Clockwork.io Raises $31 Million
Article excerpt
-Clockwork.io raises 31 million dollars; LinkedIn, Together AI, and WhiteFiber adopt its resilience software to prevent wasted GPU hours. LinkedIn prevents tens of thousands of GPU idle hours monthly; new TorchPass innovations preserve AI workload progress without code changes and accelerate reinforcement learning. PALO ALTO, Calif., October 5, 2026 /PRNewswire/ -- Clockwork.io, whose fault-tolerance software keeps AI training, reinforcement learning, and inference workloads running in the face of infrastructure failures, today announced new funding of 31 million dollars, production deployments at LinkedIn and Together AI, and increased adoption by WhiteFiber. The company also introduced two new functionalities for its TorchPass solution, each of which captures the state of a running distributed AI task. Multi-mode platform snapshots, an industry first for training, save the entire task across all nodes without requiring changes to the training code and preserve it for later recovery. Meanwhile, fast, asynchronous application checkpoints, which run in the background while the task is in progress, accelerate reinforcement learning. These supply updated model weights to inference replicas responsible for generating rollouts - the examples from which the model learns - allowing these replicas to spend less time waiting or working with an outdated model.Fault Tolerance Has...
Keep reading with a free account
The rest of this article, and every signal for Together AI, is in your free account.
Extracted from this sentence
Together AI is launching TorchPass as a service on its GPU clusters.
