QumulusAI secures $124M AI infrastructure deals
Analysis based on 6 articles · First reported Jun 11, 2026 · Last updated Jun 11, 2026
The agreements establish long-term recurring revenue for QumulusAI, positively impacting its financial stability and growth prospects. This also signals strong demand for specialized AI inference infrastructure, benefiting hardware providers like Nvidia, Lenovo, Supermicro, and Cisco.
QumulusAI, an AI cloud infrastructure company, announced securing over $124 million in 3-year customer subscriptions, including $21.9 million in upfront commitments. These agreements are with Hyperbolic and another leading AI inference platform, supporting deployments of 1,280 Nvidia Blackwell GPUs. QumulusAI will provision 160 Lenovo and Supermicro bare-metal servers, with Cisco Nexus One powering the networking. The workload-optimized deployments aim to reduce AI inference costs by approximately 20% compared to standard configurations, addressing the specific needs of large-scale open-source AI models. Mike Maniscalco, CEO of QumulusAI, highlighted the shift from one-size-fits-all AI infrastructure, while Jasper Zhang, CEO of Hyperbolic, emphasized the performance, efficiency, and scalability provided by QumulusAI's infrastructure.
Set up alerts, explore entity relationships, search across thousands of events, and build custom intelligence feeds.
Open Dashboard