Can AI Solve a City’s Problems? Sakai Builds a Co-Creation Pipeline Around Residents, Researchers and Startups

“Use AI to solve local problems” is an easy sentence to put in a strategy. The difficult work begins when the problem becomes specific.

Whose problem is being solved? What data can be used? Who is accountable when the model is wrong? Does a pilot become a service with a budget and operator, or disappear when the subsidy ends? Can an elderly resident use it without a smartphone? Does it save public employees time, or simply create another system they have to supervise?

Sakai City is trying to build an institutional answer to those questions. On October 8, the city will co-host the “SAKAI AI Co-Innovation Kickoff” with Osaka Metropolitan University and SoftBank. On September 14, Sakai announced the startups and speakers for the event, which is intended to bring companies, researchers, students, investors, financial institutions and public agencies together around AI, social problems and new business creation.^[1][2]

The distinction matters: the September announcement did not unveil a completed three-party AI product or a formal joint venture. Sakai describes the October meeting as a place to share problems and technical efforts and to create encounters from which future co-creation projects might emerge.^[1]

A 50-person kickoff rather than a mass conference

The October 8 program is scheduled at Osaka Metropolitan University’s Innovation Academy Smart Energy Building in Sakai’s Naka Ward. Mayor Hideki Nagafuji will open the event. SoftBank executive Kunimitsu Asanuma and Osaka Metropolitan University professor Hirokazu Fukuda are scheduled to give keynote talks, followed by researchers and startups working across AI, medicine, agriculture, dental care and wireless medical devices.^[1]

Why Sakai? The data center is only part of the answer

Sakai explicitly points to a new physical fact in the city: SoftBank’s large AI data-center project at the former Sharp Sakai plant. SoftBank acquired the site with the aim of converting it into AI infrastructure and making computing capacity available not only for its own generative-AI work but also to companies, universities and research institutions.^[3]

SoftBank’s 2026 plan describes a 140-megawatt Osaka Sakai AI data center capable of housing an AI computing platform rated at 110 exaFLOPS. The company is building what it calls an AX Factory around AI transformation, and has also announced battery manufacturing on the site as part of a broader industrial cluster.^[4][5]

But GPUs do not solve municipal problems by proximity. A data center does not automatically reduce an elderly resident’s transportation burden or a ward office’s phone queue.

Nakamozu is being repositioned from a university district to a test bed

Sakai’s Basic Plan 2030 designates Nakamozu as an innovation-creation hub. The district combines Osaka Metropolitan University’s Nakamozu campus with the Sakai City Industrial Promotion Center, S-Cube startup center, the Chamber of Commerce and rail connections to central Osaka. The city says it wants researchers, students, startups, SMEs and larger companies to create new business together.^[6]

Osaka Metropolitan University opened in April 2022 through the merger of Osaka City University and Osaka Prefecture University. The former prefectural university’s Nakamozu campus became part of the new university and remains a major engineering, agriculture and systems-science base.^[7]

Sakai and the university renewed their comprehensive partnership in March 2024, inheriting an agreement dating to 2008 and committing to use the university’s broad research capacity for social-problem solving and regional development.^[8]

Since July 2022, Sakai, the university and the local industrial-promotion organization have also placed a dedicated coordinator inside the university to connect research seeds and researchers with Sakai companies. In March 2026, that program ran a seminar for local firms on the characteristics and actual adoption of AI technology.^[9]

“Make Sakai a demonstration field for startups”

The AI kickoff sits on top of an existing city policy rather than starting from zero. Sakai’s Startup Demonstration Promotion Program explicitly uses the phrase “make Sakai a demonstration field for startups.”

In fiscal 2026, the city received 52 proposals from around Japan and selected four. Support can include access to public facilities or cooperating private sites, matching with local businesses, a subsidy of up to ¥1 million, introductions to city departments with relevant administrative problems and publicity support.^[10]

One selected company is AdDice, which is testing a voice-analysis AI and interactive affirmation application intended to explore whether spoken interaction can reduce loneliness and isolation and support self-conditioning. AdDice’s CEO is also scheduled to appear at the October AI event.^[1][10]

Dental MaaS has already operated outside City Hall

Another October participant, O-Gai Holdings, is based in Sakai and develops digital dentures and a mobile dental-care platform. On February 25, the company worked with Sakai City and Tamanoi Vinegar to run a dental screening demonstration using its O-Gai medical MaaS vehicle at Sakai City Hall and Xavier Park.^[11]

That project is not itself the new three-party AI initiative. It nevertheless illustrates the model Sakai is trying to cultivate: combine a local problem, startup technology, public space, a private-sector partner and an evaluation opportunity.

The problems belong to a city of 800,000 people

Sakai’s estimated population was 800,041 on August 1, 2026. It is a designated city with responsibilities spanning welfare, health, transportation, education, disaster management and industrial development.^[12]

The city counted 228,856 residents aged 65 or older in September 2025, an aging rate of 28.3%. It projects that ratio to rise to 33.7% by 2040. Residents aged 75 and over numbered 143,213, while 42,167 were 85 or older.^[13]

That demographic pressure creates both the case for AI and a constraint on it. Demand for care, health and mobility grows while the workforce available to provide services becomes tighter. Yet a city cannot respond by moving every service onto a digital interface and assuming everybody will follow.

Sakai has already tested AI mobility — and learned that a pilot is not deployment

In Senboku New Town, Sakai and its smart-city consortium carried out three rounds of AI on-demand bus trials. Instead of fixed routes and timetables, riders booked trips and an AI system was intended to optimize vehicle routing.^[14]

The city’s current page contains an important update: service that had been planned from November 2025 through October 2026 was not implemented. Sakai says it is considering the possibility of implementation under a new scheme and new conditions.^[14]

That should not automatically be labeled a failed technology. It does demonstrate a central lesson in civic innovation. A routing algorithm may work while the service model does not yet work.

AI is also entering Sakai City Hall

Sakai’s ICT Innovation Promotion Office says fiscal 2026 priorities include expanding the use of data and generative AI, sharing practical use cases, opening regular consultation opportunities and using data analysis to improve decision-making and productivity.^[15]

City strategy documents describe broader use of Microsoft Copilot, department-specific internal chatbots and experiments with AI summarization, classification and analysis of recorded telephone calls.^[16]

Sakai’s public-private demonstration program also includes an AI system to assist fire-investigation work by supporting document preparation, case search and accumulated investigative knowledge, with possible use as a foundation for wider fire-department digital transformation.^[17]

That makes AI governance an infrastructure requirement

Municipal data are not ordinary commercial data. Welfare, medical, child, tax, consultation and disaster-response records can affect fundamental interests of residents.

Japan’s Digital Agency revised the national government’s generative-AI procurement and utilization guidelines in June 2026. The document explicitly frames promotion and risk management as two sides of the same policy: public organizations should use AI, but only with governance around procurement, data, evaluation and operational responsibility.^[18]

Local governments are institutionally distinct from central ministries, but the questions are similar. What information may be sent to a model? Is it retained for training? Can an answer be audited? Which decisions require a human? Can data be exported if the vendor changes?

Do not reshape the public problem to fit the AI

There is a subtler policy risk: technologies can make governments prioritize problems that are easy for the technology to address.

Loneliness may involve housing, poverty, family relationships and care needs. A conversational AI may provide one useful layer without solving those causes. A transportation gap may result from low population density and driver shortages; an optimization model does not automatically create a profitable service. Limited dental access involves clinicians, reimbursement and continuity of care as well as mobile equipment.

AI should be selected to fit a public problem, not used to redefine the problem until it fits the AI.

For startups, City Hall is a difficult but valuable laboratory

Municipal pilots are hard for young companies. Public procurement, privacy, annual budgeting, resident communication and political accountability are conditions startups may never encounter in a normal corporate proof of concept.

They are also valuable constraints. Sakai’s 52 applications for four 2026 demonstration slots suggest demand for access to real civic settings.^[10]

SoftBank brings more than GPUs

SoftBank’s most visible local asset is the Sakai AI data-center project. Its potentially more important contribution to co-innovation may be commercialization and operational infrastructure.

The company already sells communications, cloud, security and AI services to corporate customers. If a model developed by a university or startup proves useful, SoftBank can potentially contribute network architecture, identity, cloud operations, security and scaling expertise.^[3][4]

That strength creates a corresponding governance issue. A municipality that builds essential services tightly around one provider can create vendor lock-in. Interoperability, data portability and an exit plan should therefore be designed at the beginning of a public-private AI project, not after procurement.

The university’s role may be to measure rather than merely invent

Osaka Metropolitan University spans engineering, medicine, agriculture and the social sciences. Its contribution to municipal AI can therefore extend beyond model development to evaluation: what should be measured, whether a medical outcome is meaningful, whether a behavior actually changes and whether benefits are distributed fairly.

The university’s Startup Creation and Support Center runs gap-funding, business-plan and researcher-company programs intended to move academic knowledge toward practical use.^[19]

Smart cities have struggled with the gap between pilots and permanence

Sakai has made Senboku New Town a priority smart-city area and created a public-private consortium that describes government and private participants as equal partners. Its stated goal is not merely experimentation but settlement and sustainable development of new services in the community.^[20]

The key word is permanence. A technology demo and a five-year welfare or transit service are fundamentally different operational achievements.

Five numbers that should determine whether Sakai’s AI co-creation works

Japan.co.jp’s analysis is that attendance, networking and startup counts will be poor measures of the October kickoff’s eventual value.

If Sakai wants to become an AI city, residents should be able to see the results

Sakai is assembling unusually complementary pieces: an 800,000-person real-world city, a major public university, an established cluster of business-support institutions in Nakamozu, startup test-bed programs, years of smart-city experimentation in Senboku and a large new AI data center.

Those ingredients do not automatically create an “AI city.”

What turns computing capacity into public value is a process: publish the problem, find partners, test at small scale, document failure, measure outcomes, protect resident data and move only successful approaches into sustainable budgets or businesses.

The SAKAI AI Co-Innovation Kickoff can be the first meeting in that process. Its own name is a useful reminder: a kickoff is not the result.

Can AI solve a city’s problems? The answer will not be found in a model benchmark or the exaFLOPS of a local data center. It will be found in whether an older resident can get where she needs to go, whether an isolated person reaches support and whether a municipal employee gains more time to deal with the cases that require a human being. That is the standard by which Sakai’s experiment should eventually be judged.