It is reported that the co-founder and technical lead of the embodied AI startup SynapX will soon be leaving the company. This departure of the technical co-founder—who was responsible for algorithm models—comes less than a year after the company’s inception.
The startup’s founding team attracted significant investor interest—raising over a billion yuan—thanks to the co-founders’ backgrounds at Horizon Robotics and their prior experience in the intelligent driving sector.
However, the founding team currently appears to lack strong cohesion.
The departing technical co-founder is a veteran talent in domestic machine learning and computer vision research. Having previously worked on fingerprint and facial recognition technologies at various tech companies, he later entered the intelligent driving industry to lead perception technology, where he connected with the startup’s core co-founder.
Amid the booming embodied AI sector, the pair chose to pivot from intelligent driving to this new field. They successfully formed a team and secured funding, subsequently releasing their own embodied AI models and hardware products.
Yet, they seem to have fallen into a common pattern within the embodied AI industry: high turnover among technical co-founders.
Beyond the frenzy of fundraising, a recurring trend in the embodied AI industry this year has been the high departure rate of technical partners.
Technical co-founders have left dozens of embodied AI startups, and the phenomenon is becoming increasingly common. The reasons behind these frequent departures are complex.
In some cases, technical progress stalled—such as failing to deliver model demos on schedule—particularly for those transitioning from other sectors without specific expertise in embodied AI models. Internal accountability measures often put technical co-founders at a disadvantage, leading to them being marginalized or replaced by high-profile “star” hires brought in from the outside.
In other instances, ambitious technical co-founders—confident in their mastery of the technology—decided that rather than playing second fiddle to someone else, they would strike out on their own to raise capital and become the “boss.”
There are also cases where co-founders lacked deep prior relationships; they had simply come together as a “hastily assembled team” because they all recognized the investment boom in embodied AI and wanted to launch a startup. When co-founding teams lack mutual trust, they are prone to fighting over the spoils—especially during periods of soaring valuations and fundraising—leading to internal friction that can drive technical co-founders to quit.
A contact in the investment community noted that the underlying motivation for co-founding an embodied AI startup is crucial, as it dictates the company’s approach in every respect.
With the recent boom in fundraising, many seasoned technical executives are eager to launch their own startups, yet their motivations vary widely. Some co-founding teams are driven by a long-term vision, enabling them to work together steadily and effectively. Others, however, are simply chasing the hype surrounding embodied AI in hopes of a quick payday; they tend to prioritize external PR while engaging in internal factionalism and power struggles.
However, the recent cooling of the embodied AI investment market marks a turning point, shifting the focus squarely onto scenario-based delivery. Previously, companies would simply showcase model demos—often securing investment from investors who didn’t fully grasp the technology but were willing to back it anyway. Now, with IPO markets tightening, commercialization has become the top priority and the primary focus for investors.
Merely showcasing model demos is no longer as effective for fundraising as it once was. Most investors lack the technical expertise to understand or evaluate these demos; instead, they excel at assessing commercial viability—evaluating factors such as target application scenarios, the potential scale of delivery, and revenue generation.
In short, the embodied AI sector has moved past the “kindergarten” phase—where no formal coursework or exams were required—and has entered the “elementary school” phase, characterized by intense academic competition.