The Long Run• Apr 21, 2026• 1:10:08Interview

Ep199: Martin Burke on Making Small Molecule Medicines for the AI Era

From The Long Run

Martin Burke•Professor of Chemistry and Director of the Molecule Maker Lab at the University of Illinois at Urbana-Champaign

Executive Summary

Martin Burke's pioneering 'block chemistry' enables the rapid, automated synthesis of small molecules by using a standardized, iterative process, significantly reducing a major bottleneck in drug discovery.
His latest startup, Excelsior Sciences, has raised $95 million to build a platform that combines this automated synthesis with AI, treating molecules as 'tokens' to create foundation models for predicting chemical properties.
This integrated platform has reduced the design-synthesize-test cycle in drug discovery from months to a matter of days, enabling a Silicon Valley-style approach of rapid iteration.
Burke's academic research has led to multiple startups and clinical candidates, including a next-generation antifungal (now at Elyon) that avoids the kidney toxicity of its predecessor, amphotericin B.

12 quotes

Concerns Raised

The historically slow, manual, and bottleneck-prone nature of traditional organic synthesis.
The severe toxicity of powerful existing drugs like amphotericin ('amphoterrible').
The high failure rate and strategic pivots of biotech startups can delay or derail promising science.

Opportunities Identified

Dramatically accelerating drug discovery by reducing iteration cycles from months to days.
Creating the first true foundation models for chemistry by linking automated synthesis with AI.
Developing novel 'molecular prosthetics' to treat diseases of protein deficiency.
Democratizing molecular synthesis to unlock innovation from a broader scientific community.

Key Themes

Automating Small Molecule Synthesis

The core innovation discussed is 'block chemistry,' a method for making complex molecules by repeatedly adding simple chemical building blocks. This modular approach is highly amenable to automation, transforming the slow, bespoke art of organic synthesis into a fast, predictable, and scalable process.

Automating synthesis removes a critical bottleneck in drug discovery, enabling faster iteration cycles and making it feasible to generate the large, high-quality datasets required for modern AI/ML approaches.

The Convergence of AI and Physical Automation in Drug Discovery

The conversation highlights a new paradigm where automated robotic labs physically create molecules, and the resulting data is fed back to AI models in a rapid feedback loop. Excelsior Sciences is designed to treat chemical blocks as 'tokens' for AI, analogous to how large language models learn from text, with the goal of building predictive foundation models for chemistry.

This convergence promises to dramatically accelerate the discovery of new medicines by moving beyond in-silico prediction to a system that can rapidly test and validate those predictions in the real world, creating a true learning cycle.

Function-First Drug Discovery and Molecular Prosthetics

Burke's work is driven by a philosophy of creating small molecules that can perform protein-like functions, a concept he calls 'molecular prosthetics.' This approach aims to restore missing functions (e.g., in cystic fibrosis) rather than just inhibiting overactive ones, expanding the therapeutic potential of small molecules.

This represents a shift in therapeutic strategy, opening up new ways to address diseases of deficiency and potentially tackle targets previously considered 'undruggable' with conventional small molecule inhibitors.

The Non-Linear Path of Academic Entrepreneurship

The discussion traces the journey of Burke's discoveries from university lab to multiple startups (Revolution Medicines, Funga/Elyon, Excelsior). It illustrates the complex realities of commercialization, including company pivots, terminated programs, and intellectual property returning to the university before finding a new home.

This provides a realistic view of biotech translation, showing that valuable science can persist through business failures and strategic shifts to eventually reach patients, highlighting the importance of resilience and a robust academic-industrial ecosystem.

Democratizing Molecular Innovation

A long-term vision expressed is to break the 'specialist bottleneck' in synthetic organic chemistry. By making the tools to create molecules automated and user-friendly, the field could be opened up to a much wider range of scientists and innovators, sparking an explosion of creativity.

Lowering the barrier to entry for creating novel molecules could fundamentally change innovation not just in medicine, but also in materials science, agriculture, and other industries, much like standardized tools democratized software development.

Get started free

Topics

Martin Burke Block Chemistry Automated Synthesis Small Molecule Drug Discovery AI in Drug Discovery Foundation Models Medicinal Chemistry Molecular Prosthetics Excelsior Sciences Revolution Medicines Amphotericin B Antifungals Cystic Fibrosis Techbio Biotech Startups

Processed May 4, 2026 yt-dlp + mlx-whisper + Gemini

You're reading a preview

Get started free →