优步借自动驾驶数据掘金 不造车只做数据服务商

Uber is launching a new initiative to profit from self-driving technology by collecting and analyzing real-world driving data, rather than developing its own autonomous vehicles. The company will work with its 50,000 global fleet partners to install exterior sensor kits that capture road conditions, weather, and unexpected obstacles, providing valuable information for AV developers. Uber has already partnered with Waabi to deploy 25,000 robotaxis in a $1 billion deal, and remains tight-lipped about other partners like Waymo. This shift comes after Uber sold its self-driving unit to Aurora in 2020, following a 2018 fatal incident involving one of its autonomous test vehicles.

优步正推出一项新计划，通过收集和分析真实驾驶数据从自动驾驶技术中获利，而非自主研发自动驾驶车辆。该公司将与其全球 5 万家车队合作伙伴合作，安装外置传感器套件，记录路况、天气及突发障碍物，为自动驾驶企业提供高价值信息。优步已与 Waabi 达成合作，以 10 亿美元的协议部署 2.5 万辆自动驾驶出租车，对 Waymo 等其他合作伙伴则未透露细节。此次转型源于 2018 年其自动驾驶测试车致人死亡事故后，优步于 2020 年将自动驾驶部门出售给了 Aurora。

Uber’s CTO Praveen Neppalli Naga stated that high-quality real-world data, not software or hardware, is now the biggest bottleneck for autonomous driving. The company aims to gather "long-tail data"—rare, unpredictable events like debris in the road or sudden pedestrian appearances—that simulations struggle to replicate. This data could become a major new revenue stream for Uber, which plans to charge AV partners for access, as consumer trust in self-driving cars hinges on comprehensive real-world training to ensure safety.

优步首席技术官普拉文・内帕利・纳加表示，自动驾驶的最大瓶颈已非软件或硬件，而是优质的真实世界数据。公司计划收集 “长尾数据”，即路面杂物、行人突然出现等罕见且不可预测的事件，这类场景是模拟技术难以复现的。这些数据有望成为优步的重要新收入来源，公司计划向自动驾驶合作伙伴收取数据使用费，因为消费者对自动驾驶汽车的信任，取决于全面的真实场景训练以保障安全。

Industry experts like Zachary Greenberger, CEO of Nexar, acknowledge the data opportunity but warn of challenges for Uber. Fleet drivers, who are professional, are less likely to encounter the extreme, rare scenarios that generate unique data. Greenberger also notes that Uber would need to rapidly deploy hundreds of thousands of sensors to produce useful data at scale. Despite these hurdles, Uber remains optimistic, viewing autonomous vehicles as a trillion-dollar long-term opportunity that its data strategy will help accelerate.

Nexar 首席执行官扎卡里・格林伯格等行业专家认可数据的价值，但也对优步面临的挑战提出警示。作为专业从业者的车队司机，较少遭遇能产生独特数据的极端罕见场景。格林伯格还指出，优步需快速部署数十万枚传感器，才能产出规模化的有效数据。尽管存在这些障碍，优步仍持乐观态度，将自动驾驶汽车视为万亿美元级的长期机遇，而其数据战略将助力这一领域加速发展。