Outpost Bio has raised $3.5m in a pre-seed round to develop computational models of human microbiology, aiming to help pharma and consumer health companies better predict how microbial communities influence interventions.

The round was co-led by Merantix Capital and Seedcamp, with participation from OpenSeed VC, Defined, and family offices and angel investors.

Human microbiology examines how microbial communities living in and on the body metabolise drugs, transform nutrients and shape health outcomes. While microbiome research has expanded rapidly, translating complex datasets into actionable insights remains a challenge for R&D teams.

Outpost Bio is developing what it describes as “frontier models” focused on microbial interaction layers rather than solely on human DNA. The company argues that many biological foundation models overlook how microbial communities modify drugs and other interventions.

Central to its approach is a “Lab-in-the-Loop” platform that integrates automated experimentation with machine learning. The system creates a feedback loop in which models learn from experimental outputs and guide subsequent testing. According to the company, this enables the generation of proprietary, human-derived functional datasets at scale.

Outpost Bio said the platform is designed to help pharma partners predict how microbial communities may alter drug metabolism, support formulation design and generate evidence for regulatory submissions. The technology is also positioned for use in food and consumer health, where companies are seeking to understand how prebiotics, botanicals and other ingredients reshape microbial ecosystems.

Jenny Yang, co-founder and chief executive officer at Outpost Bio, said: “We’re building the most comprehensive dataset in human microbiology. Microbial communities can dramatically alter drugs and other interventions, yet this layer has been largely ignored because the data hasn’t existed at scale. For the first time, we can move beyond correlations to reveal causal pathways.”

The founding team combines backgrounds in microbiology, data engineering and machine learning, with prior roles at organisations including Amazon Web Services and Basecamp Research. Advisers include former leaders from Ginkgo Bioworks, the Genome Sciences Centre and the DARPA Biological Technologies Office.

Tom Wilson, partner at Seedcamp, said: “This is a rare team that combines deep microbiology, machine learning, and company-building experience. Jenny, Alex, and the founding team have both the scientific rigor and operational insight required to build a new layer of biological infrastructure.”

Adrian Locher, co-founder and general partner at Merantix Capital, added: “Faster wet-lab data generation, lower sequencing costs, and more powerful machine learning make this the right moment to build predictive models of the microbes that live within and on us.”

The company said the new capital will be used to scale its experimental and modelling platforms, with a focus on expanding its proprietary datasets and advancing predictive capabilities for drug and ingredient development.

As microbiome research matures, investors are increasingly backing platforms that aim to move beyond descriptive correlations toward mechanistic and predictive models. Outpost Bio’s strategy reflects a broader push to integrate wet-lab automation with AI systems in order to generate datasets tailored to specific commercial and clinical questions.