Sophont, a startup building foundation models that reason across pathology slides, brain scans, clinical notes, and lab results, has closed a $9.22 million seed financing. The round was led by Kindred Ventures with participation from Upfront Ventures, Delphi Ventures, and prominent AI figures including Jeff Dean, Logan Kilpatrick, Lukas Biewald, and Clément Delangue.
The company’s platform ingests vast unlabeled clinical datasets to create “multimodal” AI backbones—systems that see the whole patient rather than isolated data types. Later this year, Sophont plans to release its foundational model APIs, enabling med-tech firms and pharma R&D teams to fine-tune capabilities for use cases such as symptom triage, biomarker discovery, and patient selection in clinical trials.
Founded and led by 22-year-old CEO Tanishq Abraham—who earned his PhD at 19—and co-founder/CTO Paul Scotti, formerly Head of NeuroAI at Stability AI, Sophont collaborates with academic medical centers worldwide. It publishes in top venues such as Nature Biomedical Engineering and NeurIPS.
With this oversized seed round—well above typical $1 million to $4 million investments in healthcare AI—Sophont will expand its research team, accelerate model development, and launch pilot programs with leading hospitals, health-tech companies, and pharmaceutical partners.
KEY QUOTE:
“Current medical AI is like the parable of the blind men and the elephant—each model sees one part, but none sees the whole patient. We’re building an AI that can finally see the elephant.”
Tanishq Mathew Abraham, PhD, CEO
