NVIDIA RTX PRO Blackwell Series GPUs
Designers, developers, data scientists, and creatives can now build and collaborate with agentic AI on workstations and servers to NVIDIA Blackwell RTX PRO.
For millions of professionals across the world, groundbreaking AI and graphics performance redefines visualisation, simulation, and scientific computing from desktop to data centre.
With its ground-breaking accelerated computing, AI inference, ray tracing, and neural rendering technologies, NVIDIA today unveiled the NVIDIA RTX PRO Blackwell series, a revolutionary generation of workstation and server GPUs that are revolutionising workflows for professionals in AI, technology, creativity, engineering, and design.
From developing physical AI that powers autonomous robots, vehicles, and smart spaces to agentic AI, simulation, extended reality, 3D design, and complex visual effects, the NVIDIA RTX PRO Blackwell series gives professionals in a variety of industries the newest and greatest compute power, memory capacity, and data throughput at their fingertips, whether they are using data centre GPUs, mobile workstations, or desktop computers.
Among the new lineup are:
- NVIDIA RTX PRO 6000 Blackwell Server Edition data centre GPU
- Desktop GPUs include RTX PRO 6000 Blackwell Workstation Edition, Max-Q Workstation Edition, 5000, 4500, and 4000.
- Laptop GPUs from Blackwell are RTX PRO 5000, 4000, 3000, 2000, 1000, and 500.
NVIDIA Blackwell Technology Arrives in Data Centres and Workstations
The remarkable performance, efficiency, and scale of NVIDIA RTX PRO Blackwell GPUs enable the possibilities of generative, agentic, and physical AI.
Features of NVIDIA’s RTX PRO Blackwell GPUs include:
- The NVIDIA Streaming Multiprocessor will power the next ten years of advancements in AI-augmented graphics with its up to 1.5x faster throughput and new neural shaders that include AI within programmable shaders.
- Fourth-Generation RT Cores: With NVIDIA RTX Mega Geometry optimisations, these cores can produce complex 3D designs and photoreal, physically correct scenes with up to twice the performance of the previous generation.
- Fifth-Generation Tensor Cores: These cores can execute up to 4,000 AI trillion operations per second and enable NVIDIA DLSS 4 Multi Frame Generation and FP4 precision, ushering in a new age of AI-powered graphics and facilitating the quicker execution and prototyping of larger AI models.
- Faster, Larger GDDR7 Memory: Increases capacity and bandwidth up to 24GB on laptops and up to 96GB on workstations and servers. For everything from taking on enormous 3D and AI projects to investigating expansive virtual reality worlds, this makes it possible for applications to operate more quickly and work with bigger, more complicated datasets.
- Ninth-Generation NVIDIA NVENC: With additional capability for 4:2:2 encoding, this technology speeds up and enhances the quality of video encoding for professional video applications.
- Sixth-Generation NVIDIA NVDEC: Supports 4:2:2 H.264 and HEVC decoding and can double the throughput of H.264 decoding. Professionals may employ sophisticated AI-powered video editing tools, speed up the ingestion of video data, and enjoy high-quality video playback.
- Data-intensive applications benefit from fifth-generation PCI Express’s doubled bandwidth for CPU memory data transfer speeds.
- The DisplayPort 2.1 supports 4K/480 and 8K/165 Hz displays. More bandwidth facilitates smooth multi-monitor setups, while improved dynamic range and colour depth improve video editing, 3D modelling, and live broadcasting colour accuracy.
- Multi-Instance GPU (MIG): MIG technology allows a single GPU to be securely divided into up to four instances (6000 series) or two instances (5000 series) in the RTX PRO 6000 data centre and desktop GPUs and 5000 series desktop GPUs. The purpose of fault isolation is to maximise performance and flexibility by preventing workload interference and ensuring safe, effective resource allocation for a variety of workloads.
The newest NVIDIA Blackwell Max-Q technologies, which use AI to continuously and intelligently optimise laptop performance and battery economy, are also supported by the new laptop GPUs.
NVIDIA RTX PRO Blackwell GPUs provide previously unheard-of speed and efficiency in producing immersive experiences, digital twins of real-world landscapes, and breathtaking images using neural rendering and AI-augmented tools. By providing designers and engineers with outstanding performance for intricate modelling, rendering, and visualisation, GPUs are designed to improve processes for 3D computer-aided design and building information models.
The passively cooled RTX PRO 6000 Blackwell Server Edition offers up to eight GPUs per server and is designed for enterprise data centres. The RTX PRO 6000 Blackwell Server Edition performs well for next-generation AI, scientific, and visual computing applications in healthcare, manufacturing, retail, and media and entertainment that require data centre compute density and scale.
Additionally, NVIDIA vGPU software may be used in conjunction with this potent data centre GPU to power AI tasks across virtualised environments and provide distant users with high-performance virtual workstation instances. In the second half of this year, NVIDIA vGPU support for the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU is anticipated.
“NVIDIA GPU-based ray-tracing product, Cyclops, has been tested on the NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition GPU,” stated Martha Tsigkari, senior partner and head of applied research and development at Foster + Partners. The performance of the new NVIDIA Blackwell GPU has surpassed all previous tests. For instance, it has outperformed NVIDIA RTX A6000 GPUs by five times when used with Cyclops. Additionally, rendering speeds rose fivefold, enabling real-time feedback on the performance of the design solutions from tools like Cyclops as it build them.
This led to intuitive yet well-informed decision-making from the very beginning of the conceptual process.GE HealthCare’s senior executive and general manager of platforms and digital solutions, Rekha Ranganathan, said, “RTX PRO 6000 Blackwell technologies’ early review by the engineering team shows promise. for up to 2x GPU processing time improvement on reconstruction algorithms, which could lead to significant benefit for customers.”
According to Jeff Hammoud, chief design officer at Rivian, “NVIDIA RTX PRO 6000 Blackwell Workstation Edition GPUs enable incredibly sharp and photorealistic graphics.” The technology provided the degree of clarity required for immersive vehicle design assessments when used with Autodesk VRED and a Varjo XR4 headset. NVIDIA employed two potent 600W GPUs using VR SLI, which enabled us to reach the maximum pixel density and the most breathtaking images it have ever seen in virtual reality to NVIDIA Blackwell PCIe Gen5 support.
SoftServe industry solutions director Shaun Greene claimed the NVIDIA RTX PRO 6000 Blackwell Workstation Edition GPU “processing power has tripled NVIDIA productivity” with 96GB memory, large AI models like Llama 3.3-70B and Mixtral 8x7b, the NVIDIA Omniverse platform, and industrial copilot. “Since workstations can now manage AI workloads that were previously only feasible in the cloud or on rack servers,NVIDIA seen instant performance improvements. This opens up new possibilities for interactive demos and production workloads in retail, manufacturing, and industrial edge applications.”
A vast ecosystem of AI-accelerated apps developed on NVIDIA CUDA and RTX technology is accelerated by RTX PRO GPUs, which run on the NVIDIA AI platform and have the newest Tensor Cores and more memory. Inferencing is now faster than ever to the newest AI-based content creation tools and new reasoning models, such the NVIDIA Llama Nemotron Reason family of models and NVIDIA NIM microservices that were presented today. Additionally, developers can quickly create, optimise, launch, and scale new AI applications from workstations to the data centre or cloud with the help of more than 400 NVIDIA CUDA-X libraries.
By using an NVIDIA RTX PRO Blackwell GPU for local prototyping and the NVIDIA Omniverse and NVIDIA AI Enterprise platforms, NVIDIA Blueprints, and NVIDIA NIM, which provide enterprise-level support for user-friendly inference microservices, businesses can expedite the development and implementation of AI. With the RTX PRO 6000 Blackwell Server Edition, they can also run these apps at scale on the best universal data centre GPU for AI and visual computing, providing ground-breaking acceleration for the most taxing compute-intensive corporate workloads.
Accessibility
Leading data centre system partners, such as Cisco, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro, will soon offer the NVIDIA RTX PRO 6000 Blackwell Server Edition in server configurations.
AWS, Google Cloud, Microsoft Azure, and CoreWeave are among the first cloud service providers and GPU cloud providers to offer instances powered by the NVIDIA RTX PRO 6000 Blackwell Server Edition later this year. Additionally, ASUS, GIGABYTE, Ingrasys, Quanta Cloud Technology (QCT), and other international system partners will offer the server edition GPU in data centre platforms.
In April, global distribution partners PNY and TD SYNNEX will make the NVIDIA RTX PRO 6000 Blackwell Workstation Edition and NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition available. In May, manufacturers BOXX, Dell, HP Inc., Lambda, and Lenovo will make them available.
In the summer, BOXX, Dell, HP, Lenovo, and international distribution partners will offer the NVIDIA RTX PRO 5000, RTX PRO 4500, and RTX PRO 4000 Blackwell GPUs.
Later this year, NVIDIA RTX PRO Blackwell laptop GPUs will be offered by Dell, HP, Lenovo, and Razer.
