Paris-based startup Poolside AI has officially announced a $400 million funding round to build a specialized large language model (LLM) designed exclusively for software engineering. The round was co-led by U.S. venture capital firm Bain Capital Ventures and French firm DG (Digital Garage), signaling significant investor confidence in the growing market for vertical AI solutions.
Poolside was co-founded by Jason Warner, the former Chief Technology Officer of GitHub, and Eiso Kant. Their mission is to create a foundational model trained from the ground up on trillions of lines of code and technical documentation. The goal is to develop an AI agent capable of handling complex software development tasks, from initial design to debugging and maintenance, far surpassing the capabilities of current-generation code completion tools.
Unlike generalist models from OpenAI or Google that are trained on a wide range of internet data, Poolside’s model will have a deep, native understanding of programming languages and software architecture. Warner stated that the aim is to create an AI partner that can reason about large, complex codebases and collaborate with human developers on a sophisticated level. This specialization, the company believes, will provide a crucial performance edge in the technical domain.
The massive funding round highlights the intense capital requirements for training frontier AI models and positions Poolside as a serious contender in the AI-assisted coding space, challenging established players like GitHub Copilot, Amazon CodeWhisperer, and Google’s Gemini for Code. The investment will be used to scale the company’s research and engineering teams and to secure the vast computational resources needed for model training. As AI becomes more integrated into the software development lifecycle, specialized models like Poolside’s could fundamentally change how code is written and managed.
English 
Deutsch