Core Strategic Thesis
The fundamental insight driving this venture is that biology's rate-limiting step has always been experimental data generation — not ideas. Autonomous labs eliminate this bottleneck by closing the Design-Build-Test-Learn (DBTL) loop at machine speed. But building the robots themselves is not the moat. The moat is in three things: the experimental graph (the typed, provenance-signed record of how every datapoint was produced), proprietary clinical data (Apollo Hospitals biobank), and an agent mesh that turns natural-language hypotheses into auditable wet-lab execution across heterogeneous partners.
Our thesis in one sentence
LabOS is the agentic operating system for distributed biology — a multi-agent mesh that orchestrates multiomics computation, partner wet labs, and Apollo Hospitals' clinical biobank into closed-loop discovery. We do not own robots; we own the experimental graph and the India-tuned models running on top of it.
The Three-Layer Stack — LabOS at the Center
Layer 1 · OmicsOS (the AI brain feeding LabOS)
Multi-modal model integrating WGS/WES, scRNA-seq, spatial omics, proteomics, metabolomics, epigenomics. Generates hypotheses and ranks targets. Frontier-model wrappers in Year 1; fine-tuned, India-cohort-trained bio model on sovereign cloud by Year 2 (see LabOS fine-tuning roadmap).
Layer 2 · LabOS — the Agentic Operating System FLAGSHIP
A five-agent mesh (Hypothesis → Design → Execution → QC/Validator → Learning) that translates a goal into a typed experiment graph, routes nodes to the cheapest qualified partner (Syngene, Aragen, MedGenome, Ginkgo Cloud Lab, FAL Pods), captures provenance-signed results, and feeds the loop. LIMS-native: ingests from and writes to STARLIMS, LabWare, Benchling, LabVantage. No robots owned. The experimental graph is the moat.
Layer 3 · ClinicalBridge (regulatory translation)
Regulatory-grade packaging for CDSCO (India), TGA (AUS), PMDA (Japan), FDA/EMA bridging studies. Biomarker panels, IVD companion-diagnostic dossiers, Phase 4 pharmacovigilance signal packages. India-specific disease area specialization.
Data Substrate · Apollo Hospitals Biobank
Anchor data partnership with Apollo — pan-India EHR, longitudinal clinical phenotyping, consented biospecimens. The India Biomarker Atlas is constructed on top of this substrate. See Apollo Partnership.
Business model
| Revenue line | Model | Typical ACV |
|---|---|---|
| LabOS Platform | SaaS per site + broker take-rate | $250K–1.5M/yr |
| Discovery Sprints | Milestone-gated programs | $0.5M–8M |
| Atlas licensing | Annual data license | $200K–2M/yr |
| PV surveillance | Recurring contract | $2.5M/yr |
| Co-development | Equity + royalty | Long-tail |
What we do not do (Phase 1)
- Own robotic labs — partner CROs and cloud labs; see Build vs Partner
- Sell diagnostics kits directly — we develop CDx with pharma partners; manufacturing via MiRXES/BioHelix at scale
- Replace Apollo clinical care — we are a research and discovery layer on consented data