Clockwork.io Raises $31M for AI Resilience

Clockwork.io, a specialist in AI fault tolerance, has raised $31 million in new funding after announcing production deployments at LinkedIn and Together AI, and expanded adoption by WhiteFiber. Its LinkPass and TorchPass products are designed to keep AI training, inference and reinforcement learning workloads running through infrastructure failures. The funding brings Clockwork.io’s total capital raised to $73 million. New TorchPass capabilities include multi-node platform snapshots and fast asynchronous application checkpoints.

Keeping GPU Workloads Running

Large distributed AI workloads can span thousands of GPUs, where a failed GPU, dropped link or frozen server can stall an entire job. Clockwork.io positions LinkPass between hardware and workload to reroute traffic around failed links, while TorchPass moves work from a failing GPU to a healthy one.

TorchPass now includes multi-node platform snapshots, which save the state of a running distributed job without changes to training code, and fast asynchronous application checkpoints for reinforcement learning.

Suresh Vasudevan, CEO at Clockwork.io, said:

“Failures are inevitable at AI scale. Losing hours of useful work to them should not be. Fault tolerance is a goodput multiplier: it keeps GPUs doing useful work instead of waiting for recovery or repeating work already done. We built our software alongside enterprises and cloud providers operating some of the largest GPU fleets, so it handles the failures they actually see. That protection belongs in the infrastructure enterprises and cloud providers rely on every day.”

LinkedIn Deployment

LinkedIn has deployed LinkPass network fault tolerance across its AI infrastructure fleet and says it prevents tens of thousands of GPU-hours of downtime each month. The software automatically reroutes traffic onto healthy paths while link, optic, cable or NIC faults are repaired.

Raghu Hiremagalur, SVP, CTO Infrastructure at LinkedIn, said:

“At AI infrastructure scale, a single network issue should never sideline healthy GPUs or interrupt running workloads. Before Clockwork.io, one InfiniBand NIC flap could remove an eight-GPU server from service, while a switch port flap could drain a second server, doubling the impact to 16 GPUs. Clockwork.io helped transform that operating model. Its network fault-tolerance technology automatically reroutes traffic onto healthy paths, allowing jobs to continue uninterrupted while link, optic, cable, or NIC faults are repaired. In aggregate, Clockwork.io prevents tens of thousands of GPU-hours of downtime per month across our fleet. By turning what were once disruptive operational incidents into manageable maintenance events, Clockwork.io has helped improve infrastructure utilization and operational efficiency.”

Reinforcement Learning Support

Together AI is bringing TorchPass to market as a service on its GPU Clusters. The companies plan to demonstrate a live multi-node training job continuing through injected network and GPU failures without restarting at the PyTorch Conference.

For reinforcement learning, TorchPass application checkpoints can move updated model weights to inference replicas sooner, while LinkPass keeps multi-server replicas operating through link failures.

Pavneet Ahluwalia, Product Lead at Together AI, said:

“Our customers grade us on goodput, the share of their GPU-hours that actually move the model forward. Node repair already detects faults and provisions replacement capacity automatically. Clockwork.io's TorchPass and LinkPass build on that foundation and are designed to keep jobs moving through GPU faults and link failures, preserving progress. We are bringing them to market as the next layer of resilience in the platform.”

Funding Supports Wider Rollout

The round was co-led by Premji Invest, Wing Venture Capital and Seligman Ventures. Clockwork.io plans to use the capital to expand its fault-tolerance suite across training, inference and reinforcement learning, grow enterprise adoption and scale delivery through cloud partners.