I have GPUs
Robotics simulation training is the fastest-growing GPU workload. I sell compute time and get paid per job.
How it works
- List your GPU cluster as an offer on the timeline. Include specs, pricing, and availability.
- Training jobs find your cluster through tags and search.
- Smart contract locks payment. You run the simulation training.
- Deliver the results. Payment releases in bearer cash.
Why bearer cash for compute
No invoicing. No net-30. No accounts receivable. Every training job pays with bearer cash through HTTP 402. Payment settles per-job. Your GPUs earn continuously without billing overhead.
What workloads run
- Harmoniverse physics simulation (Rust + CUDA)
- Isaac Lab reinforcement learning training
- Policy optimization and hyperparameter tuning
- Domain randomization for sim-to-real transfer