Augment AI has raised $252 million to develop a sub-100ms inference engine that replaces traditional RAG with a novel memory-mapped architecture for large-scale codebases. The company's approach moves beyond raw code completion to focus on cross-file dependency logic, enabling sub-second architectural refactoring by maintaining state across entire repositories. This funding signals a shift in AI coding tools toward deeper, repository-wide understanding.
Augment AI Secures $252M to Build Next-Gen Code Inference Engine
AI
April 28, 2026 · 1:50 AM