By Federico Mentoni and Jack Fox-Male Reading time: 10 minutes
The poke bowl you picked up on the run yesterday afternoon.
The protein-rich pill you’ll be having for dinner in 10 years.
Maybe even the vegan burger you tried for the first time.
All of these have been touched by AI at some point in their journey from raw data to your plate.
AI has become foundational in revolutionising food tech in the past decade. While early use cases focused on logistics and supply chain management through early adopters such as GreenTropism or SeaFoodIQ, more and more applications are coming.
But…
Why Should You Care About AI in Food Tech?
AI food tech applications address complex challenges by streamlining processes across the value chain. AI is reshaping how food is created, processed and marketed. From generative models accelerating the R&D of novel products, to machine vision systems enhancing food safety and predictive software to match consumer trends.
The global AI for food technologies market is currently sitting at $11.6 billion, and is estimated to reach $84.8 billion in 2030, growing at an impressive CAGR of 39.1%. This upward trajectory is reflected in significant investor interest, with 2024 and 2025 seeing an increasing number of funding rounds in early-stage startups, such as Starday’s $15 million Series A or Biomatter’s $6.5 million Seed Round.
New Opportunities Emerging in Food Tech
The Hello Tomorrow Challenge has a dedicated Food & Agriculture track and is the world’s longest-running deep tech startup competition. Now in its 11th edition, it receives more than 5000 applications ever year.
The Hello Tomorrow Challenge is the world’s longest-running startup competition, receiving 5000 applications every year. Analysing the biggest trends from the 2025 edition, we identified two transformative trends shaping the future of AI in food tech throughout 2025:
– Applied generative AI to reduce R&D time of novel ingredients
– Large language and processing models to grasp consumer insights
Generative AI is cutting the R&D processes of product formulation and ingredient discovery from years to weeks. AI-driven tools analyse molecular structures and biological interactions, such as protein folding and stability, or functional property optimisation for microbial and synthetic proteins. This approach allows companies to design novel protein sources and alternative ingredients with sensibly improved efficiency.
Companies like NotCo have pioneered this space. Their generative AI–based platform offers a holistic solution: they improve nutritional profiles of existing products by offering alternative compositions, and they de-risk R&D by AI-enabled data retrieval across disparate sources.
Similarly, other emerging startups are adopting approaches targeted at specific phases of the R&D value chain. AI Bobby, for instance, leverages neural AI to drastically reduce the time required to find the correct pattern for gelation processes in alternative protein development, mirroring the success seen in healthcare for drug formulation.
Tastewise, founded in 2017, uses AI to monitor real-time market trends, extract insights from dynamic databases, and forecast consumer preferences with unparalleled accuracy.
Large language models (LLMs) and natural language processing (NLP) technologies are transforming how food companies understand and predict consumer behaviour. Ventures like Tastewise and Spoonshot trained AI to monitor real-time market trends, extract insights from dynamic databases with a constant flow of data intake, and forecast consumer preferences with unparalleled accuracy.
These tools not only generate actionable marketing insights but also assist in conceptualising new product ideas tailored to specific demographic or cultural preferences. By improving flavour profiling accuracy by an average of 20% and shortening recipe formulation processes by 25%, food-specific LLMs allow brands to bring personalised, high-quality products to market faster and more efficiently.
However, despite new exciting technologies and a rapid increase in market size, unlocking this opportunity requires addressing a handful of fundamental challenges that hinder food companies from fully incorporating AI in their processes.
The Evergreen Challenges: Overcoming Data Silos and Regulatory Hurdles
Two specific challenges are yet to be solved: data silos and regulatory hurdles.
Data silos
Firstly, there is a distinct lack of accessible, high-quality datasets necessary to train AI models. Unlike sectors such as pharmaceuticals, where standardised and structured datasets are more readily available, food tech operates within a fragmented data ecosystem. Factors like regional variability in ingredients, environmental conditions, and inconsistent processing methods leave us with datasets that are extremely difficult to harmonise.
Large corporations and research institutions hold the majority of available datasets, often viewing them as a potential competitive advantage and limiting access for startups and external entities, even when they are not actively engaging in any activity related to the data. For instance, universities hold most of the global data on food shelf life, and have shown limited willingness to share it with external entities.
This is a problem, because without widespread data-sharing frameworks, the scalability and accuracy of AI applications will remain constrained, hindering their scaling potential.
Regulatory hurdles
And then we have regulatory complexity. This particularly concerns novel ingredients and alternative proteins.
In the EU especially, stringent compliance requirements and bills introducing restrictions on alternative protein R&D have deterred investment and innovation. Slow approval processes for novel ingredients and alternative proteins, which now require an average of 28 months from the first dossier submission, are contributing to the ecosystem’s general uncertainty by greatly limiting startups’ ability to iterate and scale their products, a critical requirement for securing larger funding rounds.
Because of this, many founders relocate to more favourable ecosystems, such as Singapore and the U.S., where regulatory frameworks are more conducive to experimentation and growth.
The solution? Build a Data Consortium and Streamline Approval Paths
The goal is industry-wide adoption of AI in food tech. And to solve the two challenges mentioned above, we can draw inspiration from neighbouring verticals: pharmaceuticals and fintech.
Data consortiums in the pharmaceutical industry
Data consortiums offer a promising path forward to tackle data fragmentation. These consortiums should involve collaboration between private companies, academic institutions and the public sector to pool, standardise, and share resources.
By centralising diverse datasets, such as flavour profiles or ingredients rheology, consortiums will be able to create a unified foundation for training AI models. This will not only enhance data quality and accessibility, but also reduce redundancy in data collection efforts. To ensure trust and fairness, such consortiums can adopt data-sharing agreements that protect proprietary information while promoting collective growth.
In the pharmaceutical industry, shared databases have already accelerated drug discovery and development, leading to the growth of companies like Insilico Medicine.
In 2024, the UK announced a regulatory sandbox for cultivated meat, while the US announced an AI sandbox to evaluate the feasibility of AI-driven projects before scaling them.
Streamlining approval paths in fintech
When it comes to regulatory frameworks, one solution lies in the development of simplified pathways to streamline approval processes, without necessarily compromising safety standards.
Singapore is already moving in this direction. Initially a fintech-heavy ecosystem, it pivoted towards food tech by creating an adaptive regulatory environment that encouraged innovation. Entities such as the Singapore Food Agency (SFA) and A*STAR have played a pivotal role in this transition, introducing practical guidelines for the approval of alternative proteins and cultivated meat, and the creation of a regulatory sandbox.
The UK is following the same path, through the sandbox for cultivated meat, which will include £1.6 million dedicated to funding startups operating cell-cultivated products. The U.S. is also active on this front, launching an AI sandbox in April 2024 in collaboration with the World Food Programme (WFP) to evaluate the feasibility of AI-driven projects before scaling them, ensuring ethical and impactful solutions.
The Future is on Your Plate
AI is increasingly becoming a cornerstone of innovation in food technology, reimagining everything from product development to supply chain optimisation.
To unlock the full benefits of AI, stakeholders across the food industry must collaborate. Initiatives such as data consortiums and adaptive regulatory frameworks can provide the infrastructure necessary for safe, responsible, and scalable AI integration.
Creating open, collaborative environments and forward-thinking policies will ensure that AI becomes a driver of meaningful progress in food technology, helping to pave the way for a future where food innovation is smarter, more inclusive and better equipped to meet global needs.
At Hello Tomorrow Consulting, we partner with corporations to turn deep tech into real business opportunities, supporting with new market entry, foresight & trends analysis, startup sourcing, corporate venturing strategy, and more.
Interested in working with us? Get in touch.
Or, if you’re a founder of a deep tech startup, apply to the world’s longest-running startup competition, the Hello Tomorrow Challenge.
