NVIDIA marks a major turning point in accelerating AI and graphics workloads with the launch of new data center GPUs.

NVIDIA announced on the 21st a new data center GPU, 'RTX PRO 6000 Blackwell Server Edition', leading the next-generation technological innovation.

This is the first Blackwell-based data center GPU for visual computing and enterprise AI, and is expected to revolutionize the acceleration of workloads across a wide range of industries.

The RTX PRO 6000 delivers incredible performance improvements over the existing NVIDIA L40S GPU. The inference throughput of the Large Language Model (LLM) has increased by up to 5 times, the speed of genome sequencing has increased by about 7 times, text-to-video generation has been accelerated by 3.3 times, and the rendering speed has been accelerated by more than 2 times.

This performance is expected to open up new possibilities for data analysis and simulation in various industries such as AI, semiconductors, and life sciences.

The RTX PRO 6000 is designed to support AI and graphics workloads simultaneously, featuring 96GB of high-speed GDDR7 memory and multi-instance GPU (MIG) capabilities.

This allows the GPU to be divided into four isolated instances, providing a flexible working environment.

Additionally, hardware-based security is enhanced through NVIDIA Confidential Computing to safely protect sensitive data.

The RTX PRO 6000 is used in various fields such as architecture, automobiles, finance, medicine, game development, and media, and is expected to perform particularly well in genome analysis, new drug development, content creation, and semiconductor manufacturing processes.

For example, the processing speed of the Smith-Waterman algorithm is 6.8 times faster than that of the L40S GPU, and the performance of AI models for protein structure prediction in new drug development is improved by up to 4.8 times.

The RTX PRO 6000 will be available starting in May through global cloud providers including Amazon Web Services (AWS), Google Cloud, and Microsoft Azure.

We also plan to introduce server platforms in cooperation with major system partners such as Cisco, Dell, and Lenovo.