Gasgoo Munich- At the 2026 China Auto Forum Technology Leaders Summit, Lu You addressed the industry's most persistent headaches. Lu You is general manager of products and markets at neueHCT. Despite being a joint venture founded just two years ago, the company addresses a sector caught in a blind race for computing power. Internal combustion engine (ICE) vehicles struggle to adopt affordable city-driving assist systems. Domestic autonomous driving tech often stumbles abroad due to unfamiliar road conditions and driving habits. Leveraging Horizon Robotics chips, neueHCT proposes a path forward. It involves tiered products paired with physical AI end-to-end algorithms. These algorithms balance cost, compatibility, and global adaptability.
An arms race in computing power is clearly underway. Premium models pack chips ranging from 2,000 to 3,000 TOPS, yet vehicle costs keep climbing. Lu You stated that neueHCT rejects unchecked hardware stacking. Instead, the company plans hardware in tiers based on model positioning. The lineup includes an entry-level solution around 10 TOPS. It also features a mid-to-high-tier system backed by the J6M chip offering 100-plus TOPS. A next-generation cockpit-driving integrated chip will roll out in stages. This approach balances functional needs with manufacturing costs. The team is focusing heavily on the J6M platform. It uses a one-stage end-to-end physical AI algorithm. This delivers safety and smoothness comparable to 200-to-300 TOPS systems, offsetting the hardware gap through software optimization.
Image source: 2026 China Auto Forum
ICE vehicles have long been constrained by thermal and hardware architecture limitations. They often require expensive retrofits to support advanced city-driving features. neueHCT's solution, however, uses a single-fan hardware structure. It fits ICE vehicles without extra cost. It delivers consistent performance across both gasoline and electric models. The algorithm moves beyond rigid rules to understand road logic. This allows for flexible judgment at obscured crosswalks and irregular intersections. It demonstrated stability on Chongqing's complex mountain roads without specialized training. It naturally adapts to niche overseas conditions like narrow European streets and cobblestone surfaces. Most overseas deployment issues are not hard algorithm failures. Instead, they stem from a lack of local data on driving habits and regulations. These problems can be solved by filling those data gaps.
Commercially, the company has secured over 50 model design wins in two years. More than 20 models are reaching mass production this year. Backed by Omov Group's overseas engineering team, the technology will launch in Europe, South America, and Australia-New Zealand this year. The mature passenger vehicle technology is also being repurposed for electric two-wheelers. It extends to autonomous logistics vehicles to open new markets. Lu You concluded that Chinese autonomous driving firms should not fixate on high computing power when expanding globally. Instead, steady internationalization relies on tiered cost control, ICE-EV compatibility, algorithm generalization, and adaptation by local teams.
