flâneur

Better Together: How Voltage Park Built NVIDIA SuperPODs for Cursor’s RL Training

voltagepark.com · 647 words · saved by 1 readers

Scaling agentic AI for the next era of agentic coding demands infrastructure that supports reinforcement-learning pipelines and fast experimentation cycles. When Cursor needed a partner to design, deploy, and operate the compute foundation for their next venture, they trusted Voltage Park to architect a customized NVIDIA HGX B200 SuperPOD that didn’t require a provider-defined one-size-fits-all software environment. Voltage Park’s partnership with Cursor began more than a year ago with 128 HGX H100 GPUs. It has now grown into multiple InfiniBand-interconnected Large SuperPODs spanning across the Hopper and Blackwell hardware generations. Our teams along with NVIDIA and Dell, engineered a fleet specifically for the reinforcement-learning workloads Cursor runs. Jointly architected, Voltage Park operates the customized AI infrastructure, provides infrastructure observability, and 24/7 support for the environment. Composer, Cursor’s agentic model for software coding, depends on reinforceme

Better Together: How Voltage Park Built NVIDIA SuperPODs for Cursor’s RL Training Voltage Park has merged with Lightning AI. One place to build and ship AI. GraphN Sign In On-Demand Sign In AI Cloud Why Voltage Park Pricing Resources Company Back to Blog AI Cloud December 2025 December 2025 Better Together: How Voltage Park Built NVIDIA SuperPODs for Cursor’s RL Training Share Scaling agentic AI for the next era of agentic coding demands infrastructure that supports reinforcement-learning pipelines and fast experimentation cycles. When Cursor needed a partner to design, deploy, and operate the

related reading