Physical AI News (Japan)

Physical AI, robotics and compute — reported from Japan.

日本語
· Partnerships & M&A· 日本語で読む

newmo and KDDI form capital and business alliance, aim to launch autonomous taxis in Sakai in 2028

newmo and KDDI have formed a capital and business alliance to bring autonomous taxis into practical use. KDDI will provide communications, AI computing infrastructure, remote monitoring and high-precision positioning, among other services. The companies aim to launch commercial service in Sakai in 2028 and are considering expanding their operating functions nationwide.

Key facts

  • newmo and KDDI have formed a capital and business alliance to bring autonomous taxis into practical use.
  • The companies aim to launch autonomous taxi service in Sakai, Osaka Prefecture, in 2028.
  • KDDI will provide communications, AI computing infrastructure, remote monitoring and high-precision positioning, among other services.
  • The three companies are considering providing local transportation operators nationwide with functions such as dispatching, operations management and remote monitoring that they will develop and test in Sakai.
  • newmo says it operates a fleet of about 1,900 taxis nationwide.

What happened

newmo and KDDI have formed a capital and business alliance to bring autonomous taxis into practical use. The companies aim to launch commercial service in Sakai, Osaka Prefecture, in 2028. Along with KDDI Smart Mobility, they are considering providing transportation operators nationwide with functions such as dispatching and operations management that they will develop and test in Sakai.

Why it matters

The plan is to combine KDDI’s communications, computing and positioning capabilities with newmo’s autonomous driving technology and taxi operations systems to develop and test operational functions, including remote monitoring, in Sakai. The concept goes beyond autonomous vehicle technology to provide local transportation operators with a suite of functions that support operations. Whether it can be expanded nationwide will depend on adapting operations to local demand and confirming performance, including safety.

Analysis

The focus of the alliance is not just autonomous driving for taxis, but combining communications, computing, remote monitoring and positioning needed for commercial operations with dispatching and operations management in a single operating structure. KDDI’s AI computing infrastructure is already used to train newmo’s autonomous driving AI, and KDDI’s role in the alliance will also include inference infrastructure and communications linking vehicles with monitoring centers. The plan appears to connect training, driving and remote support, but KDDI says it is still considering a platform for collecting and aggregating driving data. How much of a cycle linking operations and retraining will be implemented remains to be seen. The plan to develop and test functions in Sakai and provide them to local transportation operators can be seen as an effort to turn operations in a single area into a suite of functions that can be deployed elsewhere in the country. The companies also cite coordination with AI on-demand transportation and autonomous buses and shuttles, pointing to a design that combines different mobility services suited to each area rather than focusing only on taxi operations. Nationwide expansion, however, is still under consideration, and the companies have not said which functions would be standardized across regions and which would be adjusted locally. We have covered KDDI’s AI implementation and deployment efforts in KDDI announces Physical AI strategy and KDDI and ELYZA launch Physical AI joint venture, aim to deploy more than 1,000 robots by fiscal 2030. This alliance, separate from the robot development and deployment covered in those reports, focuses on local transportation operations. It marks a shift in the KDDI Group toward implementation that combines remote monitoring, positioning and operational support, rather than providing computing infrastructure alone. Under the business structure, newmo will handle its taxi business, autonomous driving technology and driving data; KDDI will provide infrastructure such as communications, computing and positioning; and KDDI Smart Mobility will contribute its knowledge of supporting local transportation operations. newmo says it operates a fleet of about 1,900 taxis nationwide, but that figure does not refer to autonomous vehicles. Even in light of the government’s target of deploying 10,000 autonomous service vehicles by fiscal 2030, the scale of deployment cannot be inferred from this announcement. The published announcement does not specify the investment amount or ownership stake in the capital alliance, the number of vehicles or service areas in Sakai, fares, how safety will be evaluated, or the terms for managing and using driving data. Clarifying these details will be key to assessing what commercial service will look like.

Details

KDDI will provide communications links connecting autonomous vehicles with remote monitoring centers, computing infrastructure for AI training and inference, remote monitoring and high-precision positioning. newmo will handle autonomous driving technology, a dispatch app and taxi operations systems, while collecting driving data and developing autonomous driving AI. KDDI Smart Mobility will contribute expertise in remote monitoring and operational support, and the companies will consider coordination with AI on-demand transportation and autonomous buses and shuttles. The three companies are considering offering functions such as dispatching, operations management and remote monitoring, developed and tested in Sakai, in response to local mobility needs. newmo says it operates about 1,900 taxis nationwide. The government has set a target of deploying 10,000 autonomous service vehicles, including buses and taxis, nationwide by fiscal 2030.

Sources

Translated from the Japanese original: /news/newmo-to-kddi-ga-shihon-gyoumuteikei-2028-nen-ni-sakaishi-de-1d8d43. The Japanese version is authoritative if they differ.