QumulusAI启动了2026年的AI扩展,使用1 144个NVIDIA Blackwell GPUs,利用500M资金,到年底达到20,000+GPUs。 QumulusAI launched its 2026 AI expansion with 1,144 NVIDIA Blackwell GPUs, using $500M financing to scale to 20,000+ GPUs by year-end.

QumulusAI已经部署了1 144个NVIDIA Blackwell GPUs,使用其5亿美元无追索权筹资机制的第一次缩编,启动了一项2026年扩展计划,到年底时将目标对准20 000多个GPUs。 QumulusAI has deployed 1,144 NVIDIA Blackwell GPUs using the first drawdown from its $500 million non-recourse financing facility, launching a 2026 expansion plan targeting over 20,000 GPUs by year-end. 初始阶段包括760个Blackwell GPUs,另外384个B300 GPUs定于3月下旬交付。 The initial phase includes 760 Blackwell GPUs, with 384 additional B300 GPUs set for late-March delivery. 该公司的融资模式使得能够在没有传统的建筑延误或大笔预付费用的情况下分阶段和具有成本效益地扩大规模。 The company’s financing model enables phased, cost-efficient scaling without traditional construction delays or large upfront costs. QumulusAI通过将荷兰国际发展合作署的先进GPU架构整合为一个超分散的云网络,提供灵活、企业级的AI,计算出保健、金融、媒体、汽车和研究部门之间的使用机会。 By integrating NVIDIA’s advanced GPU architecture into a hyper-distributed cloud network, QumulusAI provides flexible, enterprise-grade AI compute access across healthcare, finance, media, automotive, and research sectors. 计划在未来部署B300和RTX Pro 6000平台,并通过Permian Labs USD.AI平台得到基于链式信贷市场的支持,该平台将机构资本与现实世界的AI基础设施联系起来。 Future deployments of B300 and RTX Pro 6000 platforms are planned, supported by blockchain-based credit markets via Permian Labs’ USD.AI platform, which links institutional capital to real-world AI infrastructure.