Gangnam Robotaxi Makes Its Silicon Valley Debut

SWM, Lenovo, and NVIDIA Collaborate to Unveil Next-Generation Autonomous Driving Platform with 2,000 TOPS

[SWM Gangnam Robotaxi]

[SWM Gangnam Robotaxi]

View original image

The autonomous robotaxi operating in the heart of Gangnam, Seoul, South Korea, has made its debut on the global stage in Silicon Valley, USA.


SWM, an autonomous driving technology company, announced that it would showcase its robotaxi technology at “NVIDIA GTC 2026,” the world’s largest AI developer conference, held in San Jose, USA, from March 16 to 19. This exhibition marks the first time a Korean robotaxi has been introduced on the global stage, serving as an opportunity to demonstrate both SWM’s autonomous driving technology and its global partnerships.


The main highlight of GTC 2026 is the actual “SWM Robotaxi” vehicle exhibited at the Lenovo booth. Since 2017, SWM has focused on developing autonomous driving technology and, in 2024, commercialized South Korea’s first robotaxi service in Gangnam, Seoul, collecting operational data in complex urban environments. In particular, real road driving data from the densely populated streets of Gangnam is considered a core technological asset for SWM.


Since 2025, SWM has been collaborating with Lenovo and NVIDIA to build a technology pipeline for next-generation Level 4 fully driverless autonomous driving technology. The exhibition of an actual vehicle is an extension and tangible result of the strategic partnership for joint development of next-generation autonomous driving systems that the three companies formalized at “CES 2026.” By moving beyond simple collaboration to implement real vehicles and control platforms, SWM has proven its technological competitiveness in the global market.


During the event, SWM will also unveil the next-generation autonomous driving platform “AP-700,” co-developed with Lenovo, which is based on NVIDIA’s platform. The AP-700 is a successor to the AP-500 platform currently used in robotaxis operating in Gangnam, significantly improving perception, decision-making, and control performance.


This platform is built around Lenovo’s L4 autonomous driving domain controller “AD1,” which is based on NVIDIA’s DRIVE AGX Thor vehicle AI computing platform.


With a dual DRIVE AGX Thor configuration, the system delivers AI computing performance of up to 2,000 TOPS, and is optimized to simultaneously handle generative AI and vision-language model workloads.


The AP-700 is expected to serve as the core control system for next-generation robotaxis, exemplifying global cooperation by combining SWM’s autonomous driving software, Lenovo’s automotive-grade hardware and edge computing capabilities, and NVIDIA’s AI computing technology.

At the Lenovo booth inside the NVIDIA GTC event held in San Jose, USA, key officials are posing for a commemorative photo. (From left) Bingchuan Sun, Director of Vehicle Computing Business at Lenovo; Kim Kihyuk, CEO of SWM; Yang Yuanqing, Chairman of Lenovo; Won Jonghyun, Executive Director of SWM.

At the Lenovo booth inside the NVIDIA GTC event held in San Jose, USA, key officials are posing for a commemorative photo. (From left) Bingchuan Sun, Director of Vehicle Computing Business at Lenovo; Kim Kihyuk, CEO of SWM; Yang Yuanqing, Chairman of Lenovo; Won Jonghyun, Executive Director of SWM.

View original image

Another highlight of the event was a test video utilizing NVIDIA’s next-generation autonomous driving AI model, “Alpamayo.” Alpamayo is a VLA (Vision-Language-Action) model that integrates vision, language, and action, significantly enhancing the ability to understand and judge contextual situations compared to existing autonomous driving AI models.


SWM is conducting practical application tests of the Alpamayo model through its proprietary autonomous driving data engine, “TruD.” At the event, SWM received attention for releasing both simulation and real driving footage based on the challenging urban environment of Gangnam, featuring narrow alleys and complex traffic scenarios. The video detailed how the AI learns and responds to various variables in intricate city environments.


This system is designed to continuously self-learn and advance its algorithms based on data collected from real roads. By combining NVIDIA Alpamayo’s “contextual understanding intelligence” with SWM’s technology, which has already demonstrated stable performance in Gangnam, the autonomous driving capabilities are evolving to flexibly respond to unpredictable and sudden situations.


Ki-Hyuk Kim, CEO of SWM, stated, “Our collaboration with Lenovo and NVIDIA is a significant opportunity to present SWM’s autonomous driving technology to the global market,” adding, “Based on the real-world driving data and technical expertise accumulated in Gangnam, we will showcase competitive robotaxi technology in the global market as well.”


Notably, Yang Yuanqing, Chairman of Lenovo, who visited the event in person, expressed deep interest in SWM’s robotaxi business, further solidifying the global partnership framework between SWM and Lenovo.



Taking part in GTC 2026, SWM plans to further expand its collaborative network with global AI companies and, based on the empirical data accumulated in Gangnam, will gradually pursue entry into overseas autonomous driving markets such as Japan, Southeast Asia, and the Middle East.


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing