With this year’s being the 2nd year of the NVIDIA Blackwell product cycle, aside from the enterprise world getting the more powerful “Blackwell Ultra” chips, the standard Blackwell has slowly trickled down to the “consumer/prosumer/SMEs” line of market with the release of several new products by Team Green over at GTC 2025.
RTX PRO 6000 Blackwell Server Edition
We have the RTX PRO 6000 Blackwell Server Edition designed for data centers, and it is the first GPU created for both enterprise AI and visual computing. Compared to its predecessor, the Ada Lovelace architecture L40S GPU, this new server edition delivers remarkable performance improvements of up to 5x higher throughput for LLMs in agentic AI applications, close to 7x faster genomics sequencing, 3.3x times quicker text-to-video generation, and nearly 2x the speed for recommender systems inference as well as more than 2x the rendering performance.
This is done through the use of new NVIDIA Streaming Multiprocessor tech, 4th-Gen RT Cores for RTX Mega Geometry, 5th-Gen Tensor Cores, and DLSS 4 alongside the use of high-capacity GDDR7 VRAM, the latest NVENC/NVDEC for video data, and PCIe 5.0 support.
Multi-Instance GPU
Aside from the RTX PRO 6000 Blackwell Server Edition, this feature is also enabled for desktop variants as well as 5000 series offerings that provide secure partitioning of a single GPU into up to 4 instances for the 6000 series or 2 instances for the 5000 series, allowing the efficient use of resources and safely isolated operating environment for diverse workloads.
Goodbye Project DIGITS, Hello DGX Spark and DGX Station
Remember Project DIGITS announced a couple of months ago said to bring the power of Blackwell to the homes of normal people like you and me? Well, the program has been rebranded and now consists of 2 models – The DGX Spark and DGX Station.
For DGX Spark, this is the “Blackwell at home” machine powered by the GB10 Grace Blackwell Superchip optimized for desktop form factors, these devices are capable of delivering up to 1,000 trillion operations per second of AI computing for fine-tuning and inference with the latest AI reasoning models further supported by NVLink-C2C interconnect technology to deliver a CPU+GPU-coherent memory model.
The DGX Station is another story as this is a giant box containing a GB300 Grace Blackwell Ultra Desktop Superchip and packs 784GB coherent memory for large-scale training and inferencing workloads with connectivity handled by the ConnectX-8 SuperNIC for up to 800Gbps transfers, unlocking even more performance through interlinking multiple DGX Stations with ease.
Official Team Green Sticker for Enterprise Storage
For those who want everything “officially recognized” for best compatibility across different industrial offerings, NVIDIA will be helping with that by expanding its certification program to include enterprise storage certification, ensuring streamlined AI factory deployments with accelerated computing, networking, software, and storage. The new NVIDIA-Certified Storage program validates that enterprise storage systems meet stringent performance and scalability data requirements for AI and high-performance computing workloads. Leading enterprise data platform and storage providers including DDN, Dell Technologies, Hewlett Packard Enterprise, Hitachi Vantara, IBM, NetApp, Nutanix, Pure Storage, VAST Data, and WEKA are already participating in this program.
This certification program builds upon existing NVIDIA DGX systems and NVIDIA Cloud Partner storage programs, expanding the data ecosystem for AI infrastructure. The certification programs are aligned with their deployment models and architectures, including NVIDIA DGX BasePOD and DGX SuperPOD Storage Certification for enterprise AI factory deployments with NVIDIA DGX systems, NCP Storage Certification for large-scale NCP Reference Architecture AI factory deployments with cloud providers, and NVIDIA-Certified Storage Certification for enterprise AI factory deployments with NVIDIA-Certified servers available from global partners.
Availability
The NVIDIA RTX PRO 6000 Blackwell Server Edition will soon be available in server configurations from leading data center system partners including Cisco, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro. Cloud service providers and GPU cloud providers including AWS, Google Cloud, Microsoft Azure, and CoreWeave will be among the first to offer instances powered by this server edition later this year. The server edition GPU will also be available in data center platforms from ASUS, GIGABYTE, Ingrasys, Quanta Cloud Technology, and other global system partners.
Meanwhile, the RTX PRO 6000 Blackwell Workstation Edition and Max-Q Workstation Edition will be available through global distribution partners such as PNY and TD SYNNEX starting in April, with availability from manufacturers like BOXX, Dell, HP Inc., Lambda, and Lenovo beginning in May.
The NVIDIA RTX PRO 5000, RTX PRO 4500, and RTX PRO 4000 Blackwell GPUs will be available in the summer from BOXX, Dell, HP, and Lenovo and through global distribution partners, while RTX PRO Blackwell laptop GPUs will be available from Dell, HP, Lenovo, and Razer later this year.
