CoreWeave Expands AI Cloud with NVIDIA HGX B300 for Agentic Workloads
Event summary
- CoreWeave announced the expansion of its AI cloud platform at NVIDIA’s GTC conference, integrating NVIDIA HGX B300 for enhanced performance in AI workloads.
- The new Weights & Biases capabilities streamline reinforcement learning (RL) and agent development workflows.
- NVIDIA HGX B300 offers 2.1 TB of HBM3e memory, a 50% increase over HGX B200 instances, enabling large-scale inference and model training.
- CoreWeave expects to deploy NVIDIA Vera Rubin NVL72 platform and NVIDIA Vera CPU rack in production by the second half of 2026.
The big picture
CoreWeave’s expansion aligns with the industry shift from large-scale training to iterative reinforcement learning, positioning it as a key player in the next phase of production-scale AI. The integration of NVIDIA HGX B300 and advanced Weights & Biases capabilities underscores CoreWeave’s strategic focus on enabling enterprises to build and refine autonomous agents efficiently.
What we're watching
- Performance Scaling
- How CoreWeave’s integration of NVIDIA HGX B300 will affect the scalability and efficiency of AI workloads.
- Market Positioning
- Whether CoreWeave can sustain its leadership in AI-native cloud services amid increasing competition.
- Technological Adoption
- The pace at which enterprises adopt CoreWeave’s advanced development workflows for agentic AI applications.
Related topics