Zhi Tong Cai Jing·2025-12-17 13:52Core Insights - Nebius has launched the Nebius AI Cloud version 3.1, integrating Nvidia's next-generation Blackwell Ultra computing technology and enhanced operational capabilities into its AI cloud platform [1] - The company is deploying Nvidia's Blackwell Ultra infrastructure globally, with the GB300NVL72 and HGX B300 systems already available for customers, making it the first cloud service provider in Europe to operate both platforms in a production environment [1] - Nebius is the first globally to utilize the 800Gbps Nvidia Quantum-X800 InfiniBand interconnect technology for running the GB300NVL72 system, doubling the throughput for distributed workloads [1] Capacity Management Features - Version 3.1 introduces Capacity Blocks and a Capacity Dashboard to meet enterprise demands for operational transparency, allowing customers to monitor reserved GPU capacity and availability across all data center regions [2] - The update builds on the Nebius AI cloud "Aether," enhancing transparent capacity management and expanded infrastructure to provide operational visibility and resource planning capabilities for AI applications [2] - Aether 3.1 also offers improved AI/ML development tools, expanded ecosystem integration, and refined enterprise governance controls, accelerating value realization while reducing risks [2] Competitive Landscape - CoreWeave, a competitor of Nebius, also offers Nvidia's Blackwell chips for AI infrastructure and became the first cloud platform to launch Nvidia RTX PRO6000 Blackwell server instances in July [3] - CoreWeave provides a wide range of Nvidia Blackwell infrastructure products, including the GB200NVL72 system and the HGX B200 platform [3]