card has 96GB of error-correcting GDDR7 memory, 24,064 CUDA cores and a maximum board power of 600W. Its large memory pool is intended for demanding work such as AI projects, large datasets, 3D visualisation and virtual reality environments.

NVIDIA also supports RTX PRO cards with workstation-focused drivers, enterprise support and certifications for professional applications. The company previously sold professional graphics cards under the Quadro brand; AMD’s Radeon PRO line also targets workstation users.

The GeForce RTX 5090 launched at $1,999 and has 32GB of GDDR7 memory. In tests by YouTuber der8auer, the RTX PRO 6000 was about 3% to 14% faster than the RTX 5090 in several games. Those results come from one tester and do not establish performance across games or workloads.

NVIDIA has not disclosed why the RTX PRO 6000’s listed price rose. Engadget noted the increase came amid higher memory costs and broader GPU price rises, without establishing either as the cause. NVIDIA directs buyers to its Marketplace and workstation partners, though individual enthusiasts can also buy the card.

Continue reading

More from Tech

Read next: PipePipe maintains independent path after NewPipe fork
Read next: Garmin Fenix 9 adds speed and storage, but keeps Fenix 8 battery life
Read next: macOS 27 ends support for Intel Macs, Engadget reports