RoboBoom
Automate Show· September 10, 2026· 30:40

Apptronik on How Embodied AI Builds Intelligence for the Real World | A3 Humanoid Robot Forum

Jeff Cardenas, co-founder of Apptronik, discusses how embodied AI and practical design choices could help humanoid robots scale, at the A3 Humanoid Robot Forum hosted by Automate Show.

In this talk

  1. 0:49

    From feasibility to deployment

    Apptronik began in 2016 with a focus on bringing humanoid robots into commercial use. Cardenas says the field has moved from debating whether humanoids can work to figuring out how to scale them reliably. He describes the company’s decade of development as including 10 humanoid designs and about 80 electric actuator iterations.

  2. 5:35

    Choosing the right robot form

    Cardenas argues that wheeled and legged robots can serve different stages and tasks. Wheels can make early deployments and data gathering more practical, while legs offer whole-body movement and a compact footprint for varied tasks. Apptronik’s Apollo platform supports both, and he says the industry still needs clear safety standards for legged robots.

  3. 11:34

    What it takes to scale

    He frames commercial success as a chain of connected needs: viable economics, reliable hardware, safety, manufacturing, controls, and AI. A shared platform could reduce costs through scale and attract developers, though he emphasizes that quality and dependable service matter more than simply making robots cheap. Data can come from robots, people, and simulation, with different collection methods suited to different needs.

  4. 17:07

    Applications and trust

    Near-term work includes warehouse tasks such as picking, packout, and kitting. Cardenas sees a path from back-of-house roles to customer-facing settings and eventually homes, with healthcare and elder care as promising areas. He says customers will need confidence in a robot’s safety, reliability, and the company’s ability to support it.

  5. 25:20

    Platform choices and business models

    In response to questions, Cardenas says two arms can handle more tasks than one by letting a robot stabilize an object while manipulating it, often with simpler grippers. He expects bimanual mobile manipulation to remain useful as data and AI improve. On pricing, he anticipates both outright purchases and subscription models, with subscriptions more appealing in logistics and purchases more common in manufacturing.

Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.

More talks