Uber Japan has named Hinomaru Kotsu Co., Ltd. as the fleet operator for its planned Tokyo robotaxi pilot, announcing on August 13, 2026, that the century-old taxi company will manage day-to-day vehicle operations for a service scheduled to launch in late 2026 — the first time Uber has deployed autonomous vehicle technology in Japan, and the result of a regulatory framework that made an authorized taxi partner not optional but legally mandatory.
Under Japanese law, commercial passenger transport requires a licensed taxi company to operate the service. Uber, as a platform provider, cannot run a commercial robotaxi service in Japan without an authorized operator taking on the legal role of fleet manager. Hinomaru Kotsu will fill that role, handling depot operations, vehicle cleaning, scheduled maintenance, safety inspections, battery charging, and overall fleet uptime — while Uber’s app provides the ride-matching layer and Wayve’s AI Driver software handles the driving.
The vehicles will be Nissan LEAF electric hatchbacks equipped with Wayve’s end-to-end AI system. During the initial phase, experienced Hinomaru Kotsu drivers will ride aboard as safety operators. Fully driverless service — no human in the car — is planned for a later stage, contingent on regulatory approval from Japanese authorities.
Why Japan’s Law Made Hinomaru Kotsu a Requirement, Not a Choice
The regulatory logic is structural. Japan’s Road Transport Law requires that anyone running a commercial passenger vehicle service hold a Second-Type Driver’s License and operate as an authorized taxi company. Uber’s standard ride-hailing model — where personal-vehicle owners accept fares as contractors — is not commercially viable at scale in Japan, a constraint that has shaped every foreign platform’s Japan strategy. Uber operates in Japan today exclusively through UberTaxi, which partners with licensed taxi operators in the same framework.
This is not a new obstacle Uber has discovered. It is the same regulatory wall that forced Uber to adapt its business model years ago, and the same one that explains why Japan’s dominant ride-hailing app, Go Inc., was built from day one as an aggregation layer over licensed taxi fleets rather than as a contractor-hiring platform. For the robotaxi pilot, the legal structure means Hinomaru Kotsu — not Uber — will be the authorized operator responsible for compliance with Japanese licensing requirements. Uber’s official press release makes this explicit: “This initiative is strictly designed to comply with domestic laws, regulations, and licensing requirements.”
The practical consequence for riders is straightforward: booking a robotaxi through the Uber app in Tokyo is legally identical to booking a licensed taxi — because it is one. The autonomous driving technology does not change the regulatory category of the service; Hinomaru Kotsu’s authorization does.
Hinomaru Kotsu Is Not a Stranger to Either Uber or Autonomous Vehicles
The partnership is not a cold introduction. Hinomaru Kotsu has operated Uber’s premium electric vehicle fleet in Tokyo since 2023, running 100 Tesla Model Y crossovers across the city’s 23 wards as Uber Premium’s first EV fleet in Japan. That existing relationship — with shared operational systems, charging infrastructure already in place, and a demonstrated working relationship between the two companies’ Tokyo teams — makes the robotaxi extension a natural next step rather than a new vendor procurement.
The company also has a prior history in autonomous vehicle testing. Hinomaru Kotsu partnered with autonomous vehicle startup ZMP Inc. to trial self-driving taxi technology in Tokyo’s Otemachi-Roppongi corridor, one of Japan’s earliest urban autonomous driving demonstrations.
Kazutaka Tomita, Hinomaru Kotsu’s president and representative director, framed the announcement explicitly in terms of Japan’s mobility crisis: “As we address the pressing societal challenge of driver shortages, we are committed to building a sustainable mobility service that everyone can rely on with confidence — one where people and technology complement and enhance each other.” Tomita added that the pilot aims to model how human-driven taxis and robotaxis can operate alongside rather than in competition with each other — a framing that reflects the labor-political sensitivity of any autonomous vehicle deployment in an industry where driver scarcity is already acute.
Japan’s Driver Shortage Makes Tokyo More Urgent Than Any Other Market
The demographic context behind the partnership is not incidental. Japan’s taxi driver workforce has contracted by nearly 40 percent since 2009, falling from approximately 376,000 drivers to 231,938 in 2023. The median age of a licensed taxi driver in Japan is now over 60. Japan’s health ministry recorded 686,061 births in 2024 — the first time births fell below 700,000 since records began in 1899, and the ninth consecutive year of decline. Nearly one-third of Japan’s population is now over 65.
Strict overtime rules that took effect in April 2024 capped transport workers’ additional hours, compounding the structural shortage with immediate operational impact. Japan’s government has responded by setting a target of realizing unmanned autonomous mobility services across more than 100 municipalities by fiscal year 2027, a goal enshrined in the 2025 New Capitalism implementation plan.
Sarfraz Maredia, Uber’s Global Head of Autonomous Mobility and Delivery, described the Hinomaru Kotsu partnership as the delivery mechanism for that ambition: “We are very happy to partner with Hinomaru Kotsu, a deeply trusted operator, to bring this pilot deployment to life in Tokyo.”
Inside Wayve’s AI Driver: What “End-to-End” Means in Tokyo Traffic
The technology at the center of the pilot is Wayve’s AI Driver, built on the company’s AV2.0 platform — an architecture that replaces the modular “sense-plan-act” pipeline that has dominated autonomous vehicle development since the DARPA Grand Challenges of the mid-2000s. In a conventional autonomous vehicle system, separate neural networks handle perception, prediction, planning, and control in sequence, connected by hand-engineered interfaces. Wayve’s approach trains a single end-to-end neural network to map raw sensor data directly to driving commands — steering, braking, and acceleration — with no intermediate representation required.
Because the model learns from real-world driving data using self-supervised training — without requiring human annotators to label every pedestrian and traffic cone in every frame — it can potentially generalize to new driving environments without city-specific map preparation. Wayve demonstrated this claim at scale during its AI-500 Global Road Trip, navigating more than 506 cities across Europe, North America, and Japan with a single model and without rebuilding it for each location.
The Tokyo deployment will put that generalization claim to a specific test. Japanese road environments present distinctive challenges: kanji and trilingual signage, triangular stop signs unfamiliar to models trained primarily on European roads, narrow urban streets, dense mixed-traffic conditions with cyclists and pedestrians in close proximity, and some of the world’s most demanding urban driving complexity. Wayve’s model was reportedly exposed to Japanese roads during the AI-500 road trip, and the company has opened a dedicated Japanese research hub with support from investor SoftBank to improve the model’s performance on local driving data.
How the Tokyo Vehicle Differs From the London Fleet
One distinction between the Tokyo and London deployments deserves specific attention, because the two hardware configurations are not identical despite sharing the same AI Driver software.
The London fleet — the Ford Mustang Mach-E vehicles that received Transport for London’s Private Hire Vehicle license on August 5, 2026 — operates with six surround cameras and radar only. No LiDAR sensors are present. This is central to Wayve’s “mapless AI” positioning: the London fleet demonstrates that end-to-end neural networks can drive without LiDAR or pre-built maps.
The Tokyo Nissan LEAF prototype announced at NVIDIA GTC in March 2026 has a different sensor suite. The LEAF robotaxi prototype includes high-resolution cameras providing 360-degree coverage, high-performance surround and forward imaging radar, and a forward-facing LiDAR sensor providing “complementary sensing” for scene understanding and decision-making, according to the Wayve-Nissan GTC prototype release. The vehicle also runs on NVIDIA DRIVE Hyperion — a reference compute platform featuring dual DRIVE AGX Thor processors — rather than the compute architecture in the London vehicles.
The underlying end-to-end AI software is the same platform in both cases. But the sensor hardware is not. Calling the Tokyo vehicle “LiDAR-free” or “mapless” in the same sense as the London fleet would be inaccurate. The Tokyo LEAF adds LiDAR for additional redundancy and coverage, reflecting a different engineering decision appropriate to a different vehicle platform and validation context.
Uber’s Strategic Need: What the Waymo Exit Means for This Deal
The Tokyo announcement arrives at a moment of significant strategic pressure for Uber’s autonomous mobility program. Waymo, Alphabet’s robotaxi subsidiary, notified Uber in July 2026 that it plans to launch its own independent app in Austin and Atlanta in January 2028, when the current exclusivity arrangement permits — effectively signaling an end to their exclusive partnership in those markets. The two companies had already ended an arrangement in Phoenix in 2026. The Uber-Waymo contract for Austin and Atlanta runs through May 2028.
For Uber, which has committed more than $10 billion to acquiring autonomous vehicles and taking equity stakes in AV developers, the Waymo exit removes its highest-profile US robotaxi partner and raises the stakes on every other deployment in its portfolio. Wayve represents Uber’s clearest near-term response: Uber has invested in Wayve’s 1.2-billion-dollar Series D (February 2026, valuing Wayve at $8.6 billion) and has committed an additional $300 million in milestone-based capital contingent on Wayve deploying robotaxis across its global network — giving Uber a direct financial stake in Wayve’s commercial success, not just a booking partnership.
The March 2026 MOU between Uber, Wayve, and Nissan identified Tokyo as the first step in a planned rollout across more than ten cities worldwide, including London. The Hinomaru Kotsu signing converts that framework agreement into an operational reality. If the Tokyo pilot launches successfully and scales toward driverless service, it becomes the template for Uber’s Wayve-powered deployments in other markets — and the argument that Uber’s multi-partner AV network can absorb the loss of Waymo without losing its position as the world’s leading autonomous mobility platform.
Nissan, the third party in the March MOU, brings its own urgency to the arrangement. Under CEO Ivan Espinosa’s Re:Nissan recovery plan, the automaker has identified autonomous vehicle partnerships as a strategic pillar for its return to profitability. Nissan and Wayve have already signed definitive agreements to integrate Wayve’s AI Driver into Nissan’s next-generation ProPILOT advanced driver-assistance system for consumer vehicles, with deployment planned for fiscal year 2027. Nissan has also separately announced a commercial autonomous taxi service planned for Yokohama from fiscal year 2027, using a people-mover van, through its own operational framework rather than through Uber’s network. Espinosa, who holds a mechanical engineering degree from Tecnológico de Monterrey and rose through Nissan’s product planning division, has made the Wayve partnership one of the Re:Nissan plan’s technology bets.
Tokyo Is Already Becoming a Robotaxi Battleground
The Uber-Wayve-Nissan-Hinomaru Kotsu coalition is not the only autonomous vehicle effort targeting Tokyo. Waymo — operating in a separate partnership with Go Inc. and licensed taxi operator Nihon Kotsu, rather than through Uber — began human-driven HD-mapping operations in seven central Tokyo wards in April 2025. As of March 2026, Waymo had told Tokyo media that a commercial launch could arrive within months, though no date has been confirmed. The Waymo-Tokyo deployment, if it reaches commercial service, will run through Go’s app platform — making the Tokyo robotaxi market a direct test of two very different technological and commercial architectures: Wayve’s mapless end-to-end AI on the Uber network versus Waymo’s HD-map-dependent modular system on the Go network.
Nissan itself has a competing plan: a separate commercial autonomous taxi service in Yokohama — distinct from the Uber-Wayve-Hinomaru Kotsu pilot — using a people-mover van, targeted for fiscal year 2027. That service would use Wayve’s AI Driver through Nissan’s own operational framework rather than through Uber’s network.
What Comes Next: What Still Depends on Regulators
Uber has not disclosed the specific geographic coverage area, vehicle count, or launch date within “late 2026” for the Hinomaru Kotsu pilot. The rollout will be phased, with safety operators aboard for the initial period. Fully driverless operations — no human in the car — require separate regulatory approval from Japan’s Ministry of Land, Infrastructure, Transport and Tourism under the country’s Level 4 autonomous vehicle permit framework, a process that Uber and Wayve have not publicly detailed a timeline for.
Japan’s regulatory infrastructure for Level 4 autonomous vehicles is operational: the Road Traffic Act’s 2022 amendment established a permit mechanism for specified automated driving without a human driver, and Japan’s first Level 4 commercial mobility service began in Eiheiji, Fukui Prefecture in May 2023. The government’s stated target of 100-plus municipalities with autonomous mobility services by FY2027 indicates political will to accelerate approvals. But urban Tokyo, with its density, complexity, and internationally mixed traffic, represents a categorically different regulatory challenge from the rural routes where Japan’s earliest Level 4 approvals were granted.
For Tokyo riders, the near-term reality is a supervised robotaxi service available through the Uber app, with a human safety operator alongside the AI. That is not yet the driverless future either Uber or Wayve is promising. But it is the legally available starting point — and in a city where the taxi driver median age is over 60 and the workforce is contracting every year, even a supervised autonomous pilot represents a meaningful step toward a transport system that does not depend on a shrinking pool of professional drivers.
Frequently Asked QuestionsHow does Wayve’s AI Driver technology work in the Tokyo Nissan LEAF, and is it really driverless?
Wayve’s AI Driver uses an end-to-end neural network — a single model that takes raw sensor data directly as input and produces steering, braking, and acceleration commands as output, with no separate perception, prediction, or planning modules in between. The Tokyo Nissan LEAF prototype runs on NVIDIA DRIVE Hyperion compute platform and includes cameras, surround imaging radar, and a forward-facing LiDAR sensor for complementary sensing. During the initial pilot phase, an experienced Hinomaru Kotsu safety operator will ride aboard every vehicle. Fully driverless operation — with no human in the car — is planned for a later phase, subject to regulatory approval from Japan’s Ministry of Land, Infrastructure, Transport and Tourism. The “robotaxi” label describes what the service is building toward; what launches in late 2026 is a supervised autonomous vehicle with a human safety operator present.
Why does Uber need a Japanese taxi company as partner: can’t it just run its own fleet?
No. Japan’s Road Transport Law requires commercial passenger transport services to be operated by authorized taxi companies holding the necessary licenses. Uber’s standard business model — recruiting personal-vehicle owners as contractors — is not legally viable in Japan. Every ride-hailing app that has entered the Japanese market, including Uber itself, has been required to operate exclusively through licensed taxi partners. Hinomaru Kotsu holds the legal authorization to operate the vehicles; Uber provides the booking platform and operational support tools; Wayve provides the AI driving software. The three-way structure is a legal necessity, not a business preference.
When will Tokyo riders be able to hail a fully driverless Uber robotaxi with no safety driver?
No date has been disclosed. The late 2026 launch will require safety operators in every vehicle. Transitioning to fully driverless commercial service in Japan requires a separate regulatory approval under the country’s Level 4 permit framework — a process Uber and Wayve have not publicly detailed a timeline for. Japan’s government has approved Level 4 commercial operations in limited rural environments, but urban Tokyo represents a more complex regulatory challenge. The timeline will depend on how the supervised pilot phase performs and how quickly Wayve, Uber, and Hinomaru Kotsu can build the regulatory case for removing the safety operator.
Does Wayve’s approach differ from how Waymo is preparing to launch in Tokyo?
Significantly. Waymo’s system relies on centimeter-accurate HD maps of every operational zone, built through extensive manual mapping operations — Waymo-equipped Nihon Kotsu vehicles have been manually driving Tokyo’s streets since April 2025 to collect the mapping data Waymo’s system requires before it can navigate autonomously. Waymo’s sensor suite is also more hardware-intensive: its fifth-generation vehicles use five LiDAR units, 29 cameras, and six radar sensors, paired with those pre-built maps. Wayve’s AI Driver is trained on unlabeled real-world driving data and is designed to operate in new environments without city-specific mapping preparation — what the company calls zero-shot generalization — a capability proved at scale across 506 cities without city-specific retraining. The London Mach-E fleet has no LiDAR at all; the Tokyo LEAF prototype adds a forward-facing LiDAR for additional redundancy. Whether the mapless approach will prove equally safe and capable in Tokyo’s uniquely dense and complex urban environment is the central technical question the pilot will begin to answer.
AloJapan.com