NVIDIA has announced that major global enterprises—including Disney, Foxconn, Hitachi, Hyundai Motor Group, Lilly, SAP, and TSMC—have adopted NVIDIA RTX PRO Servers, which are powered by the new RTX PRO 6000 Blackwell Server Edition graphics processing unit (GPU). According to NVIDIA, these servers are designed to provide universal acceleration for AI workloads, enabling customers to transition from general-purpose computing clusters to dedicated AI infrastructure without a full data center overhaul.
The RTX PRO Servers are based on NVIDIA’s Blackwell architecture and are engineered to accelerate a range of workloads integral to enterprise data centers. The platform targets agentic AI, physical AI, advanced design, simulation, graphics, video, and scientific computing applications. Enterprises in manufacturing, digital advertising, aerospace, telecommunications, and silicon design are deploying these servers for faster factory automation, simulation, enterprise AI integration, and AI-driven simulation across silicon and systems.
NVIDIA reports several performance improvements with the RTX PRO 6000 Blackwell Server Edition GPU. For example, the NVIDIA Llama Nemotron Super reasoning model delivers up to three times better price performance when running with NVFP4 on a single RTX PRO 6000 GPU compared to FP8 on NVIDIA H100 GPUs. For industrial and physical AI, these servers can provide up to four times faster performance in digital twin, simulation, and synthetic data workflows compared to servers outfitted with NVIDIA L40S GPUs. The servers are designed to fit into established air-cooled, PCI Express-based, x86 rack environments and support Windows, Linux, and major hypervisors for flexible enterprise deployment.
Software support for NVIDIA RTX PRO Servers includes the NVIDIA AI Enterprise platform, featuring microservices, frameworks, libraries, and tools for data center, cloud, and workstation deployment. The platform integrates with validated designs like NVIDIA Enterprise AI Factory for on-premises AI and the NVIDIA AI Data Platform for modern storage requirements in agentic AI applications. For physical and industrial AI, the servers leverage NVIDIA Omniverse libraries and NVIDIA Cosmos world foundation models to enable digital twins and large-scale synthetic data generation.
NVIDIA RTX PRO Servers are available from systems manufacturers including Cisco, Dell Technologies, HPE, Lenovo, Supermicro, Advantech, Aetina, Aivres, ASRock Rack, ASUS, Compal, Eviden, Foxconn, GIGABYTE, Inventec, MiTAC Computing, MSI, PEGATRON, QCT, Wistron, and Wiwynn. Cloud service providers such as CoreWeave and Google Cloud now offer instances accelerated by the RTX PRO 6000 Blackwell Server Edition GPU, with additional support planned from AWS, Nebius, and Vultr later in 2025.
Source: NVIDIA
