Clockwork.io Raises $31 Million to Eliminate GPU Waste in AI Workloads and Training
One tiny infrastructure glitch can literally halt thousands of interconnected GPUs. While the system agonizes through a slow, painful reboot, massive amounts of money practically evaporate into thin air. To The post Clockwork.io Raises $31 Million to Eliminate GPU Waste in AI Workloads and Training appeared first on Ventureburn .
Clockwork.io has secured $31 million in funding to develop software that minimizes GPU waste in AI workloads and training processes. The startup aims to create resilient software capable of catching training jobs before hardware failures occur. The cost of a single glitch in AI workloads can be astronomical, particularly when companies train models across thousands of interconnected GPUs.
Historically, the only fix was to restart the entire training job, resulting in the loss of hours of computations. Clockwork.io's flagship tools, TorchPass and LinkPass, serve as invisible buffers that protect workloads from hardware failures. If a network link fails, LinkPass instantly reroutes traffic, while TorchPass transfers ongoing work to a healthy GPU.
A recent addition allows for multi-node job snapshots without requiring code changes. Major industry players, including LinkedIn, Together AI, and WhiteFiber, have already adopted this technology to prevent GPU waste and ensure uninterrupted compute power. With this new funding, led by Premji Invest and Wing Venture Capital, Clockwork.io's total funding now stands at $73 million.
The company plans to use the capital to revolutionize infrastructure resilience, a critical need as AI models continue to grow in size.
Written by urgent.news from Ventureburn's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
