Aarna.ml announced the release of Version 2.0 of its GPU Cloud Management Software (CMS). This new version is designed to transform how AI Cloud providers—both public and private—optimise and scale their GPU-as-a-Service (GPUaaS). By enabling multi-tenant sharing of GPU pools, Aarna.ml’s software allows cloud providers to boost internal rates of return (IRR) by over 140% while increasing revenue and enabling users to efficiently deploy and scale their AI workloads with hyperscaler-grade cloud features.
The latest release equips AI cloud providers with tools to offer on-demand, fully isolated bare metal, virtual machine (VM), and container instances. This flexibility is crucial for public and private AI cloud providers seeking to maximise their return on investment on GPU infrastructure and extend the lifespan of their investments.
“As inferencing demand grows and aging GPUs become more prevalent, AI cloud providers will need to offer more than just reserved instances,” said Mick McNeil, Group Chief Revenue Officer at Northern Data Group. “The industry needs solutions that allow providers to offer multi-tenant on-demand GPU instances, addressing this pressing market need.”
“Beyond multi-tenancy and instance creation, Version 2.0 introduces advanced Day 2 capabilities such as observability, scale testing, and fault management,” said Amar Kapadia, co-founder and CEO at Aarna.ml. “These features, combined with cloud brokering, position our solution as the only enterprise-grade GPU cloud management offering in the market.”
Version 2.0 of Aarna.ml’s GPU CMS includes advanced features for both administrators and tenants. Administrators can onboard GPUs, create tenants, and access unified inventory and observability tools. Tenant capabilities include instance creation, management, and enhanced monitoring features. Additionally, the product includes cloud brokering features that enable AI cloud providers to trade capacity through a marketplace. The software is specifically designed to support the NVIDIA Cloud Platform (NCP) and NVIDIA Storage reference architectures.
“The availability of Aarna.ml’s GPU CMS 2.0 marks a milestone in the nascent GPU Cloud Management Software landscape,” said Roy Chua, Founder and Principal at AvidThink. “For AI cloud providers and enterprises grappling with GPU resource optimization, this solution offers a viable way to unlock the full potential of their high-value GPU investments.”