HostedAI Raises $19 Million in Seed Funding
HostedAI, based in San Jose, California, has announced a successful seed funding round, securing $19 million to further develop its GPU cloud infrastructure solutions. The funding round was led by Creandum, with additional participation from People Ventures and Repeat VC.
Optimizing GPU Utilization
HostedAI provides a software platform designed to improve the efficiency of GPU resources for cloud service providers. The company focuses on GPU orchestration, multi-tenant provisioning, and optimized utilization, enabling service providers to achieve up to five times more profit per GPU. By pooling artificial intelligence workloads on shared GPU resources, HostedAI addresses the significant underutilization often seen in current GPU infrastructure setups.
Industry Context and Challenges
In today's AI-driven landscape, GPUs are typically rented at the card, server, and rack levels, often leading to significant idle capacity. HostedAI CEO Ditlev Bredahl highlighted the inefficiency, noting that customers often use only a fraction of their rented GPU capacity, leaving about 60% of the potential unutilized. This model results in costly overheads for cloud operators, especially those categorized as "neoclouds."
Strategic Use of Funds
The newly raised funds will be directed towards expanding HostedAI's platform capabilities. This includes enhancing GPU virtualization technology, which allows multiple AI workloads to run concurrently on the same GPU. The company also plans to develop additional components such as a GPU resource exchange, known as "GPU Mesh," and a marketplace named GPUaaS.com to connect buyers and providers of GPU resources.
Expanding Reach and Partnerships
With this funding, HostedAI aims to broaden its geographic footprint and strengthen partnerships across the U.S., EMEA, and Asia-Pacific regions. The company's platform not only seeks to reduce capital expenditures for service providers but also aims to lower barriers to entry for regional infrastructure providers, offering an alternative to the dominant hyperscalers in the market.
Conclusion
HostedAI's innovative approach to GPU resource management is set to transform the economics of AI infrastructure. By improving GPU utilization and reducing waste, the company is positioned to offer a more cost-effective solution for service providers and enterprises looking to optimize their AI workloads.