Phase 1 of Toyota Woven City. The small city built on the site of the former Higashi-Fuji plant is expanding in stages.

In 2018, at a factory slated for closure, an employee asked what would become of the site. Seven years later a city stood on that land, and another year has passed. We look at which of the hopes Toyota placed in the city in 2020 have held, which have changed, and whether this small town can be useful in a world Toyota does not own.

By Sang Min Han
한글로보기 




“What will happen to this place?”
In 2018, at Toyota Motor East Japan’s Higashi-Fuji plant, which had been slated for closure, an employee put a question to Akio Toyoda, then Toyota’s president. Some people wanted to keep building cars at the company’s plants in the Tohoku region, the employee said, but family circumstances meant they could not move and would have to leave the company. With them in mind, the employee could not leave with a light heart. What did Toyota intend to do with this place?
Toyoda answered that he wanted to turn it into a connected city for testing technologies such as automated driving, one that would contribute to vehicle development for the next 50 years. It was not a plan the company had approved but a personal idea he voiced for the first time, on the spot. The plant, which began operating in 1967, built about 7.52 million vehicles over 53 years, and some 7,000 people worked there over its history.
Toyoda presented the idea to the world at CES in January 2020. Toyota had first considered another autonomy test site like M-City in Michigan, he said, before deciding to build a real city where real people would live. He called the project his personal “Field of Dreams” and borrowed the film’s line.
“If you build it, they will come.”
The city was built. Woven City officially launched in September 2025, and as of September 2026 about 160 people in roughly 80 households live there as Weavers, as residents who take part in trials are called, including people from Susono City, where the site is located, and neighboring municipalities. There are 31 Inventors, the companies and organizations that run those trials, and about 30 proof-of-concept trials centered on residents have been under way. Five months earlier, in April, there were 24 Inventors and about 100 Weavers. Residents now fill a little more than half of Phase 1’s capacity of about 300. A recruitment drive for residents of Shizuoka Prefecture drew more than 1,000 applications in August, and about 25 households are scheduled to move in around March 2027. From October 1, the visitor program was opened to the public across Japan. People have begun to come, but the rollout is now more phased than the 2020 announcement suggested.
When Phase 1 construction was completed a year ago, AEM described the city as a partial success and asked whether it would become more than another R&D center. At the least, an R&D center has been built. The question for the first anniversary is what comes next. Why did a carmaker think it needed a city, and do those reasons still hold six years on?




An intersection in Phase 1, designed to put pedestrians first. Vehicles, signals and servers are tested together.



Three hopes born under one roof

At least three expectations overlapped in the Woven City of 2020.
The starting point was Akio Toyoda’s promise to an employee at a closing plant. It was also a story that Toyota, which began as a loom maker, tells about itself. Sakichi Toyoda built an automatic loom for his mother, who worked at her hand loom late into the night. At the end of his CES speech, Toyoda reminded the audience that Toyota’s origins were in loom manufacturing and said it now hoped to use its technology to build a new kind of city. The name Woven comes from that history.
The same speech also carried a business case. Toyoda said Toyota would build the city in the virtual world first, creating “a one-of-a-kind digital operating system for our city. One that perhaps others will be able to use.” Two months later, Toyota and NTT announced a capital alliance in which each would acquire roughly ¥200 billion of the other’s shares. The plan was to build a joint smart city platform, implement it first at Woven City and on NTT land in front of Shinagawa Station in Tokyo, and then extend it to other cities.
To these was added Toyota’s way of developing products: go to the site, see for yourself and fix the problem there. As software and AI come to define a vehicle’s value, problems also arise from interactions among code, data and systems. A city where people live is a place to observe those changes in daily life.
At first, all three expectations sat in one organization. When Toyota reorganized TRI-AD into Woven Planet in 2020, the operating company Woven Alpha took charge of both Woven City and the vehicle software platform Arene.
Toyoda also decided to invest in Woven Planet personally. Asked by employees why, he said:
“There are no clear answers to creating a new future. Within a company, you have organizations where you make decisions based on logical discussions, ‘investing this much to gain such and such a level of return.’ However, in this uncharted era, not everything can be considered using such logical mechanisms. … Large organizations try to think things through logically based on successful past experiences. But I do things differently.”
Over six years, the three hopes went separate ways.
Arene went into production. After founding CEO James Kuffner stepped down in 2023, Hajime Kumabe, who came from Denso, succeeded him. Arene entered production first in the new RAV4 and then expanded to the new Lexus ES. Woven City and Arene still sit inside Woven by Toyota (WbyT), but their published results have come from different places: the city’s from trials within the town, Arene’s from production vehicles. In April this year Kumabe said that “over time, we believe Arene can become more than a vehicle platform,” growing into a broader foundation for connecting people, mobility and infrastructure.
The Toyota–NTT partnership has also shifted. In October 2024 the two companies said they would jointly invest ¥500 billion by 2030 to build a “mobility AI platform” linking people, mobility and infrastructure, aiming to begin deploying it in society around 2028. The goal was zero traffic accidents. Koji Sato, then Toyota’s president, placed the project in line with the companies’ earlier cooperation, but did not explain what had become of the 2020 smart city platform.
The city’s data infrastructure has not disappeared. This year WbyT unveiled Woven City Infra Hub, a platform that unifies data across the city, and Woven City Data Fabric, a framework for using data while respecting individual preferences and privacy. Whether these carry on the 2020 platform, and how roles are divided with NTT, cannot be confirmed. A city operating system, an external business and traffic safety can no longer be explained as a single package.
The city’s role in automated driving has changed too. The 2020 blueprint included roads for autonomous vehicles. Today, learning rare road situations requires driving data at large scale, and a town of a few hundred people cannot supply that alone. In 2025 Toyota signed a preliminary agreement with Waymo to explore collaboration on autonomous driving platforms and their application to personally owned vehicles, with WbyT taking part. Woven City tests road infrastructure and vehicles together.
The size of the city has changed as well. At CES 2020, Toyoda spoke of “roughly 2000 people... to start... and more as it grows.” Today 2,000 is the final target for the full site. Phase 1 is planned for about 300, and when Phase 2 construction is completed in November 2028 the two phases together will house about 900. What was once the starting line is now the finish line.
Yet the physical design remains. Above ground, streets are divided into pedestrian-only paths, paths shared by pedestrians and personal mobility, and lanes for autonomous mobility; below ground, a logistics route has been built. The answers of 2020 have set in concrete. The questions of 2026 have moved to the software and data that travel over it.
At the Phase 1 completion ceremony in February 2025, Toyoda, by then chairman, described the place as a city that would never be finished. A finished future is only part of what the city needs to show. The rest is whether, in the same place, it can change what is being tested, fix what fails and check again.




The Inventor Garage, converted from a former plant facility. Under the crane that once served car production, prototypes are now built.



What a small town has

Phase 1 covers about 47,000 square meters. The town’s strength lies less in its size than in the ability to design the test environment as a whole. In a real city, roads, signals, building systems and communications networks have different owners, so changing a single signal requires decisions from several agencies. At Woven City, Toyota and WbyT can set test conditions by combining roads and signals with buildings and data infrastructure.
Experiments are not confined to the residential area. At the Inventor Garage, which began operating in April, Inventors build prototypes; at the Inventor Field they validate them under controlled conditions; in the residential area they test them in daily life. The Garage reuses the former plant’s stamping facility.
Automotive engineers may think of HIL. Just as a controller is connected to a virtual environment and its responses checked as inputs change, Woven City changes the relationships among vehicles, signals, servers and building systems in physical space. The difference is that people live there. So what is being verified extends from device performance to the effect on everyday life.
The “fully controlled site” Toyoda described in 2020 is an important asset. But residents must choose whether to provide data, and Inventors bring their own technologies. What Toyota and WbyT design is the test environment. Whether results are applied to other vehicles and roads is decided by many parties outside the city.
The intersection is where this design freedom shows. According to Toyota’s pre-launch description, pedestrian signals in Phase 1 are green by default and turn red when a vehicle passes. A server receives an e-Palette’s position and speed, calculates when it will reach the intersection and requests a signal change. Setting the signal’s default to favor pedestrians, and being able to design and revise vehicle-to-signal communication together: that is what the town shows first.
In practice, signals did not always change as intended. At KAKEZAN, a technology event held at Woven City in April, Daisuke Toyoda, then senior vice president of WbyT, described traffic signals “that were supposed to turn green when a vehicle approached—but stayed red for more than 30 minutes.” Problems in the connected systems also spilled into residents’ lives. He said authentication systems had gone down, leaving people locked out or stuck on rooftops, and that one night an infrastructure error turned on every light in the city, waking sleeping residents. In another case a mobility vehicle stopped working and took three people to carry. Rather than optimizing transportation, energy, identity and mobility separately, he explained, WbyT connects them and builds and operates them itself so they can be changed to suit each experiment. Showing that individual devices work and showing that daily life holds up inside a connected system are different verification tasks. The published account, however, stops at the failures. It does not say how the causes were found and fixed, or whether the problems recurred.
Separately, Toyota has been testing the Woven City Integrated ANZEN System. ANZEN, Japanese for “safety,” combines AI Vision Engine, which interprets video; Behavior AI, which interprets and predicts behavior; and Drive Sync Assist, which supports driving according to the driver and surrounding conditions. It analyzes camera footage from vehicles and traffic signals to predict behavior and passes that information to pedestrians and drivers.
WbyT plans to release AI Vision Engine 2.0 by the end of 2026, expanding its ability to describe events in natural language and adding the ability to reason about why events occurred and to detect specific behaviors. The company’s statement that it recorded world-leading scores on more than 10 benchmarks reflects results from an internal testing environment built to each benchmark’s standards and methods; it does not indicate safety effects on public roads.
This work becomes clearer when set beside another Toyota group safety effort. The Yuimaru project in Okinawa, run by the Toyota Mobility Foundation, overlaid connected-car driving data on police accident data to identify high-risk locations. It alerted drivers through in-car screens in rental cars and worked with the Okinawa prefectural police on road countermeasures. According to the foundation’s August 2025 announcement, measures such as speed-reduction poles cut the accident rate at the former Isa intersection, an accident hotspot, by about 60 percent. It was a result achieved on someone else’s roads, with the police, through small interventions.
Hirofumi Inoue, president of Toyota’s Advanced R&D and Engineering Company, said early this year that zero accidents cannot be achieved through vehicle performance alone. Where Yuimaru warned drivers and put poles on the road, Woven City can change how signals make decisions, along with the vehicle software and the cameras and AI models, and test them together.
The town’s experiments have also moved beyond the car. As of September 24, the AI Robot Association (AIRoA) had placed household robots in 11 homes and tested more than 1,000 daily-support tasks such as tidying up and carrying items. This home testing is also set to connect with developers outside the city. In July AIRoA was selected for GENIAC, a program run by Japan’s Ministry of Economy, Trade and Industry and NEDO, and said it plans to collect about 7,000 hours of robot motion data from 35 households, including households in Woven City. The data will be used to train and evaluate vision-language-action (VLA) models and organized so that other robot makers and AI developers can use it.
AIRoA said that collecting data continuously in real homes raises issues that technology development alone cannot solve, including safety, privacy and the management of residents’ consent. Residents’ daily lives and choices shape the experiments, and the experience of managing them is itself something that has to be carried outside.
There is a case in which residents’ participation changed an operational decision. In the final of the Toyota Woven City Challenge, a call for startups, residents also voted, and reflecting support from residents and judges, the organizers increased the number of winners from three to four. Taking part in deciding which experiments come in, however, should be distinguished from improving product design.
The same conditions also create limits. About 160 people live and take part in trials here. That sample is too small to produce statistics on rare accidents, and there is no guarantee that results obtained in a place designed by one company will be reproduced on public roads with many owners. What this town can encounter first are the problems that arise when vehicles and signals, robots and buildings, residents and data all enter a test together.




The Inventor Field, where technologies are validated under controlled conditions.



Three roads out of the city

The city’s value has to be confirmed outside it. There are three ways to carry its results out.
The first is production vehicles. No example has yet been published showing that a problem found in the city changed the requirements or test items of any vehicle. Being in the same company does not automatically connect Arene with the city’s results. Early this year Jean-François Campeau, a vice president at WbyT, told the story of his mother, who drives a 17-year-old Japanese car. What matters to her is not whatever changed in the software overnight but whether the car starts in the morning. Only when risks found in the city are reflected in vehicle requirements and validation, and carry through to the reliability of production vehicles, can Woven City be said to have contributed to vehicle development.
The second is public roads. On September 3, 2026, WbyT and Shizuoka Prefecture signed an agreement to promote the development and testing of technologies and products and to bring the results into society. Within fiscal 2026, WbyT will extend ANZEN testing beyond the city by installing devices on prefectural government vehicles and collecting data. Mobility service trials centered on Mishima City are also under consideration. On public roads there are vehicles of different makes and ages, pedestrians who are not connected, and existing signals.
Inside Woven City, the conditions of roads and signals, vehicles and data could be designed together. On public roads, that authority shrinks. ANZEN’s testing outside the city is a chance to see whether the same technology remains useful under conditions it cannot control.
The third is products. Technologies have emerged that can be read alongside Akio Toyoda’s 2020 remark about an operating system “that perhaps others will be able to use.” According to materials released in August, Robot Platform, which links robots to automatic doors and elevators, is being prepared for sale. Digital KAIZEN Platform, which virtually validates proposed changes at manufacturing and logistics sites, is being tested in Japan and abroad. xEV-VPP, which uses EV batteries as an energy resource, is in pre-commercial testing with 50 vehicles. Rather than selling the city, Toyota would sell tools tested in it.
How these efforts are judged also depends on management. In October 2023 Toyota made WbyT a wholly owned subsidiary, and Akio Toyoda sold the shares he had bought personally. That year Kenta Kon, a former Toyota CFO, became WbyT’s CFO, overseeing the finances of Arene and Woven City; he returned as Toyota’s CFO in 2025 and became president in April 2026. He has made lowering Toyota’s break-even sales volume a priority. The executive who oversaw the finances of the organization Akio Toyoda backed in order to work outside the logic of returns is now responsible for the profitability of all of Toyota. Daisuke Toyoda, who led Woven City from concept to launch, returned to Toyota in August. These appointments alone do not show that investment policy has changed. How Toyota evaluates long-term research will matter to Woven City’s future.




Chairman Akio Toyoda and Daisuke Toyoda on stage at the official launch in September 2025. The man who conceived the city, and the man who led its development and launch.



Field of Dreams, and the next question

Over the year, what can be tested together in this town has begun to come into view. What has not been confirmed is whether that site will become a source of competitiveness for Toyota. Toyota has not disclosed its total investment in the city, so results cannot be weighed against cost. External revenue from its technologies, ANZEN’s safety effect on public roads, and cases in which problems found in the city fed into production vehicle development have not been confirmed either. Whether this site will change the future of a carmaker is still being tested.
At the April event, Daisuke Toyoda was asked whether Woven City would still exist 100 years from now. “I want to ensure that preserving this place does not inadvertently become our sole objective,” he answered, adding that the aim was to make full use of the city to help create a better society.
Keeping the city alive and spreading what it produces are not the same thing. The question the plant employee asked in 2018 was what would become of this land. Toyota built a small city on it, and people have begun to come. That question has been half answered. The next question is what happens on that field: what game will the people gathered there play, and will the results hold up beyond the fence?

 

AEM(오토모티브일렉트로닉스매거진)