Roland Berger describes the current state of Japan’s physical AI market, noting that only 7.5% of companies have deployed the technology in production and that many have not even considered adoption. A lack of clarity on ROI and talent shortages are barriers, while the report points to starting with small-scale production implementations, learning from them and using physical AI as a growth investment.
Roland Berger has published the results of a survey of CxOs in Japan’s manufacturing, logistics and infrastructure services sectors. The survey found that only 7.5% of companies have deployed physical AI in production, while 71.5% have not even considered adoption. The main challenges identified were the lack of a clear ROI outlook, shortages of the talent and know-how needed to drive implementation, and uncertainty over where to begin.
Physical AI is positioned not only as a means of reducing labor requirements but also as a growth investment aimed at expanding productivity. The report says it is important to begin with small-scale implementations and build up knowledge. Against the backdrop of the low overall rate of production deployment, the latest report identifies the accumulation of training data and the use of partner sharing as key factors.
The survey was conducted from 8 to 9 July 2026, with Rakuten Insight serving as the research agency. Respondents were 200 CxOs in manufacturing, logistics and infrastructure services across Japan, and valid responses were collected through a multiple-choice online survey.
The report also presents investment plans aimed at expanding the physical AI market in Japan, with approximately ¥10.5 trillion in public- and private-sector investment projected through fiscal 2040. Roland Berger calls for a phased expansion through learning and data accumulation, rather than limiting efforts to short-term adoption.