Pixeom
Category: AI in Biotech / Drug Discovery
Oita University spinout developing PSM (Proximity Score Matrix), a patented 3D molecular encoding technology that lets AI "see" molecular structures directly for drug discovery and chemistry research. Pixeom was founded in 2025. The company is led by Shigeru Matsuoka (松岡茂). Based in Oita City, Oita Prefecture, Japan. Team size: 1-10. Latest round: Seed. Key investors include Lifetime Ventures.
- Founded
- 2025
- Headquarters
- Oita City, Oita Prefecture, Japan
- Team size
- 1-10
Value proposition
Converts 3D molecular structures into high-information numerical representations (PSM) that AI can learn from directly, avoiding the information loss and hallucination caused by converting molecules to strings/symbols. Aims to establish "chemical vision AI" as a new standard for AI-driven drug discovery.
Products and solutions
PSM (Proximity Score Matrix) molecular representation technology, PSM Explorer structural analysis workspace with similar-structure search, roadmap toward property/activity prediction, de novo molecular design, and a drug discovery platform (SaaS targeted for 2028).
Unique value
Patented (Japan Patent No. 7715333) 3D molecular encoding that lets AI directly "see" molecular geometry rather than infer it from symbolic representations — addressing the root cause of hallucination in AI drug discovery rather than just improving model performance.
Target customer
Universities, research institutions, pharmaceutical companies, and drug discovery startups in the drug discovery and chemical research fields.
Industries served
Pharmaceuticals / drug discovery, chemistry research, life sciences
Technology advantage
PSM encodes geometric information (distance, element, stereochemistry) of molecules into numerical matrices, enabling structure recognition, similar-structure search, and property/activity prediction. Focuses on the input representation layer ("how AI sees molecules") rather than model architecture, positioning it as a foundational standard for chemical vision AI.
How they differentiate
While most AI drug discovery companies compete on model performance, Pixeom targets the layer before the model — how molecules are represented/encoded for AI. Its PSM approach preserves 3D geometric structure information that is lost in conventional string/symbol-based representations, aiming to reduce hallucination and enable direct structural understanding.
Main competitors
Insilico Medicine, Schrödinger, Elix, ExaWizards
Key partnerships
Oita University (spinout origin, located in Oita University Research Management Organization building), Lifetime Ventures (investor)
Major milestones
Founded February 2025, project start 2025 Q4, internal PoC design/initial validation 2026 Q1, seed funding from Lifetime Ventures September 2026, prototype verification for joint research targeted 2026 Q3, analysis support service consideration 2027 Q1, SaaS expansion targeted 2028 Q1.
Market positioning
Early-stage university spinout positioning PSM as a potential standard input format for chemical vision AI, targeting the drug discovery and chemistry research market in Japan and internationally.
Geographic focus
Japan (headquartered in Oita), with stated intent to expand collaborations with pharmaceutical and life sciences sectors both domestically and internationally.
Patents and IP
Japan Patent No. 7715333 — "立体構造同定システム、方法、プログラム" (3D structure identification system, method, program), registered.
About Shigeru Matsuoka (松岡茂)
Specially Appointed Professor, Oita University Faculty of Medicine; former Specially Appointed Associate Professor at Oita University and University of Tokyo Graduate School of Medicine; CTO of Albius Sciences Ltd. University of Tokyo graduate (BSc/MSc Chemistry); PhD in Science (Osaka University). Specialist in molecular structural chemistry applied to drug discovery.
Latest news about Pixeom
More AI in Biotech / Drug Discovery companies
Official website: https://pixeom.jp/