Nvidia RTX Pro Blackwell series pack insane power for AI and 3D work

Nvidia announced the RTX Pro Blackwell series of GPUs, including the RTX Pro 6000 Blackwell for workstations, designed for professional workloads across various industries. The launch occurred during the GTC event.

The RTX Pro 6000 Blackwell for workstations features 96GB of GDDR7 memory and requires 600 watts of power, slightly more than the 575 watts of the RTX 5090. It includes double flow-through cooling, 24,064 CUDA cores, a 512-bit memory bus, and 1792 GB/s of memory bandwidth. The GPU supports PCIe Gen 5, DisplayPort 2.1, and includes the latest Blackwell generation of RT and Tensor cores.

This GPU is specifically tailored for professionals working in areas such as game development, AI workloads, and tasks that necessitate substantial VRAM and high-processing speed. The RTX Pro 6000 will also be available in a Max-Q variant and a server edition for data centers.

Nvidia is adopting the new RTX Pro branding, which replaces both the RTX numbering scheme and previous Quadro branding. The lineup also includes the RTX Pro 5000 and RTX Pro 4000 Blackwell for desktops and laptops, along with the RTX Pro 4500 Blackwell for desktops.

Laptop versions of the RTX Pro Blackwell will consist of the 500, 1000, 2000, 3000, and 4000 models, offering up to 24GB of VRAM. These variants support Nvidia’s latest Blackwell Max-Q technologies, aimed at optimizing laptop performance and power efficiency using AI. They will compete against AMD’s Strix Halo chips, which have 128GB of unified memory shared among the CPU, GPU, and AI engines.

Nvidia has yet to disclose pricing for the RTX Pro 6000 workstation variant. It will be available through distribution partners such as PNY and TD Synnex in April, with manufacturers like Dell, HP, and Lenovo offering it starting in May. The server variant will be accessible from Cisco, Dell, HP, Lenovo, and Supermicro in the near future, with cloud providers including AWS, Google Cloud, Microsoft Azure, and CoreWeave expected to provide RTX Pro 6000 Blackwell servers later this year.

RTX Pro Blackwell GPUs unlock advanced capabilities in generative, agentic, and physical AI, delivering high performance and efficiency.

Key features include:

• NVIDIA Streaming Multiprocessor provides up to 1.5 times faster throughput with new neural shaders integrating AI into programmable shaders.
• Fourth-Generation RT Cores offer up to double the performance for photorealistic scene creation.
• Fifth-Generation Tensor Cores perform up to 4,000 AI trillion operations per second while adding support for FP4 precision and NVIDIA DLSS 4 Multi Frame Generation.
• GDDR7 Memory enhancements enable up to 96GB for workstations and servers, and 24GB for laptops, facilitating faster application performance with larger datasets.
• Ninth-Generation NVIDIA NVENC and Sixth-Generation NVIDIA NVDEC enhance video encoding and decoding capabilities, improving quality and speed for professional applications.
• Fifth-Generation PCIe support allows for double the bandwidth for efficient data transfer.
• DisplayPort 2.1 drives high-resolution displays, providing support for diverse visual tasks like video editing and 3D design.
• Multi-Instance GPU (MIG) technology facilitates secure partitioning of GPUs, enhancing performance and flexibility for various workloads.

The RTX Pro 6000 data center and server editions are optimized for enterprise deployments and can support up to eight GPUs per server. This configuration allows for high-performance solutions in AI, scientific computing, and visual computing across industries such as healthcare and manufacturing.

Nvidia vGPU™ software will be available for the RTX Pro 6000 Blackwell Server Edition in the latter half of the year, allowing for the powering of AI workloads across virtualized environments.

Availability of the RTX Pro 6000 Blackwell is expected from leading data center partners and cloud service providers later this year. The RTX Pro 6000 Blackwell Workstation and Max-Q editions will be shipped from April and May, respectively, with other workstation editions and laptop variants rolling out throughout the summer.

Featured image credit: Nvidia