51WORLD (06651): The two platforms under its umbrella, SimOne and Dataverse, have successfully completed deep integration with NVIDIA Omniverse NuRec neural reconstruction technology.

51WORLD (06651) announced that its intelligent driving simulation platform SimOne and data management and mining platform Dataverse, under the company's subsidiary Siasun Robot & Automation, have successfully completed deep integration with NVIDIA Omniverse NuRec neural reconstruction technology. From now on, 51Sim, as one of the company's three major businesses, will join hands with NVIDIA to tackle the core challenges of the autonomous driving industry and jointly build the next generation of data-driven autonomous driving simulation infrastructure. In the intelligent driving industry, real-world road test data cannot be directly used in interactive closed-loop simulation scenarios, as scene construction heavily relies on manual modeling, limiting the efficiency of large-scale verification. NVIDIA Omniverse NuRec uses 3D Gaussian splatter (3DGS) neural rendering technology to reconstruct real-world road test data of fleets into interactive simulation scenarios, fundamentally solving this industry pain point. However, the full release of NVIDIA Omniverse NuRec technology capabilities depends on deep compatibility with mature simulation platforms. The company has long been deeply involved in the end-to-end intelligent driving simulation field. The SimOne platform leads the industry in cloud-based large-scale parallel simulation execution, high-fidelity rendering of multi-modal sensors, and algorithmic closed-loop validation. The Dataverse platform is responsible for the full lifecycle management of multi-modal data for production fleets, covering collection, quality inspection, structured processing, and scene asset management. Leveraging the leading technological levels of SimOne and Dataverse in the autonomous driving industry, the company has completed the deep integration of NVIDIA Omniverse NuRec with SimOne and Dataverse, establishing a complete data-driven process from real data collection, neural scene reconstruction to closed-loop simulation execution, achieving a significant transformation from "collection to archiving" to "collection to simulation" of real-world road test data. With the rapid rise of reasoning-based intelligent driving systems represented by VLA (vision-language-action model) and world models, the intelligent driving industry is demanding unprecedented realism, diversity, and closed-loop confidence in simulation scenarios, traditional manual modeling methods are no longer sustainable. The data-driven closed-loop simulation infrastructure built by the company this time provides core support for this emerging verification demand, with significant competitive advantages and market scarcity. This integration will significantly reduce the marginal cost of scene generation, shorten project delivery cycles, and systematically enhance the company's competitive barriers. By deeply integrating into the global NVIDIA technology ecosystem, the company's brand credibility and customer stickiness in the global autonomous driving supply chain will continue to strengthen, broadening the company's business scope and revenue sources continuously.