NVIDIA has officially confirmed that its flagship RTX PRO 6000 Blackwell Server Edition GPU will arrive later this year in a streamlined 2U server configuration, offered by major enterprise server providers such as Cisco, Dell Technologies, HPE, Lenovo, and Supermicro. This marks a significant step in expanding GPU-accelerated computing to the most common rack-mounted systems in enterprise data centers.
The new 2U systems are designed to replace traditional CPU-only servers, bringing up to 45x higher performance and 18x greater energy efficiency compared to standard 2U CPU configurations. 
This means lower operational costs while dramatically improving processing capabilities for workloads like AI model training, graphics rendering, video processing, simulation, and data analytics.
NVIDIA’s Blackwell architecture underpins these systems, delivering unmatched acceleration for a broad range of enterprise applications – from agentic AI and industrial simulations to content creation and scientific computing. The RTX PRO Server lineup is positioned as the backbone for the NVIDIA AI Data Platform, a reference design for building modern storage and compute infrastructures for AI-driven workflows.
At SIGGRAPH, Dell showcased its updated Dell AI Data Platform, now integrated with NVIDIA’s reference design, alongside the PowerEdge R7725 2U servers equipped with dual RTX PRO 6000 GPUs, NVIDIA AI Enterprise software, and NVIDIA networking solutions. This expanded lineup complements the rack-mounted options first revealed at COMPUTEX, which include systems supporting two, four, or eight RTX PRO 6000 GPUs, enabling enterprises to select configurations based on their performance, efficiency, and budget needs.
With space, power, and cooling efficiency becoming critical in modern data centers, these new 2U GPU-powered servers aim to make AI acceleration more accessible, practical, and cost-effective – helping enterprises transition from CPU-heavy infrastructures to fully accelerated computing environments.