Here’s the American English translation:
During GTC 2025, NVIDIA announced the first data center GPU based on the Blackwell architecture, specifically designed to accelerate demanding applications in both artificial intelligence and graphics computing. The new NVIDIA RTX PRO 6000 Blackwell Server Edition promises to revolutionize business performance in sectors such as automotive, healthcare, manufacturing, media and entertainment, and financial services.
Unprecedented Performance for Critical Applications
According to NVIDIA, this new GPU delivers up to 5 times higher performance in large language model (LLM) inferences, nearly 7 times faster genomic sequencing, and 3.3 times quicker text-to-video generation compared to the previous Ada Lovelace L40S generation. In rendering tasks and recommendation systems, the improvements exceed double the speed.
Advanced Architecture for 24/7 Environments
The RTX PRO 6000 Blackwell Server Edition features 96 GB of ultra-fast GDDR7 memory and supports Multi-Instance GPU (MIG), allowing for up to four isolated instances of 24 GB each to run simultaneous AI and graphics workloads. Additionally, it is the first universal GPU to incorporate NVIDIA Confidential Computing technology, providing a secure hardware-based execution environment to protect AI models and sensitive data in real time.
Enterprise-Scale Deployment and Applications
The new GPUs easily integrate into high-density accelerated computing platforms for distributed workloads or as virtual workstations using NVIDIA vGPU software. With fifth-generation Tensor cores, fourth-generation RT cores, DLSS 4, and a second-generation Transformer engine compatible with FP4 precision, this GPU offers full support for the NVIDIA Omniverse and NVIDIA AI Enterprise platforms, driving advancements in image and video generation, digital simulations, and robotics.
Leading Companies are Already Investing in the RTX PRO 6000 Blackwell
Companies like Black Forest Labs, developers of the FLUX engine for image generation, have already started working with this new GPU to reduce latency and accelerate AI content creation.
Meanwhile, the cloud graphics company OTOY is optimizing its OctaneRender engine to leverage the performance of this GPU, enabling artists to integrate neural rendering with advanced generative AI.
The firm KLA, specialized in semiconductor manufacturing, will use the GPU to accelerate its inspection and metrology processes with ultra-fast AI algorithms, improving defect detection in wafer production.
Accelerating Genomics and Drug Discovery
In the biomedical field, the RTX PRO 6000 Blackwell Server Edition demonstrates significant improvements in genomic analysis and drug discovery applications. Tools like Fastq2bam and DeepVariant from the NVIDIA Parabricks suite experience up to 1.75 times acceleration compared to H100 GPUs. For the Smith-Waterman algorithm, crucial for sequence alignments, the improvement reaches up to 6.8 times.
Oxford Nanopore Technologies is collaborating with NVIDIA to bring these advancements to its real-time sequencing platforms, achieving double the speed in its basecalling processes on the Dorado platform.
Global Availability Starting in May
Platforms with the RTX PRO 6000 Blackwell Server Edition are expected to be available starting in May through cloud providers such as AWS, Google Cloud, Microsoft Azure, IBM Cloud, CoreWeave, Crusoe, Lambda, and Vultr.
Server manufacturers like Cisco, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro will also offer solutions equipped with this GPU, along with companies such as ASUS, GIGABYTE, Foxconn, Inventec, and Quanta Cloud Technology, among others.
With this move, NVIDIA reaffirms its leadership in creating hybrid solutions that combine graphical power and artificial intelligence processing to transform key sectors of the industry.
via: NVIDIA blog