Former Huawei AI lead Huang Qingqiu, now CTO of Morphi Robot, argues that data quality matters more than model architecture for embodied intelligence. After Morphi Robot raised more than RMB 1 billion (USD 148 million) in angel funding, Huang met 36Kr in Shanghai. He stated, “If data quality were infinitely high, even a very simple model could be trained to produce excellent results.” Huang was part of the second cohort of Huawei’s “Genius Youth” program, joining its automotive business unit in 2020 and becoming head of AI for autonomous driving. He led efforts applying AI to LiDAR, perception, sensor fusion, and, during the ADS 4.0 era, Huawei’s entire assisted driving system. He also built Huawei’s autonomous driving data engineering system from scratch, starting in the ADS 2.0 era and maturing by ADS 3.0. By ADS 4.0, the system supported the entire end-to-end stack. Huang believes embodied models lack sufficient generalization, and sees data engineering as the key breakthrough. He argues that a model’s capability ceiling depends on its parameter count, and most architecture improvements mainly increase efficiency at absorbing data, which can be offset with more data and longer training time. Therefore, he prioritizes data engineering over model architecture innovation, building a data flywheel for Morph.