Weights & Biases

CoreWeave Fully Connected 2026: Everything Announced in 10 Minutes

Published 2026-10-13 · Duration 9:35

Summary

CoreWeave announced a comprehensive platform, CoreWeave Forge, designed to manage the entire AI development lifecycle—the 'AI loop' (Run, observe, curate, improve, evaluate). Key announcements include the launch of CoreWeave Forge, the production readiness of the NVIDIA Vera Rubin NVL72, and new tools like CoreWeave Agent Lens and Model Distillation, which aim to improve agent reliability and streamline MLOps for AI applications.

Download summary

Key takeaways

  1. CoreWeave Forge Launch

    CoreWeave Forge centralizes the entire AI loop (Run, observe, curate, improve, evaluate) into one connected environment, allowing users to build with chosen models, frameworks, and clouds. It is available for self-service sign-up and starts with a 30-day Pro trial.

  2. NVIDIA Vera Rubin NVL72 Production Readiness 3:49

    The Vera Rubin NVL72 is announced as ready for production, with Cognition serving as the first customer. Performance comparisons showed a 4.8x improvement in throughput for inference and a 3.8x improvement in training workloads compared to previous generations.

  3. CoreWeave Agent Lens and Model Distillation 5:50

    CoreWeave Agent Lens helps understand agent behavior in production by grouping related failures into patterns. CoreWeave Model Distillation allows teams to use curated production traces to create smaller, specialized models.

  4. RL Rollouts for Reinforcement Learning 7:04

    RL Rollouts addresses the slow learning cycle in reinforcement learning by allowing weights to be loaded directly into a running deployment when a new checkpoint is ready, minimizing waiting time for serving.

Technical details

  • AI Loop Process 0s

    The core operational cycle is defined as: Run, observe, curate, improve, evaluate, and repeat. CoreWeave Forge integrates these steps into a single platform.

  • Performance Metrics (Vera Rubin) 370s

    Inference throughput improvement was measured at 4.8 times, and reinforcement learning rollout improvement was measured at 3.8 times, demonstrating cost savings and efficiency gains.

  • CoreWeave Sandboxes 424s

    Sandboxes provide an isolated environment necessary for running and scaling multiple agents in parallel, enabling the full AI loop demonstration.

  • Vera CPU Integration 301s

    Vera CPU is coming soon to the CoreWeave cloud, offering a tightly integrated platform where GPUs and CPUs are designed to work together.

Mentioned resources

  • CoreWeave Forge (Platform)
  • CoreWeave Builders Resource Center (Documentation/Demos)
  • CoreWeave Arena (Workload Testing Platform)

Channel & topics

Watch on YouTube · Back to latest

This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.