A 28-Year-Old Tōji, AI, and a New Shape for Sake Brewing
Visit a sake brewery in Japan, and you may see the tōji cup the air of the kōji room in their palm, reading temperature and humidity through the skin. “Brewing with the five senses” is a beautiful phrase. But when the person who possesses those senses is nearing their seventies, with no successor in sight, a reality emerges that beauty alone cannot resolve. As of 2026, Japan has roughly 1,100 sake breweries, representing less than half the number seen at the industry’s peak in the 1970s. Many closures are driven by a lack of successors and labor shortages. When a brewery shuts its doors, the rules of thumb stored in the tōji’s mind, such as “At this air temperature, stir the moromi 30 minutes earlier” or “When the aroma of the kōji changes like this, lower the temperature by 0.5°C,” are often lost without ever being passed on. Now, quietly but unmistakably, a movement is spreading to preserve that “vanishing intuition” as data and hand it to the next generation.
A 28-Year-Old Tōji Listening to Fermentation Through IoT Sensors
There is a brand called SAKENOVA. Its CEO and tōji, Daichi Niiyama, is 28 years old, a remarkably young age in the sake world. But his age is not the only reason he is drawing attention. He has installed IoT sensors in brewing tanks to collect real-time data on temperature, sugar content, and pH. That data is then used to train AI systems to optimize the fermentation process. The result: labor costs have been cut by about 40%, while the sake has won gold at an international competition. Any initial suspicion that efficiency might compromise flavor has been quietly addressed by the objective fact of that gold medal. Efficiency and quality are not necessarily a trade-off. Here is one example proving that point.
The Kyoto Experiment: AI Proposes, the Tōji Executes
An experiment conducted jointly by Yokogawa Electric and Kyoto craft brewery Craft Bank also produced intriguing results. AI analyzed brewing data and generated an optimal temperature schedule. When the tōji followed that schedule, the fermentation period was shortened from 14 days to 10, while quality was maintained, marking a reduction of about 28%. What matters here is that AI did not “brew in place of the tōji.” It merely proposed the temperature schedule; the final judgment and execution remained in human hands. The experiment reveals a clear relationship: AI is an excellent assistant, not the conductor.
Dassai and Fujitsu: A Quiet Question Dating Back to 1999
Asahi Shuzo, the Yamaguchi-based brewer behind Dassai, has partnered with Fujitsu to develop AI prediction models. The system learns from decades of brewing data, including tank temperature trends, rice polishing ratios, yeast types, and annual climate conditions, providing real-time adjustment suggestions during fermentation. What makes Asahi Shuzo especially interesting is that it had already abolished the traditional tōji system. In 1999, then-president Hiroshi Sakurai steered the brewery toward data-managed brewing that did not depend on a master brewer. At the time, the decision was seen as heretical. Yet Dassai went on to become a global brand. The introduction of AI can be seen as a natural evolution of that same philosophy.
AI Is Not Replacement, but Translation
What these efforts share is not automation, but visualization. Traditionally, temperature control in the kōji room depended on the tōji’s physical senses. Judgments like “It’s a little too humid” or “We should wait another half day” are often correct, but they are difficult to put into words and hard to pass down. IoT sensors and AI are tools for translating these hard-to-verbalize judgments into numbers and history. When a veteran tōji says, “The kōji is in a good mood today,” behind that phrase lies a web of parameters: temperature, humidity, mycelial growth, and the rate of change over time. IoT sensors and AI quantify and analyze those elements, turning them into something the next generation can inherit. Even if the tōji’s words disappear, their meaning remains.
In 2026, Japanese brewing stands at a quiet turning point. The era of brewing with the five senses is not ending; an era in which the five senses and data coexist is beginning. When a glass of sake is poured at a bar overseas, behind that glass lives a layered presence: the accumulated wisdom of fermentation passed down over generations, the voice of fermentation captured by IoT sensors, and the will of young makers translating between the two. (Mr. Bacchus)
This article is intended solely to explore the brewing technology, cultural heritage, and emerging AI-assisted craftsmanship in the Japanese sake industry, and does not aim to promote or encourage the consumption of alcohol. / บทความนี้จัดทำขึ้นเพื่อนำเสนอข้อมูลเกี่ยวกับเทคโนโลยีการผลิต มรดกทางวัฒนธรรม และนวัตกรรมการใช้ AI ในอุตสาหกรรมสาเกญี่ปุ่นเท่านั้น มิได้มีเจตนาเพื่อส่งเสริมหรือโฆษณาเครื่องดื่มแอลกอฮอล์ สำหรับผู้มีอายุ 20 ปีขึ้นไป โปรดดื่มอย่างรับผิดชอบ