CoreWeave, which provides cloud servers for large companies training AI models, has reached a consensus to obtain openpipeThe companies announced Wednesday that it is a startup powered by the 2-year-old Y portfolio that helps businesses develop customized AI agents through reinforcement learning.
“Reinforcement learning is becoming a key force to enhance model performance for proxy and inference tasks,” said Brian Venturo, co-founder of CoreWeave. “By combining OpenPipe’s advanced self-learning tools with CoreWeave’s high-performance AI cloud, we are expanding the platform to give AI Labs developers a significant advantage in building scalable intelligent systems.”
CoreWeave and OpenPipe did not disclose the terms of the deal. OpenPipe in Seattle in March 2024 grow up The $6.7 million seed round includes Costanoa Ventures, Y Combinator, Google DeepMind’s Logan Kilpatrick, Github co-founder Tom Preston-Werner, and co-creator of Github Copilot Alex Alex Graveley.
The deal marks the latest attempt at Coreweave to expand its stack, Weights and biases after acquiring AI developer platformMarch. OpenPipe has developed a popular open source toolkit for creating AI agents called ART (Agent Enhancement Trainer). While many of Coreweave’s biggest customers include leading AI labs such as OpenAI, the company is also trying to attract smaller businesses.
More and more AI labs and startups are building enterprise products around enhanced learning, which involves rewarding for the right-responsive AI model. Reinforcement learning has proven to be a powerful way to improve the performance of AI models on specific tasks. The idea of these enterprise products is to train AI agents specifically for company needs.
This kind of customer-specific training requires a lot of computing resources, and by obtaining OpenPipe, CoreWeave wants to have both power and provide such services. OpenPipe’s team will join CoreWeave, and OpenPipe’s customers will become Coreweave customers.
