3Billion moves from rare-disease diagnostics into therapeutic development under a national research grant.
The AMW Read
Modest ($6M) government-grant-backed pivot from AI rare-disease diagnostics into N-of-many ASO drug development is an incremental, single-company expansion, not a segment-wide shift.
3Billion moves from rare-disease diagnostics into therapeutic development under a national research grant.
3Billion (쓰리빌리언), the Seoul-based AI genetic-diagnostics company for rare diseases led by CEO Chang-Won Kum, has been selected as a joint research institution under Korea's ARPA-H-style "ARISE" program. The project, titled "Establishment and Demonstration of a Preemptive Development Platform for N-of-Many ASO Therapeutics for Rare Diseases," is led by KAIST with Asan Medical Center as co-participant. It runs from July 2026 through December 2030 and carries up to 8 billion won (~$6M) in government R&D funding, targeting a platform for antisense oligonucleotide (ASO) candidate discovery and preclinical validation.
The project's core bet is a shift from "N-of-1" rare-disease treatment, where each patient requires a separately developed therapy, to "N-of-many," which first identifies cohorts of patients sharing the same or similar genetic mutations across different hospitals, then applies one ASO strategy to the group. 3Billion's contribution is the capability it already built for diagnosis: AI-based genetic-variant interpretation and accumulated genomic and clinical data, now repurposed to define matched patient pools and build the evidentiary case for selecting therapeutic targets — diagnostic AI acting as an upstream sourcing layer for a drug pipeline rather than a standalone service.
Per the AI Market Watch index (founded 2016, $62M total funding tracked; the index covers roughly 5,000 companies, not a census), 3Billion had already reached beyond core diagnostics with an August 2026 newborn-screening contract — this ARPA-H-linked grant is a second adjacent-revenue expansion in as many months, though the sum involved (~$6M over 4.5 years) signals early platform-building, not a funded drug pipeline. For builders, it's a template: a genomic-diagnostic dataset and variant-interpretation model carry real reuse value in therapeutic candidate discovery. For investors, non-dilutive multi-year government co-development grants are a differentiated, lower-risk funding channel for rare-disease AI companies relative to pure equity rounds — worth tracking as a repeatable pattern.
