Clockwork.io raises $31M to keep AI clusters running through failures

Clockwork Systems raised $31 million (total $73 million) for software that boosts GPU cluster efficiency. The startup launched TorchSnap, multinode snapshots of distributed AI workloads that need no code changes and complement its LinkPass and TorchPass tools.

• Round co-led by Seligman Ventures, Wing Ventures and Premji Invest; total funding now $73M
• TorchSnap captures multinode inference snapshots without developer code modifications
• Cluster recovery after failures typically leaves GPUs idle for up to 90 minutes
• LinkedIn deployed LinkPass across its GPU fleet, saving thousands of GPU-hours monthly