In a recent development, AI startup CentML has successfully raised $27M in seed funding. The round saw participation from investors, including Nvidia Corporation, Google’s venture fund Gradient Ventures, Radical Ventures, Deloitte Ventures, and Thomson Reuters Ventures.
- Background: Founded by Gennady Pekhimenko, a machine learning PhD holder from Carnegie Mellon University and an Assistant Professor at the University of Toronto, CentML aims to address the critical challenges hindering AI development.
- Addressing GPU Shortages: The startup is working to alleviate the shortage of Nvidia graphic processor units (GPUs), vital for processing vast data quantities necessary for training AI systems. Industry analysts predict that the GPU supply could remain constrained until 2024 due to surging prices.
- Maximizing GPU Utilization: CentML has developed software that helps predict task processing times across different hardware types. This tool can identify underutilized system areas and redistribute tasks to expedite processing, thus enhancing GPU utilization and reducing costs.
- Impact of CentML Technology: Based on the startup’s research, the average GPU utilization in the market is around 30%. CentML’s technology can potentially accelerate systems by up to 8-times, creating a substantial impact on clients.
- Expansion Plans: CentML plans to open a Silicon Valley office and double its employee count from 30 in the coming year.
Pekhimenko, also CentML’s CEO, emphasized the growing gap between computing power and model sizes in AI. He stated, “The size of AI models grew 10 times in the last decade, and there’s a desperation for compute, and chipmakers can’t supply it fast enough.”
Amid these challenges, startups like CentML are being backed to find innovative solutions to overcome these constraints.