A DPU doesn't run the model a it moves data
A DPU doesn't run the model a it moves data. We plan to target DPUs to offload data movement in the inference pipeline (transfer, networking, staging), not for model compute; the model runs on the NVIDIA and AMD accelerators. The inference layer the category is missing a a single engine, rethought from first principles, that replaces a stack of vendor-locked runtimes. Not a bag of hand-tuned