BioStack Nets $9.5 Mn To Build Training Datasets For Medical AI

The startup plans to utilise the fresh capital to continue expanding its presence in the US and other international markets, scale its tech stack and ramp up hiring

The fundraise comes as India is becoming a training ground for collecting datasets for AI models, driven by huge local population and market opportunity for selling the data

Founded in 2025, BioStack uses fragmented clinical information across systems to create proprietary datasets for research, biotechnology, and the development of AI models

AI-powered healthtech platform BioStack has raised $9.5 Mn (₹90.4 Cr) in its seed funding round led by Peak XV Partners, Afore Capital, DCVC, Emerson Collective, SV Angel, Verdict Capital, Streamlined Ventures, and Y Combinator. 

The startup plans to utilise the fresh capital to continue expanding its presence in the US and other international markets, including India, Brazil, and Argentina. A chunk of the capital will also be used to scale its tech stack and ramp up hiring across engineering and research teams.

Founded in 2025 by Sanat Mishra and Parth Patwa, BioStack claims to be building the data infrastructure for the next generation of medical AI. 

It combines fragmented clinical information across systems such as electronic health records, medical imaging, clinical notes and medications. The startup then uses this data to create proprietary datasets for research, biotechnology, and the development and training of medical AI models.

“We’re building the infrastructure that connects all of the clinical information and makes it useful, both for clinicians making decisions today and for the researchers and AI developers building the future of medicine,” said BioStack cofounder and CEO Mishra.

BioStack currently works with multiple hospitals across the US, Brazil and Argentina. In India, it is also connecting cancer patients with an investigational cancer vaccine program. 

The startup also claims that it offers use cases such as AI model evaluation, disease modeling, biomarker discovery, patient stratification and treatment-response analysis.

Besides offering custom workflows to protect patient privacy, the startup claims to have so far processed 10 Cr de-identified patient records, more than “several hundred thousand” fragmented X-Rays, MRIs and CT scans, and “several million” telehealth audio transcripts. 

For context, de-identified patient records are medical datasets that have had all personal and direct identifiers removed so that the individual cannot reasonably be linked to the data.

The fundraise comes as India becomes a training ground to collect datasets for powering AI models and physical AI products. Startups are tapping India’s huge population to collect recordings of workers performing physical tasks, processing the footage, and selling the resulting datasets to frontier AI labs and robotics companies.

For instance, the likes of Awign, Objectways and Human Archive are building pipelines to supply first-person footage of human activity to AI developers. Just a day ago, instant home services startup Pronto’s founder Anjali Sardana launched Provision Labs, a US-based venture to collect data for physical AI systems.

At the heart of all this is the Indian AI training dataset, which is projected to become a $1.3 Bn opportunity by 2033.

Source: Inc42