JAPAN Trends and Developments Contributed by: Haseru Roku and Yoshiteru Matsuzaki, Nagashima Ohno & Tsunematsu
Vertical AI applications: market entry and established capabilities Vertical AI investments target industry-specific appli- cations. Transaction structures reflect market-specific factors: whether established global leaders already serve the market, data regulatory requirements and the strategic positioning of Japanese enterprises. SoftBank’s SB TEMPUS joint venture with Tempus AI – a JPY30 billion combined investment announced in June 2024 – focuses on precision medicine and genomic testing for cancer treatment. The transac- tion structure reflects a distinctive market condition: Tempus, a leading US-based precision medicine plat- form, had not previously entered the Japanese market despite global expansion. The joint venture structure reflects partnership necessity: entering an unfamiliar regulatory environment – with Pharmaceuticals and Medical Devices Agency (PMDA) oversight – while maintaining governance alignment with the US parent company. This suggests that market entry strategy, rather than data exclusivity, appears to be a primary driver of transaction structure in this case, demon- strating how joint venture structures enable market entry into regulated sectors without full acquisition. AI agents for business process automation: collaboration-first approach AI agent implementation for business process auto- mation remains predominantly in the early stages of development. Technology and commercial immatu- rity favour collaboration between business enterprises and technology companies. SoftBank’s SB OpenAI Japan joint venture devel- oped Cristal Intelligence for enterprise AI agents, yet such large-scale strategic investments remain limited. NEC’s collaboration with Google Cloud, announced in August 2025, exemplifies broader market patterns: integrating AI agent functionality with established cloud infrastructure and foundation models. These partnerships prevail because AI agent technol- ogies remain technically immature, and standardisa- tion frameworks have yet to solidify. Cross-industry partnerships indicate widespread recognition that AI agent deployment requires sustained ecosystem development rather than immediate market consoli-
dation. Current market dynamics remain collabora- tion-dominant as the primary strategic focus. Physical and embedded AI: market position and industry structure Physical AI investments – AI systems integrated with mechanical platforms – present a key layer 4 excep- tion. While most layer 4 investments favour partner- ships due to technological uncertainty, physical AI requires integration with established (over a period of decades) hardware platforms, manufacturing ecosys- tems and regulatory standards – creating industry bar- riers that necessitate different transaction structures. Market positioning and existing industry structure cre- ate asymmetric strategic imperatives. New entrants lacking established robotics capabilities face challenges penetrating technology-dense mar- kets through partnerships alone. SoftBank’s JPY818.7 billion ABB Robotics acquisition in October 2025 exemplifies this necessity: ABB’s integrated platform – combining manufacturing, supply chains, customer relationships and decades of IP – cannot be replicated through partnerships. For technology investors with- out a hardware presence, platform acquisition enables physical AI deployment and market access. Established robotics manufacturers – possessing customer relationships, manufacturing infrastruc- ture, regulatory compliance and technical standards – face different imperatives. Rather than wholesale platform consolidation, they enhance existing prod- ucts through selective partnerships and targeted acquisitions. Yaskawa acquired Tokyo Robotics in March 2025 for humanoid expertise while establish- ing Astellas Pharma as a joint venture for biomedical applications. Fanuc maintains NVIDIA partnerships for deep learning while establishing Kitov AI for vision inspection, while Kawasaki partners with Dexterity for autonomous truck loading. These structures enhance capability without reorganising established systems, representing a true competitive advantage. This divergence suggests that industry structure plays a dominant role alongside layer 4 characteris- tics. Physical AI implementation intersects with sec- tors where incumbent manufacturers possess durable advantages through installed bases, regulatory com-
184 CHAMBERS.COM
Powered by FlippingBook