Generalist uses off-the-shelf arms

This shows Generalist is trying to win robotics the way Android won phones, by being the intelligence that can run on many bodies instead of the company that has to design every motor, joint, and sensor itself. Buying off the shelf arms lets it move faster across customer workflows, because engineers spend time on perception, control, and task adaptation rather than on building a new manipulator platform and manufacturing stack for each deployment.

• This hardware light approach fits how Generalist sells. It scopes one workflow, recreates it with the customer’s own parts, fine tunes the base model, and proves repeatability before production. That process benefits more from fast software iteration across many arm types than from owning robot arm design.
  1 sacra
• The tradeoff is less hardware control than vertically integrated rivals. Figure’s Helix 02 controls walking, manipulation, and balance as one system, and Figure highlights palm cameras and tactile sensors built into Figure 03. 1X says it designs and manufactures critical NEO components in house. Those companies can tune the robot body around the model.
  2 figure 3 1x
• Generalist is closer to the software brain layer that plugs into existing industrial hardware. That matters in a market where installed robot bases already exist. Related research shows companies like Skild are pursuing OEM channels through ABB and Universal Robots, and Scale has integrated training tooling into Universal Robots’ installed base of more than 100,000 deployments.
  1 sacra 4 sacra 5 sacra

Going forward, the prize is becoming the default intelligence layer for commercial robot arms and future embodiments. If Generalist can keep proving high success rates on new workflows without requiring custom hardware, it can spread faster through OEM partnerships and deployments, while vertically integrated rivals push the frontier on performance through tighter hardware and software co design.