CoreWeave targets GPU utilization in continuous AI post-training
GPU utilization during post-training depends partly on how efficiently infrastructure moves data and loads updated models. As enterprises continually refine AI agents, reducing delays between training rounds can help keep that process moving. The need for continual improvement is pushing CoreWeave Inc. to build a full-stack AI cloud for the agent lifecycle. You.com Inc. provides […] The post…
CoreWeave, a GPU cloud provider, is focusing on maximizing GPU utilization during continuous AI model improvement post-training. This is crucial for enterprises as they refine AI agents, aiming to keep the process efficient. CoreWeave's new platform, Forge, incorporates model deployment, evaluation, and improvement capabilities.
The company's reinforcement learning Rollouts feature, currently in preview, allows for repeated cycles of generating training responses and updating models. To enhance data movement and GPU utilization, CoreWeave AI Object Storage now enables cross-region writes, allowing post-training jobs to write results back for others to utilize.
You.com, an AI-driven search company, has partnered with CoreWeave to provide agents with a search layer. Using RL Rollouts, they successfully post-trained Nemotron 3.5 Lightning in eight hours, demonstrating the potential for high-accuracy results and lower total cost of ownership. This rapid improvement is no longer limited to just frontier labs, according to CoreWeave's chief product officer, Saurabh Sharma.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
