Ankit Singhal
| Ankit Singhal | |
| Occupation | Entrepreneur, CEO |
|---|---|
| Known for | Co-founder and CEO of Anthrogen; protein language models |
| Education | Columbia University (B.A., Biophysics and Applied Mathematics) |
Ankit Singhal is an American entrepreneur and the co-founder and CEO of Anthrogen, an artificial intelligence research laboratory focused on training large-scale protein foundation models for biologics discovery and development. The company was part of Y Combinator's Summer 2024 batch and is based in San Francisco.[1] This article concerns Ankit Singhal, the Anthrogen co-founder. Other individuals share this name, including a finance executive at AuthBridge and an energy policy researcher.
Early life and education
Singhal attended Columbia University, where he was a Science Research Fellow and Named Scholar. He studied biophysics and applied mathematics at Columbia College. His research at Columbia focused on non-hemolytic peptide design using generative adversarial networks (GANs), protein thermostability, and chemical assays. That work placed him at the intersection of computational modeling and wet-lab biology before he graduated.[2] No record of published preprints or peer-reviewed papers from this period has been independently confirmed as of early 2026.
Career
Singhal co-founded Anthrogen alongside Connor Lee, who serves as chief technology officer. The company describes itself, in its own materials, as an AI research lab building "the network behind biological intelligence." Anthrogen's core platform is designed to generate novel proteins and peptides with specified functions, using large foundation models trained on protein sequences and structures. The company's stated goal is to speed up protein design processes that have traditionally relied on lengthy trial-and-error experimentation.[1]
In November 2024, Anthrogen was reported to have raised a $4 million seed funding round. The company's stated applications span new therapeutic development, industrial catalysts for sustainable manufacturing, and broader molecular innovation.[1] No subsequent funding announcements had been independently confirmed as of early 2026.
Anthrogen developed and released Odyssey, which the company describes as a family of protein language models scaled to 102 billion parameters. According to Anthrogen, Odyssey supports multi-objective design goals and introduces novel architectural frameworks for scaling and generalization. It's one of the larger parameter counts publicly claimed for a protein-specific language model, though independent benchmarking of the model against alternatives such as ESM or EvolutionaryScale's ESM3 has not been reported in the scientific press. The model's API was made available in early access in 2024, with full availability status as of early 2026 not confirmed by independent sources.[2]
The company's team also includes Steven Strutt, who serves as Head of Platform. Strutt previously worked in molecular and cell biology at the Doudna Lab at the University of California, Berkeley, associated with Jennifer Doudna, where he focused on protein assay development, high-throughput workflows, functional genomics, and CRISPR-Cas9 research.[2]
Notability and sourcing note
As of early 2026, substantive coverage of Singhal and Anthrogen in independent news outlets, peer-reviewed journals, or analyst reports remains limited. The primary sources for this article are Y Combinator's company directory and Anthrogen's own website, both of which are promotional in nature. Readers should treat specific claims about model performance, parameter counts, and funding figures as unverified by third parties until independent reporting confirms them.
References
- ↑ 1.0 1.1 1.2 "Anthrogen – Y Combinator". 'Y Combinator}'. Retrieved 2026-03-18.
- ↑ 2.0 2.1 2.2 "Anthrogen". 'Anthrogen}'. Retrieved 2026-03-18.