Yao Song / 姚颂

Yao Song is the central guest in 173: 对话姚颂：深鉴、东方空间、再出发，「天才少年」十年后, where LateTalk presents him as a 1992-born continuous hard-tech founder. His path runs from 清华大学 electronic engineering and Wang Yu / 汪玉’s lab to DeePhi Tech / 深鉴科技, Orienspace / 东方空间, and Striding AI / 正行创新.

The source uses Yao as a decade-long sample of Chinese hard-tech entrepreneurship. His first startup gave him Hard Tech Fundraising and Strategic Acquirer Fit experience; the Xilinx / 赛灵思 acquisition then created a Post-Acquisition Founder Identity problem because financial success did not give him a durable operating mission. Orienspace / 东方空间 added large-project strategy and Chinese commercial-space execution, while Striding AI / 正行创新 is his attempt to build a long-lived company around physical intelligence as a system stack.

Key Claims

• Yao says he chooses directions by combining important human problems with personal enjoyment in turning a technical starting point into useful products.
• His Founder Risk Calibration is bounded: he supports going all in with time and attention, but not destroying the option to keep acting if the first attempt fails.
• He sees physical intelligence as more complex than a model or robot body because data, compute, hardware, scenes, remote operation, and product delivery all have to work together.
• He treats transparency around robot demos as a long-term operating boundary, especially when industry videos blur teleoperation and autonomy.

Connections

• DeePhi Tech / 深鉴科技, Xilinx / 赛灵思, and Wang Yu / 汪玉 — first startup and acquisition path.
• Orienspace / 东方空间 and Gravity-1 Rocket / 引力一号 — commercial-space path.
• Striding AI / 正行创新, Physical AI, and Embodied AI — current physical-intelligence path.
• Founder Motivation Evolution, Founder Risk Calibration, Post-Acquisition Founder Identity, and Hard-Tech Direction Selection — founder-pattern concepts grounded by the episode.