NEW YORK: CoreWeave Inc. (NASDAQ: CRWV), the AI Hyperscaler™, said Thursday it has entered into a definitive agreement to acquire OpenPipe Inc., a developer-focused platform specializing in reinforcement learning (RL) for training AI agents.
The acquisition strengthens CoreWeave’s position in the AI infrastructure market by integrating OpenPipe’s advanced RL frameworks into its high-performance cloud ecosystem. Financial terms of the deal were not disclosed.
OpenPipe’s technology enables developers to post-train AI agents on custom workflows, allowing models to learn from experience and improve accuracy, reliability and performance over time. Its open-source toolkit, Agent Reinforcement Trainer (ART), is widely adopted across AI labs and enterprises seeking scalable, autonomous systems.
“Reinforcement learning is emerging as a pivotal force to strengthen model performance on agentic and reasoning tasks,” said Brian Venturo, CoreWeave’s co-founder and chief strategy officer. “By combining OpenPipe’s self-learning tools with our AI cloud, we’re giving developers a powerful edge in building intelligent systems.”
The deal follows CoreWeave’s recent acquisition of Weights & Biases and marks another step in its strategy to vertically integrate its AI stack. With OpenPipe’s RL and fine-tuning capabilities, CoreWeave aims to offer greater flexibility for customers training and adapting models to novel tasks.
“Together with CoreWeave, we can accelerate our vision and bring our technology to a wider set of enterprises eager to build more reliable, higher-performing and cost-effective AI systems,” said Kyle Corbitt, OpenPipe’s co-founder and CEO.
CoreWeave’s platform is purpose-built for AI workloads, offering scale and performance optimized for innovation across industries.