BioStack
Category: AI in Healthcare
BioStack (BioStack Platforms, Inc., YC Spring 2026) turns fragmented, de-identified hospital records into multimodal training datasets, evals and reinforcement-learning environments for medical AI. BioStack was founded in 2025. The company is led by Sanat Mishra. Based in San Francisco, US. Team size: 11-50. Total funding raised: $9.5M. Latest round: Seed. Key investors include Peak XV Partners, DCVC, Afore Capital, Emerson Collective, SV Angel, Verdict Capital, Streamlined Ventures, Y Combinator.
- Founded
- 2025
- Headquarters
- San Francisco, US
- Team size
- 11-50
- Total funding
- $9.5M
Value proposition
Links fragmented clinical data (notes, labs, imaging, medications, treatments, outcomes) from hospitals in the US, India, Brazil and Argentina into longitudinal patient records (100M+ de-identified records). Packages them as post-training data, benchmarks, reward functions and simulation environments so AI models can practise real clinical decision-making.
Products and solutions
Longitudinal multimodal clinical datasets (EHR, X-ray/MRI/CT imaging, pathology, ECGs, telehealth audio transcripts), healthcare RL simulation environments, evals, benchmarks and reward functions for medical AI, custom privacy-preserving data workflows for provider systems, preclinical/drug-discovery data pipelines
Unique value
Real, messy longitudinal patient data from a global hospital network, delivered as workflow-aligned RL environments rather than static datasets. Models diagnose, order tests and adjust treatment, then are scored against real outcomes, guidelines and clinician review.
Target customer
Frontier AI labs, human-data companies, biotech and drug discovery companies, clinical researchers; hospital and provider partners
Industries served
Healthcare, medical AI, biotech/drug discovery, clinical research
Technology advantage
Proprietary hospital partnership network across 4 countries with 100M+ de-identified records and diverse populations. Several hundred thousand imaging studies and millions of telehealth transcripts linked to pathology and outcomes. Founding team combines cancer genomics with GenAI/RL research (CTO has 40+ papers).
How they differentiate
Healthcare-specific, and built on real longitudinal hospital records from both emerging and developed markets rather than generalist data vendors or annotation marketplaces. Offers the full post-training loop (data, environments, evals, rewards) instead of raw datasets alone.
Main competitors
Scale AI / Surge AI (RL environments, post-training data), Mercor (expert data, RL environments), Truveta (de-identified EHR data network). Competitors are analyst-inferred.
Key partnerships
Hospital partners across the US, India, Brazil and Argentina (unnamed), in India, connects cancer patients with an investigational cancer vaccine program
Major milestones
Oct 2025: founded, followed by two pivots (multi-agent biology research systems, then in-house preclinical data generation). 2026: joined Y Combinator Spring 2026 batch. Oct 22, 2026: announced $9.5M seed and a global clinical data infrastructure network.
Growth metrics
100M+ de-identified patient records; 17 customers and prospects including top AI labs (at YC launch); early six-figure clinical data contract with a large human-data company; team size 11 per YC
Market positioning
Early-stage data and post-training infrastructure layer for medical AI, selling to frontier AI labs and biotech
Geographic focus
US-headquartered; data operations in the US, India, Brazil and Argentina
About Sanat Mishra
Cancer genomics researcher at Stanford, Yale, Carnegie Mellon and the Max Planck Institute; early researcher with top AI labs on RL tasks and benchmarks across healthcare and biotech; BS IISER Mohali, MS Carnegie Mellon.
Latest news about BioStack
More AI in Healthcare companies
Official website: https://www.getbiostack.com