By Maurizio Massini, sales director, MF Tecno Packaging Systems
There is a scene in a laboratory at the Waltham Petcare Science Institute that tells the story of where pet food is heading better than any statistic. An algorithm analyzes millions of micro-observations on the oral microbiome of cats and dogs, while a few desks away a team of nutritionists designs the geometry of a new dental snack. This is not science fiction: It is how Mars Petcare develops its products today. What until a few years ago was a sector driven by chemistry and extrusion is being transformed into a data-driven industry.
Pet food is living a paradoxical moment. On one hand, it is one of the most resilient global food industries, capable of growing even when household spending contracts. On the other, it is one of the most under pressure: pet owners demand more sophisticated recipes, more sustainable ingredients, cleaner packaging and fresher formats. The producers' response is coming from an unexpected place: artificial intelligence.
For years, AI in pet food was a conference topic. Today it has entered every link in the value chain: from ingredient discovery to packaging line maintenance, from personalized diet design to real-time product quality monitoring. Petfood Forum 2026, the sector's most important global event, placed personalization and AI-enabled automation at the center of its program. No longer an option, but the operational vocabulary of the years ahead.
The new craft of the formulator: Ingredients suggested by a machine
For decades, developing a new dog food recipe meant combining nutritional expertise, cost constraints, palatability and raw material availability. It was a craft disguised as science. Today that craft has a new collaborator.
Darren Logan, vice president of research at the Mars Advanced Research Institute, told FoodNavigator how the multinational is using artificial intelligence to discover entirely new ingredients, starting from the search for bioactive compounds in plants and low-environmental-impact protein sources. This is not about replacing researchers. It is about allowing them to explore a chemical space too vast for a human team to cover at market speed.
The same logic is found in laboratories designing personalized diets. Tara Zedayko, chief scientific officer at Ollie Pets, described at the AFIA Pet Food Conference during IPPE 2026 in Atlanta a system that allows subscription members to upload a simple photograph of their dog. From that image, an algorithm extracts useful indicators to calibrate the ration. This is the endpoint of a journey that began in 2024 with the acquisition of DIG Labs, a company specializing in AI-powered diagnostics applied to pet health. Foodback Loop, the proprietary system developed by Ollie, integrates millions of real-time data points to recommend recipes and portions tailored to each individual dog.
On the scientific front, academic research is moving in parallel. A paper published in February 2026 in the journal Informatics presented a hybrid machine learning and linear programming system to formulate wet dog food recipes calibrated on breed-specific conditions. The model cross-references data on ingredients, veterinary diets and hereditary diseases, and proposes formulas optimized under nutritional constraints. It is a signal of where the sector is heading: no longer generic recipes valid for broad categories, but diets that appear written for the individual animal.
The pet food industry today is in the same phase that pharma was 20 years ago, when computational chemistry began replacing months of laboratory work with simulations run in a few hours.
AI is not only changing how existing recipes are formulated. It is enabling a generation of ingredients that, until a few years ago, did not exist. BioCraft Pet Nutrition, founded in 2016 by biochemist Shannon Falconer, cultivates ingredients from animal cells intended exclusively for pet food. The company uses artificial intelligence to understand cell biology more efficiently and reduce the number of experiments required before achieving a commercially viable result. A strategic choice, described in an interview with Food Technology magazine published in January 2026: less laboratory, more simulation, faster time to market.
From shelf to bowl: personalization as the new standard
For the end consumer, the most visible face of this transformation is personalization. Fresh direct-to-consumer brands such as The Farmer's Dog in the U.S., Butternut Box and Bella & Duke in the United Kingdom, and Pooch & Mutt, part of the Czech group Vafo, have built subscription models in which each dog receives a formulation calibrated to age, weight, activity level and health conditions.
Pooch & Mutt recently acquired BIOME9, a microbiome research company, and opened its own laboratory on the AberInnovation campus in Wales. The at-home Gut Health Test offered to consumers uses science, research and AI to analyze the dog's microbiome and update its diet accordingly.
The movement is global. In China, PETKIT presented at CES 2026 in Las Vegas a connected device ecosystem — water fountains, robotic wet food dispensers, smart litter boxes — all equipped with AI cameras with facial recognition of individual pets. The data collected is not a detail. It reconstructs a continuous health dashboard for each animal, identifying early signs of urinary conditions in cats or behavioral changes in dogs. This is new territory, where the boundary between food and domestic medical device becomes thin.
The global pet tech market, according to forecasts cited by industry analysts, is expected to grow from $19.1 billion in 2026 to $52.9 billion in 2035. For major pet food producers, this means one precise thing: soon the bowl will be an information node, not just a container. And the companies that today build the right partnerships with pet tech will be those that tomorrow can offer truly tailored recipes.
Behind the scenes: AI enters the factory
While the spotlight falls on recipes and premium brands, a less visible but equally profound revolution is underway in factories. Pet food producers are integrating IoT, analytics and machine learning into extrusion, mixing and packaging lines. NorthWind Technical Services is developing AI capabilities within its PlantIQ platform for predictive analytics, golden batch management and cleaning cycle optimization. This is the same trajectory described by PMMI, the global association of packaging machinery manufacturers, in its 2026 report "Building an AI Advantage in Packaging Equipment": knowledge transfer and predictive maintenance are the two technologies with the highest expected impact in the near term.
Predictive maintenance, in particular, is becoming the most mature use case. Recent studies conducted on food packaging lines have documented 50% reductions in conveyor belt failures and drops of up to 45% in unplanned downtime during the first 12 months of adoption. For an industry like pet food, where large formats exceed 30 kilograms and the prolonged heat of extrusion places every component of the lines under mechanical stress, the difference between a predictable stoppage and an unexpected one is measured in tens of thousands of euros per hour.
On this ground, not only global players but also an Italian and European ecosystem of machinery manufacturers that has long served the world's major pet food brands are being measured. Companies like MF Tecno, an Italian manufacturer specializing in complete weighing, packaging, palletizing and end-of-line systems, work today with pet food clients across Europe, North Africa and Latin America, integrating remote assistance, remote diagnostics and preventive maintenance modules onto their systems designed to reduce machine downtime.
The observation coming from the field is consistent with what is seen at industry events: pet food is becoming an industry where line flexibility — meaning the ability to switch rapidly from one format to another and to manage ever-new materials (recyclable mono-materials, laminated paper, bio-based films) — matters as much as product formulation. It is at this point of contact between recipe and plant that the promise of artificial intelligence is truly tested.
The new geography of innovation
Looking at the map, AI innovation in pet food traces a fairly precise quadrilateral. The U.S. remains the epicenter for research, venture capital and direct-to-consumer models. Mars has announced a $1 billion plan on technology and AI, Nestlé Purina has expanded its R&D structure, and Hill's continues to be Colgate-Palmolive's primary growth engine.
The UK leads on fresh raw and biotech (Meatly, Pooch & Mutt). Germany concentrates research on microbial proteins and precision fermentation, with Marsapet having brought MicroBell to market, the first dry dog food featuring FeedKind microbial protein. China has transformed pet tech into an autonomous industry. Brazil and Southeast Asia, particularly Thailand, have become manufacturing hubs where the digitalization of production lines runs alongside the expansion of productive capacity.
Within this framework, Italy plays a specific role. Few major brands, but of global significance (Monge, Farmina, Morando, Agras), alongside an ecosystem of packaging and process technology suppliers that has accompanied the international growth of the main European players. It is a less conspicuous industry, but a decisive one: when a brand announces a new format based on alternative protein, behind it there is almost always a packaging line designed in Emilia, Veneto or Lombardy.
What to expect from 2026 onwards
Analysts' indications converge on three trajectories.
First: AI applied to formulation will move from pilot to standard adoption. Major multinationals are building dedicated internal teams, while specialized software providers (BESTMIX, SWARM Engineering, Format Solutions) are making their platforms accessible to medium-sized companies as well.
Second: Predictive maintenance and the digitalization of production lines will become the main terrain of competitive advantage on cost per packaged kilogram.
Third: High-end personalization, today the prerogative of direct-to-consumer brands, will come to mass retail shelves through hybrid models combining digital profiling, modular product configurations and variable-data printed packaging.
What is less debatable is the underlying direction. Pet food in 2026 no longer resembles pet food from 10 years ago. Recipes are written by cross-referencing microbiome data and generative models, ingredients can come from bioreactors rather than slaughterhouses, packaging lines tell the maintenance engineer what to change before something breaks, and the pet owner receives at home a pouch packaged specifically for their animal. This is the point at which a century-old industry becomes something else — probably the greatest transformation the sector has ever experienced.
Mars Petcare, Nestlé Purina, Hill's: The giants are at work, but the revolution is not theirs alone. It is made of biotech startups in London and Berlin, university laboratories in Canada and Lithuania, subscription operators in New York and Singapore. Pet food is learning to think. It is up to those who produce it to decide whether to follow or lead the change.
Maurizio Massini is sales director of MF Tecno Packaging Systems and a member of the Massini family, founders of the Massini Industries Group. He joined the family business in 2002 and has since contributed to the international growth of MF Tecno, specialized in weighing, bagging, palletizing and end-of-line packaging technologies, and of MIAL, focused on bulk material conveying and complete production plants for pet food and animal feed. Based in Perugia, Italy, he leads MF Tecno's commercial strategy across the pet food, animal feed and food sectors, with installations in more than 70 countries worldwide.
