Toronto-based CentML, a cutting-edge startup focused on optimizing computing power for AI development, has successfully raised $27 million in a seed round of funding. The investment was led by prominent backers including Nvidia Corp. and Gradient Ventures, a Google venture fund. Deloitte Ventures and Thomson Reuters Ventures also participated in this crucial funding round. CentML aims to address the persistent shortage of graphic processor units (GPUs), a bottleneck that has hindered the rapid growth of the AI field.
Solving AI development bottlenecks
CentML, established just last year by Gennady Pekhimenko, a PhD in machine learning from Carnegie Mellon University and now an assistant professor at the University of Toronto’s computer science department, has been gaining attention for its innovative approach to AI development. The startup focuses on building software solutions that enhance the efficiency of machine learning systems.
One of the major challenges facing AI development is the limited availability of GPUs from manufacturers like Nvidia. These specialized processors are essential for handling the vast amounts of data required to train and operate AI systems. Analysts predict that the supply shortage for GPUs, coupled with soaring prices, may persist until at least 2024. Recognizing this challenge, prominent organizations are placing their bets on startups like CentML to devise innovative solutions.
CentML’s software aims to tackle the GPU shortage by optimizing the utilization of available hardware resources. The technology developed by CentML allows for the prediction of processing times for various tasks using different hardware setups. It constantly monitors systems to identify areas of underutilization, considering factors such as cost, power consumption, and emissions.
Once inefficiencies are pinpointed, the software automatically redistributes tasks to expedite processing, thereby maximizing chip usage and reducing costs. According to CentML, the average utilization rate for GPUs across the market currently stands at approximately 30%, highlighting the potential for substantial improvements. The startup claims that its technology can accelerate systems by up to 8 times, a transformational advancement for its customers.
Gennady Pekhimenko, the CEO of CentML, emphasized the growing disparity between the increasing size of AI models and the available computing power. He stated, “The size of AI models grew 10 times in the last decade, and the gap between compute and the size of the model is growing. There’s a desperation for compute, and chipmakers can’t supply it fast enough.” This recognition of the urgency of the situation underscores the significance of CentML’s innovative approach.
Investment backing for expansion
With the successful seed funding round, CentML is now poised for expansion. The startup plans to establish an office in Silicon Valley, a strategic move to attract top talent in the tech industry. Over the next 12 months, CentML aims to double its workforce, currently consisting of approximately 30 individuals. This expansion will enable the company to further develop its software solutions and continue addressing the critical challenge of GPU shortages in the AI sector.
