百度开放 Apollo 自动驾驶平台

2017 年春天的自动驾驶赛道上，Waymo 是公认的领跑者，走的是封闭自研的路：技术攥在自己手里，一辆辆测试车在公路上积累里程。百度手里也有牌——它在自动驾驶上布局很早，但单打独斗显然拼不过美国巨头。于是百度换了一种打法：不藏了，把技术开放出去。

2017 年 4 月 19 日，百度发布 Apollo 计划，一口气开放感知、定位、规划等核心模块，配套高精地图与仿真工具。道理很简单：与其自己闭门造车，不如把底盘共享出来，让全世界的车企和开发者都来用你的框架，用生态的规模对抗单一对手的深度。首批合作名单里，已经出现多家车企和供应商。

Apollo 的增长速度惊人。开放平台吸引了大量开发者，很快成为全球规模最大的自动驾驶开源生态之一。对很多想切入自动驾驶、又没有全栈能力的车企来说，Apollo 提供了一条捷径。中国厂商在这个赛道上的话语权，很大程度上就是从 Apollo 开始建立的。

当然，开放不是万能药。自动驾驶最难的从来不是框架，而是安全验证与真实场景的海量数据；Apollo 也经历了路线调整、商业化波折。但它验证了一个重要的行业思路：当技术差距没有大到无法跨越时，生态竞争可以改写强者通吃的剧本。

回看 Apollo，它的意义在于为自动驾驶行业提供了一条与 Waymo 不同的路径假设——封闭与开放，哪条路通向终点在当时没人知道。多年后，当 Robotaxi 开始在中国城市落地，当各家车企的辅助驾驶普遍成熟，人们回望 2017 年那场发布会，会记得那是中国自动驾驶第一次以「开放」的姿态，向世界亮出了自己的底牌。

In the spring of 2017, Waymo was the undisputed leader on the self-driving track, running a closed, in-house play: technology kept under lock and key, test cars quietly piling up miles on public roads. Baidu held cards too—it had started early on autonomous driving—but going it alone against an American giant looked like a losing game. So Baidu changed tactics: stop hiding the technology and open it up.

On April 19, 2017, Baidu unveiled the Apollo program, opening core modules for perception, localization, and planning at once, complete with HD maps and simulation tools. The logic was simple: rather than building in secret, share the chassis and let automakers and developers around the world build on your framework, using the scale of an ecosystem to counter the depth of a single rival. The initial partner list already included several automakers and suppliers.

Apollo grew at a startling pace. The open platform drew developers from everywhere and quickly became one of the world's largest open self-driving ecosystems. For automakers wanting to enter self-driving without full-stack capability, Apollo offered a shortcut. China's voice on this track was, to a large extent, established from Apollo onward.

Of course, openness is not a magic cure. The hardest part of self-driving was never the framework but safety validation and massive real-world data; Apollo also went through route adjustments and commercial ups and downs. But it validated an important industry idea: when the technology gap is not insurmountable, ecosystem competition can rewrite the winner-take-all script.

Looking back, Apollo's significance lies in offering the self-driving industry a path hypothesis different from Waymo's—closed versus open, and no one knew which would reach the destination. Years later, when robotaxis began landing in Chinese cities and assisted driving matured across automakers, people looking back at that 2017 launch would remember it as the moment China's self-driving first showed its cards to the world with an open posture.