The customer calls the vehicle—and the empty PV5 drives to them

Kia announced on October 16, 2026 that a PV5-based remote-driving carsharing pilot began in Pangyo, South Korea, on October 15. It will operate for approximately three months through the end of the year. The service is limited to employees of companies located in Pangyo Second Techno Valley; it is neither a national consumer launch nor a commercial driverless mobility network open to anyone.

A qualified user books the PV5 Passenger through a dedicated app. An operator at a remote control center drives the empty vehicle to the pickup point. The customer takes possession of the vehicle and drives normally during the rental. After parking and ending the rental, a remote operator drives the now-empty car back to the depot.

This changes an important part of the carsharing workflow: where a vehicle is picked up and returned, and how delivery and retrieval are handled. The company expects convenience and operating benefits, but it did not disclose measured savings or fleet utilization results from the pilot. The verified change is an operating method, not a proven financial outcome.

No driver is inside the moving car. That does not make it autonomous

Remote driving, also called teleoperation, means a person outside the vehicle uses a control system and 4G/5G mobile connectivity to operate the vehicle. Kia's announcement describes a driver who is not on board but continues to control the car from an external center. The empty cabin is therefore not evidence that the vehicle is making all its own driving decisions.

Automated driving involves onboard systems performing driving tasks under defined operational conditions. The Pangyo demonstration is different: a human remote operator controls the vehicle during delivery and retrieval, while the customer operates it during the rental. Calling this a fully autonomous carsharing rollout would conflate two substantially different operating models.

The difference extends to responsibility. Communication quality, remote driver training, escalation procedures and traffic safety are essential when a person drives from another location. The experiment is not an example of eliminating human driving altogether; it is a test of moving part of that work outside the cabin.

Four core partners reveal why a vehicle alone is not a service

The pilot consortium includes Kia, SUM, Mobiops and the Korea Road Traffic Authority. Kia provides the PV5 vehicles and leads the project; SUM develops the remote-driving solution; Mobiops operates the customer-facing service; and the Road Traffic Authority assists with regulatory review and related issues.

Gyeonggi Province, KT, Umos One, the Advanced Institute of Convergence Technology and the Gyeonggi vehicle-rental industry association are listed as supporting organizations. Vehicle engineering, remote control, wireless communications, day-to-day fleet operations and regulatory safety involve different capabilities, even when the customer sees a single car arrive.

The important strategic observation is that manufacturing the PV5 does not by itself create a sustainable mobility service. Managing interfaces among the vehicle, the control center, customer bookings, emergency procedures and legal constraints becomes part of the product. Kia's project is therefore an operating-system experiment as much as a vehicle technology demonstration.

The pilot follows earlier technology demonstrations and commercial planning

This was not a sudden announcement. Kia's official chronology states that it obtained a regulatory sandbox exemption for remote driving in April 2025 and took part in a consortium with Socar, KT and SUM from November 2025. Tests were carried out in places including Jeju and Hwaseong, with Kia reporting a public-road remote-driving demonstration in Jeju.

In July 2026, Kia then signed a business commercialization agreement with KT and SUM. Those earlier steps concerned developing and validating teleoperation and creating a path toward commercial use. They did not establish that a broad public carsharing network was already operating at scale.

The October 2026 project moves that effort into a restricted real-user setting. A remote-driven car must not only move correctly; the reservation, dispatch, handover, customer driving, return, collection and emergency process must also work repeatedly. That is a different kind of evidence from a successful single demonstration.

When the signal degrades, the operational design becomes the product

A remote-driving system depends on communications between a human operator and a moving vehicle. Interruption of a booking-app connection is an inconvenience; loss or deterioration of a control link while a vehicle is on a road is a safety-critical operational problem. A viable service must therefore plan for more than ideal driving conditions.

Kia says the consortium has established a triple communication network to mitigate link degradation and disconnection. It also reports multiple layers of safety management, including training for remote drivers and emergency response dispatch. These are concrete control measures the company says it has prepared for the pilot.

However, a triple network is not proof that every hazard has been eliminated. Reliability depends on how the vehicle, the operator, the communications layer and field responders act together under unexpected circumstances. The company plans to collect operating data through year-end to assess safety and service feasibility; the final validation has not yet been reported.

The human driving task has moved—not necessarily disappeared

Traditional vehicle delivery and retrieval can require a person to enter the car and drive it between locations. In Kia's model, an external operator performs part of that movement from a control center. The customer still drives the vehicle while using the carsharing service.

That may change the skills and workflow involved. Operators may require training in remote vehicle controls, interpreting screen-based road information, recognizing communications failures, following escalation protocols and coordinating with staff who handle vehicle handover and field incidents. These are plausible capability implications of the described system.

But Kia did not publish how many remote drivers it employs, how many it plans to hire, or how many cars one operator can safely handle. The release does not establish job losses, hiring expansion, or a productivity gain per employee. The observed fact is the relocation of some vehicle movement tasks from inside the car to a remote operating center.

The economic unit is the whole rental journey, not just the car

A customer-facing carsharing product connects bookings, dispatch, vehicle handover, use, parking, returns, fleet recovery and preparing the car for the next booking. Remote driving changes delivery and collection, but its costs and customer benefits are embedded in the entire service process.

To judge the economics, a company would need data on vehicle utilization, delivery waiting time, recovery time, distance traveled without a paying customer, remote-operator labor, control-center infrastructure, data connectivity and safety operations. The company has not published these metrics as measured gains for the new pilot.

Kia has suggested that remote driving using existing mobile networks and conventional vehicles could require less investment than full autonomy. That statement should not be confused with a verified profit margin or unit-economics case for the Pangyo deployment. The real test is whether the workflow delivers reliable convenience at a sustainable total operating cost.

BANSEOG VIEW | The revealing change is where the human responsibility sits

The PV5 pilot demonstrates a specific division of labor: a human operator drives the empty vehicle remotely to and from the customer, and the customer drives the rental journey. There is no human driver in the cabin during dispatch or recovery, but this is not a fully autonomous driving system. The distinction is central to its engineering, safety and workforce model.

Business strategy must identify which tasks were automated, which tasks were physically relocated, and who remains responsible for decisions and exceptions. A manufacturer, teleoperation developer, telecom provider and service operator contribute different capabilities. Without measured staffing and operating data, it is premature to predict employment reductions or large efficiency gains.

Kia intends to test the system through late 2026 and expand service toward commercialization in 2027. Those plans are not accomplished outcomes. What the experiment already reveals is a new organizational question: when driving moves out of the vehicle, can the manufacturer and its partners build the control, safety and customer-service structure necessary to make the entire operation work?

BANSEOG VIEW

Banseog View — Work moves outside the vehicle before people disappear

Kia's Pangyo pilot is a limited B2B carsharing experiment: an offsite person drives an empty PV5 for pickup and return, and the customer drives the rental leg. That is different from full autonomy.

Kia, SUM, Mobiops and the Korea Road Traffic Authority divide vehicle, teleoperation, service and regulatory functions. Communications redundancy and training are safety measures, not proof of zero risk.

Operator headcount, per-operator fleet capacity, utilization gains and cost savings have not been disclosed. Commercial expansion in 2027 is prospective. The strategic question is which capabilities and responsibilities must move with the task.

SOURCES

Primary sources and references

Prepared Oct 18, 2026. Official Kia / Hyundai Motor Group press release dated Oct 16: pilot began Oct 15 in Pangyo Second Techno Valley, for tenant-company employees only, through end of 2026 (~three months). Kia PV5 Passenger booked by app; 4G/5G human teleoperator at control center brings empty vehicle to customer and retrieves it from the drop-off point; user drives the rental leg. NOT full autonomy. Consortium includes Kia (project and vehicle), SUM (remote driving), Mobiops (operations), Korea Road Traffic Authority (regulatory review) and supporting organizations. Triple communication network, remote-driver training and emergency field-response system described. Earlier sandbox and road demonstrations occurred in 2025; July 2026 commercialization MOU with KT/SUM. 2027 scale-up is an intention, not confirmed live commercialization. Kia newsroom repeats same corporate release, not an independent second source. No disclosed staffing figures, remote-operator-to-vehicle ratios, measured utilization increase or financial outcome. Business capability assessment is Banseog's independent inference, not a company hiring announcement.