Enveda
Last reviewed
Jun 8, 2026
Sources
11 citations
Review status
Source-backed
Revision
v1 · 1,653 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 8, 2026
Sources
11 citations
Review status
Source-backed
Revision
v1 · 1,653 words
Add missing citations, update stale details, or suggest a clearer explanation.
Enveda (Enveda Biosciences, legally Enveda Therapeutics) is an American AI drug discovery company that uses machine learning and high-resolution mass spectrometry to decode the chemistry of nature and turn natural molecules into new medicines. Founded in 2019 and headquartered in Boulder, Colorado, the company builds what it describes as one of the world's largest searchable libraries of natural-product molecules and trains foundation models that predict molecular structure and biological activity directly from mass-spectrometry data. In September 2025 Enveda closed a $150 million Series D financing that pushed its valuation above $1 billion, giving it unicorn status, and added former Pfizer chief scientific officer Mikael Dolsten to its board of directors.[1][2]
Enveda's premise is that the natural world contains an enormous, largely uncharacterized chemical space that has been refined by billions of years of evolution but remains difficult to access with conventional laboratory methods. Historically, isolating and identifying a single natural compound from a complex biological sample, such as a plant extract, has been slow and labor-intensive, which led much of the pharmaceutical industry to deprioritize natural-product discovery in favor of synthetic chemistry.[3][4]
The company's approach combines metabolomics, tandem mass spectrometry, and machine learning to annotate the chemical structures of thousands of molecules in a sample at once, without isolating each one individually, and to link those structures to biological activity. Enveda has said this lets it characterize roughly 10,000 molecules at a time rather than one at a time, and that it advances prioritized platform hits to development candidates several times faster than industry averages.[3][4] The output is a drug-discovery pipeline of small molecules that are derived from, or inspired by, natural chemistry.
Enveda was founded in 2019 by Viswa Colluru, who serves as chief executive officer.[1][3] Colluru previously worked at the drug-discovery company Recursion, and several Recursion-affiliated investors backed Enveda early. The company received seed funding in 2020 and is headquartered in Boulder, Colorado, with additional laboratory operations in Hyderabad, India.[5][2]
The company frames its mission around the idea that nature represents the overwhelming majority of possible chemical entities and that most of that chemistry has never been examined for medicinal value. In public statements accompanying its 2025 financing, Enveda described nature as having "been running the most sophisticated R&D program on Earth for billions of years," while "nearly all of its chemistry remains unexplored."[1]
Enveda's platform is built on tandem mass spectrometry (MS/MS), a technique that fragments molecules and measures the masses of the resulting pieces, paired with machine-learning models that interpret those fragmentation patterns. The central challenge the company addresses is that mass spectra are difficult to map back to chemical structures, and only a small fraction of spectra collected in metabolomics experiments are ever confidently identified. Enveda's models aim to "read" spectra and "translate" them into chemical structures and properties at scale.[6][3]
In May 2024, Enveda announced a collaboration with Microsoft and revealed a foundation model called PRISM, short for Pretrained Representations Informed by Spectral Masking. PRISM was trained on roughly 1.2 billion small-molecule MS/MS spectra, which Enveda described as the largest training set of small-molecule mass spectra assembled to that point. About half of the spectra came from public repositories such as GNPS, MetaboLights, and Metabolomics Workbench, and about half from Enveda's internal platform, together representing on the order of 85 billion tokens.[6][7]
PRISM uses a self-supervised technique the company calls Masked Peak Modeling, adapted from the masked-language-modeling approach used by BERT. During training the model randomly masks the masses of about 20 percent of the peaks in each spectrum and learns to predict the missing values from the surrounding context, which Enveda says teaches it the underlying "grammar" of mass-spectrometry data without labeled examples. The learned representations are then used for downstream tasks such as de novo structure prediction and property prediction, where the company has reported relative accuracy gains of roughly 7 to 16 percent on property prediction and about a 23 percent improvement in retrieving close matches from reference spectral libraries.[7]
These models sit on top of a large, curated database of natural chemistry. Enveda has reported assembling a searchable library of natural-product molecules numbering in the millions, drawing on tens of thousands of plant and other natural sources, which it mines computationally to prioritize candidates for drug development.[4][3]
Enveda's lead program is ENV-294, an orally available small molecule that the company says it discovered from nature using its AI platform and that represents a new chemical class with what it has described as "kinase inhibitor-like and steroid-like" behavior. ENV-294 is being developed for inflammatory and immune-mediated conditions, with atopic dermatitis as the lead indication and additional interest in asthma and inflammatory bowel disease.[1][2]
In November 2024, Enveda announced that ENV-294 had become its first nature-derived, AI-discovered candidate to enter clinical trials. In its September 2025 Series D announcement, the company reported that it had dosed the first patient in a Phase 1b trial of ENV-294 in atopic dermatitis, following earlier Phase 1 data, described as available around May 2025, that showed no serious adverse events.[8][1] Beyond ENV-294, Enveda has described a broader pipeline spanning immunology and metabolic disease, including an obesity program built around a hormone-mimetic approach, and at the time of its 2024 Series C it characterized its pipeline as comprising about 10 development candidates, with additional programs and investigational new drug (IND) filings planned.[9][1]
Enveda has raised capital through a series of rounds since 2020. Reported figures, which are drawn from company announcements and trade-press coverage, are summarized below.
| Round | Date | Amount | Lead investor(s) |
|---|---|---|---|
| Seed | 2020 | Undisclosed | True Ventures, Village Global |
| Series A | June 2021 | $51 million | Lux Capital |
| Series B / B1 | December 2022 to April 2023 | $68M, rising to about $119M total | Kinnevik, existing investors |
| Series C | November 2024 | $130 million | Kinnevik, FPV Ventures |
| Series C (extension) | February 2025 | brought Series C to $150 million | Sanofi (strategic investment) |
| Series D | September 2025 | $150 million | Premji Invest |
The $51 million Series A in June 2021 was led by Lux Capital, with participation from Two Sigma Ventures, Hummingbird, Catalio Capital, and existing backers including True Ventures and Wireframe Ventures.[5] The Series B, announced in December 2022 as a $68 million combined equity and debt financing, was extended in April 2023 through a Series B1 that brought the round to roughly $119 million.[10]
In November 2024, Enveda closed a $130 million Series C led by Kinnevik and FPV Ventures, with investors including Baillie Gifford, Premji Invest, Lingotto, Lux Capital, Dimension, True Ventures, Cresset, The Nature Conservancy, and Henry Kravis, bringing total funding at that point to about $360 million.[8][9] In February 2025, Sanofi made a strategic investment that brought total Series C financing to $150 million.[11]
The $150 million Series D, announced on September 4, 2025, was described by the company as oversubscribed and was led by Premji Invest, with participation from new and existing investors including Baillie Gifford, Kinnevik, Lingotto, Peakline Partners, FPV, Socium Ventures, Dimension, Level Ventures, Henry Kravis, IA Ventures, and Lux Capital. The round took Enveda's valuation above $1 billion, and the company reported total funding to date of $517 million.[1][2]
Viswa Colluru is Enveda's founder and chief executive officer.[1][3] Alongside the Series D, Enveda appointed Mikael Dolsten to its board of directors. Dolsten previously served as chief scientific officer and president of worldwide research and development at Pfizer, where he is credited with advancing more than 150 drug candidates into clinical studies and overseeing 36 drug approvals during his tenure.[1][2] His addition was presented by the company as validation of its move from a discovery platform toward a clinical-stage drug developer.
Enveda is frequently cited as a leading example of AI-driven natural-product drug discovery, a field that aims to revive interest in nature as a source of medicines by using computation to overcome the historical bottlenecks of isolating and identifying natural compounds.[3][4] Its PRISM foundation model is notable as one of the largest machine-learning models trained specifically on mass-spectrometry data, and the company's progression of a nature-derived, AI-discovered molecule (ENV-294) into mid-stage clinical testing is one of the more concrete demonstrations that such platforms can yield clinical candidates.[6][8] At the same time, as with other AI biotech companies, Enveda's ultimate significance depends on clinical outcomes that remain to be established, and many of its platform and pipeline metrics are company-reported rather than independently validated.