- Nvidia reveals Blackwell Ultra at GTC 2025
- New flagship GPU offers more power and efficiency than previous generations
- There is also a new Nvidia RTX Pro 6000 Blackwell Server Edition designed for company workload
Nvidia has taken the wrapping of its latest Blackwell Ultra Flagship GPU hardware as it appears to assert its place as the global leader in AI Computing.
The new Blackwell Ultra Hardware revealed at its NVIDIA GTC event in San Jose, offering more power and efficiency in the data center than previous generations, as the company seems to establish itself as the backbone of future AI development and implementation.
Nvidia, who declares that it is “built for the age of the AI Reasoning,” says the Blackwell Ultra can help democratization of AI passing around the world, enabling more organizations to enjoy the benefits that such calculations can cause.
Nvidia Blackwell Ultra
Nvidia noted that the rapid growth of AI use cases all over the world in the last few years has led to a huge demand for calculation, as companies and consumers are both clinging for more.
“AI has made a huge jump – reasoning and agentic AI demand orders more computer performance,” said Jensen Huang, founder and CEO of Nvidia.
“We designed the Blackwell Ultra in this moment-It is a single versatile platform that can easily and effectively perform prior, post-training and reasoning AI inferences.”
The increase in reasoning models in particular has led to a boom in requirements for a full-tuned offer that is cost-effective, but also gets the job done.
Built on the initial Blackwell architecture revealed in the GTC 2024, Blackwell Ultra will offer 1.5x more FP4 inferences and is set to be available built -in by NVIDIA partners in the second half of 2025.
It will be present in a number of new offers from NVIDIA, including the upgraded GB300 NVL72 rack, as it says offers improved energy efficiency and servicability with bandwidth speeds of up to 130 TB/s.
Nvidia RTX Pro 6000 Blackwell Server Edition
However, that was not all when it comes to Ny Blackwell hardware, as the company also revealed the NVIDIA RTX Pro 6000 Blackwell Server Edition, designed for company workload as Multimodal AI -Inferens, immersive content creation and scientific computing.
NVIDIA says the new release offers huge progress on the previous generation of ADA Lovelace architecture L40S GPU, which gives up to 5x higher LLM inference flow to agent AI applications, almost 7x faster genome sequencing, 3.3x speedups for text-to-video generation, almost 2x faster inferring 2x speedups for rendering.
Each RTX Pro 6000 can also be divided into as many as four fully insulated cases with 24 GB each to run simultaneously AI and graphics work loads, with 96 GB of Ultrauty GDDR7 memory and support for Multi-Body GPU.
They can also be configured in high-density of accelerated computer platforms for distributed inference workloads or is used to deliver virtual workstations using NVIDIA VGPU software, for graphics-intensive applications or to operate AI development.
“With RTX Pro 6000 Blackwell Server Edition, companies across different sectors – including architecture, automotive, cloud services, financial services, game development, healthcare, manufacturing, media and entertainment and retailing – enable the performance of the breakfast for work load such as multimodal generative AI, data analysis, technical simulation,” -guupte.
