Kirin Holdings is significantly expanding the use of artificial intelligence in its beer development process, marking a notable shift in how new beers are designed within one of Japan’s most established brewing groups. According to IntelliNews, the company plans to introduce AI-supported beers to the market from March 2026, using data-driven models to shorten development cycles and improve flavour precision.
The AI system analyses multiple taste-related variables such as bitterness, body and richness, translating desired sensory profiles into specific chemical and ingredient combinations. Drawing on around 20 years of internal data covering ingredients and consumer tasting results, the technology allows Kirin’s brewers to quantify how individual components influence flavour outcomes and to adjust recipes more efficiently when issues such as excessive bitterness occur.
Kirin argues that traditional beer development relies heavily on the intuition and experience of individual developers, which can slow innovation and lead to inconsistent results. By combining accumulated brewing know-how with machine-led analysis, the group expects faster iteration and more predictable outcomes. Consumer tests cited by the company showed that beers developed with AI support were rated more positively than those created through conventional methods.
While Kirin is also applying advanced AI technologies beyond beer, including collaborative research with Fujitsu on food functionality and health-related compounds, the brewer stresses that in beer production the role of AI is strictly supportive. Final decisions on flavour direction, brand positioning and market launch will remain with human brewers, ensuring that craftsmanship continues to define the character of Kirin’s beers.
As an early example of artificial intelligence being applied to beer recipes, experiments with generative AI were already underway several years earlier. In 2023, ChatGPT was used in the United States to generate complete beer recipes that were subsequently brewed and sold commercially, including guidance on malts, hops and flavour balance. One such project demonstrated that “ChatGPT created a recipe good enough to brew and sell,” illustrating that AI-driven recipe design was already moving from theory into practical brewing applications at an early stage (inside.beer, 21.03.2023).
