Last updated: 16-07-2026

AI in Hybrid Food Development: Inside Red Rabbit and Crespel & Deiters' Formulation System

Red Rabbit and Crespel & Deiters debuted an AI-linked product development system at Food ingredients Europe 2025, connecting Red Rabbit's nextfood.ai platform to Crespel & Deiters' Functional Ingredient Navigator (FIN) database, compressing early-stage concept and recipe work from several weeks to a few days, according to nextfood.ai co-founder Jochen Matzer.

Both companies have confirmed speakers on the Hybrid Foods Europe programme, Jochen Matzer and Christopher Busch, Head of Research & Development at Crespel & Deiters, which makes this a live, current example of AI-driven formulation work already reaching the same ingredient suppliers and category teams the September conference is built for.

What the System Actually Does

Nextfood.ai works through a sequential workflow: a large trend and insight database identifies emerging market opportunities, a product generator produces draft concepts complete with design and formulation suggestions, an assessment module tests those concepts against synthetic consumer groups to flag weak ideas early, and a real-lab step validates the surviving concepts under production conditions. Layered onto Crespel & Deiters' FIN database, which maps how specific ingredients affect structure, stability, and processing for its Loryma ingredient brand, the combination lets developers move from a market trend to a testable formulation in a single continuous sequence rather than a series of disconnected R&D handoffs.

Why This Matters for Hybrid Specifically

Hybrid meat and dairy formulation is unusually iteration-heavy, since every ingredient substitution changes taste, texture, and cost simultaneously, and Matzer has argued that companies that cannot keep pace with AI-supported development will fall behind on cycle time alone. Compressing that iteration loop from weeks to days matters more for hybrid than for single-ingredient plant-based products, because hybrid formulation is fundamentally a multi-variable optimisation problem: animal-to-plant ratio, ingredient choice, and inclusion level all move together. This is the same multi-variable challenge addressed in the wider B2B hybrid food and ingredients opportunity session that Christopher Busch presents alongside Andre Limmer at Hybrid Foods Europe.

Synthetic Consumer Testing: Promise and Limits

The assessment module's use of synthetic consumer groups to pre-screen concepts before physical testing is the most commercially significant, and most unproven, part of the system: it can catch obviously weak concepts early and cheaply, but it cannot yet replace the sensory and functional testing that determines whether a hybrid product actually performs on shelf. Real-world validation, cooking performance, shelf life, colour stability, and the animal-to-plant protein ratio consumers actually notice, still requires the kind of hands-on tasting work built into Hybrid Foods Europe's Innovation Plaza sessions with PlanetDairy, Farm Dairy, Smaqo, and Innovate.NU. AI-accelerated concept generation and physical sensory validation are complementary steps, not substitutes for each other.

Where This Fits in Europe's Broader AI-for-Food Agenda

Nextfood.ai and FIN are a commercial, company-level example of a trend the European plant-based innovation community has been tracking at policy level: a proposed €50 million EU investment through 2035 in AI-driven food innovation infrastructure, provisionally named AlphaFood, aims to build an open-access predictive platform for ingredient functionality modelled on the AlphaFold protein-structure database. Where AlphaFood is a proposed public research infrastructure, nextfood.ai and FIN show the same underlying idea, AI-accelerated mapping between ingredient properties and product outcomes, already operating as a commercial service between two named companies. The gap between the two is instructive: public AI infrastructure for food is still a funding proposal, while private AI tooling for formulation is already live in the market.

Comparison: Public AI Infrastructure vs. Private AI Formulation Tools

Take-Home Messages

Commercial

• AI-accelerated concept development is already commercially available, not a future capability; category teams should be evaluating it now, not waiting for it to mature.

• Cycle-time compression matters most for hybrid formulation specifically, given its multi-variable optimisation problem.

• Synthetic consumer testing can cut early-stage failure cost but should not replace physical sensory validation before launch.

• The gap between proposed public AI infrastructure (AlphaFood) and live commercial tools (nextfood.ai) shows private industry is currently moving faster than EU-level research infrastructure.

  
  

Technical

• The FIN database maps ingredient-to-processing relationships specifically for Crespel & Deiters' Loryma ingredient range, meaning outputs are only as broad as the underlying ingredient data.

• A real-lab validation step is built into the nextfood.ai workflow, addressing the most common criticism of AI-generated food concepts: that they don't survive contact with actual production.

• The system's assessment module benchmarks concepts against synthetic consumer personas, a technique still unproven against real purchase behaviour at scale.

• Hybrid-specific formulation challenges, taste, texture, and cost moving together, make this class of tool more valuable for hybrid than for single-ingredient reformulation projects.

Verdict & Next Step

FoodConNext Foundation's own Global Research Ecosystem has tracked AI-driven formulation tools moving from lab curiosity to commercial deployment inside eighteen months, and the Red Rabbit-Crespel & Deiters system is one of the clearest examples yet in the hybrid space specifically. Both companies bring this work directly to the retail and food service and product innovation days at Hybrid Foods Europe, Van der Valk Zuidas, Amsterdam, 14-16 September 2026. Innovation Plaza slots, where AI-generated concepts meet physical sensory testing, are filling ahead of September. Register now to see where formulation speed and sensory validation actually meet.

About Gerard Klein Essink

Gerard Klein Essink is Founder & CEO of FoodConNext Foundation. He has run an international plant-based foods and proteins community for more than 20 years, published numerous industry reports, and written innovation reports on proteins for the Dutch government. He has advised the Canadian government on its pulse strategy, produced strategic outlook reports for Pulse Canada and the Australian Grains Research & Development Corporation, and leads FoodConNext Foundation's 100+ member Global Research Ecosystem. He has led multiple Horizon EU protein-transition projects, including Metaphor, Prominent, Plenitude, Profuture, HealthFerm, and Giant Leaps, and spoke at the European Parliament and European Commission in June 2026.

About FoodConNext Foundation

At FoodConNext Foundation, we believe that the future of food lies at the intersection of innovation, sustainability, and global collaboration. Our foundation is dedicated to accelerating the transition toward more resilient and responsible food systems by connecting key stakeholders across the agri-food ecosystem.

Our Mission

FoodConNext Foundation exists to bridge gaps in the global food system — bringing together entrepreneurs, researchers, policymakers, and investors to co-create solutions that address some of the world's most pressing challenges, including food security, sustainability, and nutrition.