IN Brief:
- Mars and Unreasonable Group have selected 13 growth-stage businesses for the third Unreasonable Food cohort.
- Technologies include fermentation proteins, functional ingredients, rapid contamination detection, data platforms, and plastics recycling.
- The companies will work with Mars teams, mentors, investors, and commercial contacts as they pursue scale-up.
Mars and Unreasonable Group have selected 13 growth-stage businesses for the 2026 Unreasonable Food programme, bringing together technologies spanning fermentation, functional ingredients, food safety, nutrition data, and packaging circularity.
The third annual cohort includes ventures from the United States, UK, Netherlands, France, Portugal, and India. The companies are meeting in Windham, New York, from 12 to 18 September before joining the wider programme network of Mars teams, mentors, and investors.
The selection is broader than a conventional ingredient accelerator. Several participants are developing food inputs, while others focus on contamination detection, sustainability data, clinical evidence, and recycling infrastructure. That range reflects the number of technical and commercial hurdles that sit between a promising food technology and routine use in large-scale production.
Superbrewed Food is working with microbial ingredients intended to combine digestive-health functionality with protein nutrition, while Myota develops prebiotic fibre blends designed around differences in gut microbiomes. Holobiome is mapping interactions between diet, biology, and gut bacteria with the aim of identifying microbes and foods for future consumer and health applications.
BIOVIT is developing vitamins and minerals extracted from plants and fungi, and Kuli Kuli has built a supply chain around moringa. Lasso Labs takes a processing-led route, producing fibrous structures from simple ingredients to create meat-like texture without relying on high-temperature processing.
Protein development is represented by Nutrition from Water, which grows microalgae in fermentation tanks, while Dutch company Revyve is producing yeast-derived ingredients designed to perform functions normally supplied by eggs, including emulsification, gelling, and binding. Those functions are central to products such as sauces, bakery lines, prepared foods, and meat alternatives, where ingredient replacement has to survive mixing, heating, cooling, and storage rather than simply match a nutritional target.
Commercial qualification is often where these technologies slow down. An ingredient that performs in a development kitchen may behave differently under industrial shear, continuous pumping, large-batch heating, or rapid cooling. Food manufacturers also need repeatable specifications, allergen information, microbiological controls, shelf-life data, and reliable supply before a new material can enter a standard recipe.
Spore Bio addresses another production constraint. The French company is developing rapid contamination detection intended to identify problems on factory surfaces within minutes rather than waiting days for conventional microbiological results. Faster results could shorten the gap between environmental sampling and sanitation decisions, particularly in facilities operating tightly scheduled production cycles.
Speed alone is not enough for factory adoption. A rapid system has to demonstrate suitable detection limits, repeatability, and sampling performance, while sanitation teams need clear procedures for interpreting results. False positives can trigger unnecessary cleaning and downtime; false negatives carry the more serious risk of allowing contamination to remain undetected.
Data businesses make up another part of the cohort. HowGood provides ingredient-level sustainability information for brands and retailers, while FoodHealth scores products using nutrient density and ingredient quality. Radicle Science runs large-scale controlled trials on supplements and wellness products, adding an evidence-generation route alongside the physical processing technologies.
Without, the cohort’s circularity and recycling venture, is developing processes for multilayer plastics and mixed textiles. Multilayer packaging remains difficult to recover because structures designed for barrier performance, sealing, and mechanical strength often combine materials that are costly to separate after use.
That challenge is particularly relevant to food packaging. Manufacturers can only reduce material complexity so far before oxygen, moisture, grease, aroma, or light barriers begin to suffer. Recycling systems able to handle more complex structures could widen the range of viable end-of-life routes, although food-contact approval and recycled-material consistency remain separate requirements.
Unreasonable Food was established in 2024 as a collaboration between Mars and Unreasonable Group. The programme targets growth-stage businesses rather than very early concepts, aiming to connect selected companies with corporate teams, mentors, investors, and potential commercial partners.
The first cohort has reported nearly $200 million in additional revenue since joining the programme, alongside reductions in greenhouse-gas emissions and water use. Those figures cover a mixed portfolio and do not establish the performance of any individual technology, but they show that the initiative is intended to continue beyond a short accelerator event.
Corporate programmes of this kind can shorten some routes to industrial validation because startups gain access to people who understand procurement, quality systems, factory constraints, and commercial scale. They do not remove the underlying qualification burden. Food businesses still need to know whether a technology meets regulatory requirements, performs reliably on existing equipment, and can be supplied at the required cost and volume.
The 2026 cohort therefore brings together businesses at different points in that transition. Fermentation proteins and functional ingredients have to prove manufacturing economics; contamination technology has to prove reliability inside hygiene systems; data platforms need methodologies that customers can defend; and recycling businesses must deliver material quality consistently enough for converters and brand owners.
Mars and Unreasonable Group are using the programme to concentrate those problems in one commercial network rather than around a single technology theme. The useful measure of the 2026 cohort will come after the introductions and mentoring sessions, when customer trials, supply agreements, factory validation, and investment decisions show which technologies can move into routine food production.
