Aivres KR6268 Delivers Robust Compute Power, Memory and Throughput Enterprises Require for AI Industrial Revolution

The Aivres KR6268 6U 8-GPU AI server delivers robust compute power, memory and throughput for the world’s most demanding workloads, from large language model training to advanced AI graphics. How might companies leverage these transformative features in real world deployments?

• KR6268 can deliver universal acceleration for workloads spanning agentic and physical AI to advanced design, scientific computing, analytics, simulation, graphics rendering and video processing.
• With characteristics for sustainable data center operations, KR6268 can enable companies to build AI factories in space-, power- and cooling-constrained environments. These enterprise-ready systems provide breakthrough performance, energy efficiency, and seamless compatibility to integrate with existing infrastructure, allowing for upgrades to the data center without the complexity, cost and disruption of complete overhauls.
• The air-cooled, PCIe-based architectures of the Aivres platforms deliver enhanced security, manageability and serviceability for IT managers in refreshing their companies’ systems.

Aivres will demonstrate its KR6268 AI server in Booth A60 at OCP Global Summit 2025 at the San Jose Convention Center from Oct. 13-16, as part of its “Total Infrastructure Solutions to Power the AI Era” showcase.

The KR6268 incorporates NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, featuring the latest NVIDIA Blackwell architecture innovations: fifth-generation Tensor Cores; second-generation Transformer Engine with support for NVFP4 precision; fourth-generation NVIDIA RTX cores; enterprise-grade scale for multi-user AI deployments, using vGPU and NVIDIA Multi-Instance GPU (MIG) technology for four fully isolated instances per GPU; and improved performance per watt.

The recently announced NVIDIA RTX PRO Servers are a new category of on-premises, enterprise data center infrastructure built on NVIDIA RTX PRO 6000 Blackwell GPUs, bringing incredible performance for agentic and physical AI—up to 6x better performance for LLM inference versus the previous generation L40S. These servers offer transformative capabilities with lower cost of ownership compared to CPU-only systems for accelerated workloads like data analytics, video processing, and rendering —with up to 45x better performance and 18x greater energy efficiency.

This RTX PRO Server from Aivres enables the automation of complex agentic AI tasks using NVIDIA Nemotron models  at up to 3x better price performance on NVIDIA RTX PRO 6000 Blackwell GPU compared to NVIDIA H100. The more accurate reasoning at a lower cost delivered by the Aivres platform enables enterprises not only to supercharge their workloads but also re-imagine their possibilities for the burgeoning AI era: accelerating insights with advanced simulations and AI reasoning, automating industrial processes with AI agents and digital twins, and many more.

Aivres is working with NVIDIA to help the world’s leading-edge enterprises harness the revolutionary power of AI. Companies across industries such as healthcare, manufacturing, geoscience, retail, media and live broadcast, for example, leverage immense compute power, memory and throughput for critical AI applications with Aivres KR6268 and NVIDIA RTX PRO 6000 Blackwell GPUs. Through integration and innovation with NVIDIA’s leading-edge AI technology, Aivres offers AI platforms for every scenario—inference, training, large language models (LLMs), generative AI, digital twins, real-time graphics, etc.—enabling enterprises of any industry to tackle their most challenging workloads.