signal/raiseTrack rounds
SeedHealthtech & bio · San Francisco

BioStack raises $9.5m seed round to turn hospital records into training data for medical AI

The year-old Y Combinator company says it has gathered more than 100 million de-identified patient records from hospitals in the US, India, Brazil and Argentina, and is backed by Peak XV, DCVC and Emerson Collective among others.

s/rFunding deskNorth America · 2 min read

BioStack, a San Francisco start-up that turns scattered hospital data into datasets for training medical AI, has raised $9.5m in seed funding. The company announced the round on 22 October 2026.

The company's release said the round was led by Peak XV, Afore Capital, DCVC, Emerson Collective, SV Angel, Verdict Capital, Streamlined Ventures and Y Combinator. It listed all eight firms under that wording and did not single out a main lead or separate the leads from other participants. Dealroom, which reported the raise citing chief executive Sanat Mishra's comments to Axios, did not name the investors.

BioStack did not disclose a valuation.

“The next frontier of healthcare AI will depend not only on better models, but on access to high-quality, longitudinal data that reflects how patients are treated over time.”Pushpak Kedia, Vice President, Peak XV

What BioStack does

Patient information is usually split across hospitals, clinics and separate IT systems. BioStack builds workflows that pull together notes, lab results, imaging, medications, treatments and outcomes for each patient, then removes identifying details to create its own datasets. It sells access to AI and biotech companies, which can use the data to build and test models, run clinical research, model disease, look for biomarkers, group patients and study how they respond to treatment.

The company says its network holds more than 100 million de-identified patient records from hospitals in the US, India, Brazil and Argentina. It said it has combined several hundred thousand X-rays, MRI and CT scans with pathology and clinical records, and it pitches the result as material for training AI agents that help doctors weigh treatment options. Dealroom described the product as datasets and training environments in which AI labs can train models on clinical tasks.

In India, BioStack said it is helping to match cancer patients with an investigational cancer vaccine programme, so that their data can support drug developers while eligible patients may get access to the experimental treatment.

Dealroom noted that BioStack's scoring pipeline and accuracy have not been independently checked and are described only in the company's own materials. The release named no customers.

Who is behind it

BioStack was founded in October 2025 and has about 10 staff, according to Dealroom. It went through Y Combinator's Spring 2026 batch.

Mishra, the chief executive, is a former cancer genomics researcher. His co-founder Parth Patwa, the chief technology officer, previously worked as a generative AI scientist at Amazon Web Services, Dealroom reported.

Pushpak Kedia, a vice president at Peak XV, said in the release that progress in healthcare AI would depend on access to long-term patient data as well as on better models, and that BioStack could become an important data source for medical AI.

What the money is for

BioStack said it would use the funding to expand in the US and abroad, develop its technology platform and grow its engineering and research teams.

The company is entering a busy area. Dealroom pointed to Amigo AI's $11m round for training clinical agents and Snorkel AI's $350m raise for AI data infrastructure as related deals.

Company website: getbiostack.com