Nvidia has unveiled the PRO lineup in its latest Blackwell series, showcasing the flagship prosumer card, RTX PRO 6000 “Blackwell,” based on the GB202 GPU architecture and featuring 96GB of GDDR7 memory.
In comparison, the RTX 5090 offers 32GB of VRAM and operates on 575 watts, while the RTX PRO 6000 requires 600 watts of power. The GB202 chip includes 24,064 CUDA cores, enhancing AI workloads and ray tracing with 125 TFLOPS FP32 and 4000 TOPS of AI performance at FP4 precision. Its 96GB GDDR7 ECC memory significantly improves upon the RTX 5090’s 32GB, providing an impressive 1.8 TB/s memory bandwidth for demanding professional applications.
The RTX PRO lineup also features a Max-Q variant and a server edition for data centers. Targeting professionals in game development, AI, and VRAM-intensive tasks, the RTX PRO 6000 debuts Nvidia’s new RTX PRO branding, replacing the previous RTX/Quadro naming. In addition, Nvidia introduced the RTX PRO 5000, 4500, and 4000 Blackwell GPUs for desktops and laptops.
Laptop RTX PRO models (3000-5000) offer up to 24GB VRAM and utilize Blackwell Max-Q for AI-driven power efficiency, competing with AMD’s Strix Halo unified memory architecture. Pricing for the RTX PRO workstation series has not been announced yet.
RTX PRO 6000 Blackwell Launching This Summer
Regarding availability, Nvidia confirmed that the RTX PRO 6000 Blackwell Server Edition will soon be accessible in server configurations from major data center partners, including Cisco, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro. Cloud providers and data centers will offer RTX PRO 6000 Server Edition instances later this year.
Workstation models will launch in April through distributors and in May via manufacturers. Other RTX PRO Blackwell GPUs will be available this summer for desktops and later this year for laptops through various partners.
Are you excited about the Nvidia RTX PRO 6000 card? Do you think it will overcome the RTX 5090 bricking issues? Share your thoughts in the comments below.
