AI Lab Goodfire Raises $150M for 'Interpretable' Models
AI research lab Goodfire has raised $150 million in a Series B funding round, reaching a valuation of $1.25 billion. The company focuses on designing AI models with built-in interpretability, aiming to address transparency and ethics in artificial intelligence.
- The Series B round was led by B Capital, with participation from new investors including DFJ Growth, Salesforce Ventures, and former Google CEO Eric Schmidt. - This latest funding brings Goodfire's total capital raised to over $200 million, following a $50 million Series A in April 2025 and a $7 million seed round in August 2024. - The company is a public benefit corporation founded in 2023 by CEO Eric Ho and CTO Dan Balsam, who previously worked together at AI recruitment startup RippleMatch, and Chief Scientist Tom McGrath, who founded the interpretability team at Google DeepMind. - Goodfire's team includes researchers and engineers with backgrounds from OpenAI, Google DeepMind, Stanford, and Harvard. - In a practical application of its technology, the lab partnered with Prima Mente to analyze an epigenetic model, leading to the identification of a novel class of Alzheimer's biomarkers. - The company has also demonstrated the ability of its techniques to improve model behavior, successfully reducing hallucinations in a large language model by 50%. - The new capital is intended to help build out the company's "model design environment," a platform designed to allow users to debug and intentionally engineer AI models.
