AI Compute Foundation Evolves: Aivres Announces 6th Gen AMD EPYC™ General-Purpose Server, AMD Instinct™ MI350P PCIe AI Server for Generative and Enterprise AI
SAN JOSE, California–Advancing AI–July 22, 2026– In lieu of AMD’s Advancing AI 2026 conference, Aivres today announced two new server platforms based on AMD technologies – a general-purpose server powered by the 6th generation AMD EPYC™ “Venice” processor, and an enterprise AI server equipped with AMD Instinct™ MI350P PCIe GPUs, to deliver a robust compute foundation for enterprise generative AI applications.
AI adoption is evolving well past training and inference into the territory of complex agentic workloads, making the CPU increasingly vital for scaling and managing data center infrastructure. The AMD “Venice” processor is purpose-designed to meet the moment: as the industry’s first high-performance compute product built on TSMC’s 2nm advanced process, “Venice” serves as the backbone for data movement, network orchestration, storage management, and system coordination within AI infrastructure. Aivres’ upcoming KR2280-E4 2U server fully unleashes the performance potential of the Venice platform, enabling enterprises to scale from AI experimentation to production-grade deployment with confidence.
This product announcement comes at a pivotal moment for AI infrastructure — the industry conversation has shifted from accelerator-level comparisons to a holistic evaluation of system-level architecture competitiveness. Aligned with this trend, Aivres leverages its general-purpose server platform to translate Venice’s advantages in power efficiency and system balance into tangible AI compute services for customers.
In addition to the general-purpose KR2280-E4 server with the next-generation AMD CPU, Aivres is introducing the KR6268-MI350P, a 6U enterprise AI server equipped with 8 AMD Instinct MI350P PCIe cards which enable organizations to integrate agentic and generative AI capabilities into existing enterprise infrastructure without costly overhauls. The server leverages the MI350P to deliver up to 4.6 PFLOPs peak performance and outstanding throughput while fitting within customers’ current rack, power, and cooling infrastructures to provide a formidable, scalable performance boost at leadership TCO.
Aivres remains committed to delivering high-performance, high-reliability compute infrastructure. These new additions further strengthen its AI product portfolio, empowering enterprises to build flexible, efficient, and sustainable computing capabilities for the generative AI era.