€30 million. That is what mbiomics, a Munich-based techbio company, raised across three tranches of a Series A that closed last month. The money will fund a pharmaceutical-grade live bacterial product designed through machine learning — something the biotech industry has promised for a decade but has not yet delivered at scale.
The company's platform builds defined microbial consortia: specific combinations of bacterial strains selected and optimised through AI/ML models, proprietary analytics, and high-throughput co-cultivation screening. The output is a live biotherapeutic product (LBP): an oral capsule containing a precisely engineered mix of bacteria, manufactured to GMP standards, that can be prescribed like any other drug.
Fecal microbiota transplants (FMTs) proved the concept. Modulating the gut microbiome can change disease outcomes. But FMT is a black box. You cannot scale a stool transplant. You cannot patent it. You cannot control its composition batch to batch. That is the gap mbiomics and a handful of peers are trying to close.
Lead candidate MBX-116 targets advanced melanoma as a co-therapy with immune checkpoint inhibitors, where the gut microbiome has been shown to influence response rates. A Phase 1b study is planned for 2027. Beyond oncology, the pipeline extends into autoimmune and neurodegenerative conditions, all connected through the gut-immune axis.
The field is moving. The human microbiome therapeutics market now exceeds $1.6 billion. MRM Health raised €55 million last year for its own rationally-designed consortia platform. Pendulum Therapeutics brought an AI-designed probiotic for type 2 diabetes to market. Seres Therapeutics received FDA approval for VOWST, the first defined bacterial consortium product, and has Breakthrough Therapy designation for its ICU candidate SER-155.
What makes mbiomics different is the technology stack. Most microbiome companies pick strains based on correlation: this bacterium is depleted in sick patients, so replace it. Mbiomics builds from mechanism: AI models predict which combination of strains, at what ratios, will produce a targeted immunological or metabolic effect. The platform handles the complexity of maintaining a multi-strain consortium through manufacturing, which has been the graveyard of earlier attempts.
The investors, MIG Capital and Bayern Kapital, are backing this with the understanding that the translation risk has shifted from biology to engineering. The biology is established. The question is whether you can manufacture a living therapeutic at the scale, consistency, and cost that makes it a real alternative to small molecules and biologics.
What to watch. IND submission timing for MBX-116: any delay extends cash runway risk for a pre-revenue company. Partnership deals: a licensing agreement with a pharma partner holding checkpoint inhibitor assets would validate the platform and provide non-dilutive capital. As we wrote in July, the intersection of AI and microbiome science is drawing serious institutional attention. The Phase 1b safety readout in 2027 will be the first real test of whether AI-designed consortia work in humans, not just in silico.
