The round
Enveda announced a $311m Series E on 23 September, led by Catalio Capital Management with Iconiq and others participating. The round values the company at $2bn, double where it stood twelve months earlier.
Enveda was founded in 2019 by Viswa Colluru, an early employee at Recursion Pharmaceuticals. Its approach is narrower than most of the AI-drug field: rather than designing molecules computationally from scratch, it uses machine learning to find promising compounds that already exist in plants and microbes, and to work out what they are and what they might treat.
What is in the clinic
The money is for clinical work rather than discovery, which is the more informative part of the announcement. Enveda has several candidates in human trials, including a drug aimed at severe skin conditions and a treatment designed to hold weight off after a patient stops taking GLP-1 medication.
That second one is a well-chosen target. GLP-1 drugs work while people take them, and a substantial share of the weight comes back when they stop. A therapy that addresses the discontinuation problem has an obvious market and a clean endpoint, which makes it a good candidate to test the underlying method against.
The unresolved question
Natural products are how a great deal of medicine was found historically, and the reason the field moved away from them was practical: identifying and characterising a compound from a biological extract was slow and expensive. Automating that with machine learning is a legitimate use of the technology, and it is closer to a measurement problem than a generation problem, which is generally where these methods work better.
The caveat is the one that applies to the whole sector. No drug discovered with AI has yet been approved by the US Food and Drug Administration, or by any other major regulator. The pipeline has been filling for years and the readouts are still ahead.
Enveda is now valued on the assumption that its candidates will clear trials that nobody in this field has yet cleared. A $2bn valuation on a company whose lead programmes are still in human testing prices in success that has not happened. The number that will settle whether the method works is not the raise; it is the first phase-two result from a drug that a model picked out of a plant.
