Goodfire AI Raises $150M to Decode How Foundation Models Think

Goodfire AI, an interpretability research lab founded in 2024, has secured $150 million in Series B funding at a $1.25 billion valuation to advance mechanistic interpretability—the field of reverse-engineering neural networks to understand their internal workings. The company, founded by Tom McGrath (who previously co-founded the interpretability team at Google DeepMind), Eric Ho, and Daniel Balsam, is building tools that move mechanistic interpretability from academic demonstrations to production-scale systems.