He Raised $70M to Cure Every Disease With AI

The Gradient Descent interview with Sam Rodriguez reveals how his $70 million fundraising round fuels Edison Scientific’s mission to build an AI scientist capable of curing every disease. Starting in 2022, the team launched a multi‑agent system that completed the entire scientific discovery loop—hypothesis generation, experiment design, data analysis—and produced a Nature‑published treatment for dry age‑related macular degeneration. Their latest platform, Cosmos, has already enabled users to generate between 20,000 and 30,000 novel findings, demonstrating that large‑language‑model‑based agents can move beyond text generation to real‑world biomedical breakthroughs.

Rodriguez argues that biology’s talent bottleneck, not capital, limits progress, so he created a focused research organization—a nonprofit that operates like a lean biotech startup. By open‑sourcing tools and attracting philanthropic capital, the model accelerates high‑risk projects such as brain‑connectomics mapping that traditional for‑profits avoid. Once the technology proves viable, a for‑profit spin‑out provides the scaling engine needed to serve pharma executives hungry for AI‑driven drug repurposing. This hybrid approach reshapes the biotech landscape, offering investors a pathway to profit while preserving the open science ethos that fuels rapid innovation.

The conversation also highlights where AI currently shines and where it falters. Verifiable, high‑throughput tasks such as screening thousands of protein‑parasite interactions or drafting experiment plans are handled effortlessly, yet the agents still produce poor prose when asked to write press releases. These limits remind biotech leaders that AI augments—not replaces—human expertise, especially when experimental loops are costly and time‑consuming. As more pharma companies integrate multi‑agent platforms, the competitive advantage will hinge on combining AI’s data‑synthesis speed with rigorous validation pipelines, accelerating the path from discovery to market.