Skip to content
Advertisement
Alterra Labs raises $12.5 million seed round to train medical AI in simulated clinics

Alterra Labs raises $12.5 million seed round to train medical AI in simulated clinics

The New York startup builds simulated medical environments where AI models can fail safely before they reach real patients

Alterra Labs, a New York startup, has closed a $12.5 million seed round to build simulated medical environments for testing and training AI models.

The round was led by Oak HC/FT, with Flare Capital Partners and Company Ventures also participating.

A flight simulator, but for clinical AI

Alterra Labs builds simulated clinical settings, then runs models through medical tasks inside them. The first goal is to find where a model breaks. The second is to turn each breakdown into a training exercise the model can learn from.

Advertisement

Humans with real expertise sit at the center of the process. Doctors and scientists help design the tasks, and they also evaluate the decisions the models make. Their judgments become feedback that developers can use to improve their systems.

The simulations span several areas of healthcare work. They include patient care, imaging, electronic health record (EHR) systems, and clinical trials.

Who’s buying, and who’s backing it

Alterra Labs works with a mix of customers across the health sector. Those include health systems, pharmaceutical companies, biotech firms, and government entities.

The fresh capital is meant to expand the company’s evaluation and training platform.

The investor lineup leans heavily toward healthcare specialists. Oak HC/FT, which led the round, focuses on healthcare and fintech, while Flare Capital Partners is a healthcare-focused firm. Company Ventures rounds out the syndicate.

Alterra Labs has also joined the Responsible AI Institute. There, it participates in the TrustX Health working group, which concentrates on verifying and testing agentic AI technologies in health environments.

Agentic AI refers to systems that take actions on their own rather than simply answering questions. In a medical context, that might mean a model completing steps in a workflow.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.