Health AILongevity Science
Avinasi Labs
Accelerating a longevity research program by solving the data access problem first.
The Challenge
The health companies that will lead in AI are being decided right now. The models being built today will define competitive position for the next decade. The single biggest variable in whether those models are any good is the data they were trained on.
Longevity research makes that tension visible. The models that matter in this space need real-world, longitudinal health data: sleep, biomarkers, lifestyle behaviours, energy patterns, tracked across the same individuals over time. That data exists. But accessing it through conventional channels means navigating institutional data sharing agreements, ethics board timelines, and regulatory frameworks that were not designed for the speed at which AI development moves.
Avinasi Labs had a clear research agenda and the technical capability to execute on it. What they could not afford was to wait.
The Solution
Avinasi partnered with OpenDataLabs to source Primary Source Datasets: real-world health and wellness data contributed directly by users who chose to participate. Sleep data, biomarker logs, energy and recovery patterns, lifestyle signals, linked at the individual level across multiple domains.
Because contributors consent directly, the data is available without institutional gatekeeping. No data sharing agreements to negotiate. No ethics board queues. Every record traces back to an explicit consent event, and the provenance is documented end to end.
That matters for speed. It also means the data reflects what real people actually chose to share, which is a different quality of signal than data extracted from institutional pipelines never designed with AI training in mind.
The Result
Avinasi now trains its longevity foundation model on multi-domain health data sourced directly from real contributors. The data captures the individual-level variation and behavioural signals that institutional datasets consistently lack.
Their research program moved faster than it would have through conventional channels, built on data with full visibility into where it came from and how it was obtained.
Why It Matters
Every health company with an AI research agenda is racing against the same clock. The models being built right now will define which companies lead this space for the next decade. The single biggest variable in whether those models are good is the data they were trained on.
Institutional data acquisition was not built for this moment. The timelines are too long, the datasets too narrow, and the gap between what those pipelines produce and what serious AI development requires keeps widening.
Primary Source Datasets gives health AI teams a direct route to real contributor data. No gatekeeping. Full visibility into consent and provenance. Multi-domain signals linked at the individual level, from people who actively chose to participate.
The companies that lead in health AI will be the ones that moved fast and built on data they could actually stand behind. That is the decision in front of you right now.
Data is the lifeblood of longevity research. Working with OpenDataLabs gives us access to diverse, high-quality datasets while honoring the sovereignty of contributors.
Winnie Qiu, Co-Founder, Avinasi Labs