Boston Dynamics gives Atlas robot a new four-finger hand

By Splitfeed EditorialLast updated Oct 2, 2026, 04:40 UTC

On October 1 and 2, 2026, Boston Dynamics announced a redesigned hand for its Atlas humanoid robot that replaces the previous five-finger arrangement with a four-finger version by eliminating the pinkie. The decision was not arbitrary: engineers at the company reportedly taped their own pinkie and ring fingers together for a full day to experience firsthand what manipulation tasks became difficult or impossible before committing to the new configuration. The redesigned hand is intended to support more complex industrial tasks as Boston Dynamics continues to position Atlas for factory deployments.

The change matters because it reflects a core tension in humanoid robotics design, how closely a robot's body needs to resemble a human body to be useful in environments built around human workers. Removing the pinkie simplifies the mechanical system, likely reducing points of failure and manufacturing complexity, while the engineers' self-experiment suggests the team was genuinely uncertain about the trade-off and wanted empirical grounding before finalizing the design. Whether four fingers is the right number for the range of grasps and manipulations industrial settings demand remains an open question, and the answer will only emerge once Atlas is working at scale in real facilities.

None of the three camps, Pro-AI, Anti-AI, or Middle Ground, had published reactions to this story at the time of writing. The Pro-AI camp would typically frame a development like this as evidence of rapid, thoughtful iteration in robotics, pointing to the engineering process as a sign of maturing capability. The Anti-AI and robotics-skeptic camp would be expected to question whether humanoid robots can truly substitute for human dexterity in demanding industrial contexts, and might highlight that dropping a finger is an acknowledgment of current limitations. The Middle Ground camp would likely treat this as a reasonable engineering compromise, neither a breakthrough nor a failure, and would call for real-world performance data before drawing conclusions.

The clearest next signal to watch for is deployment reporting from the industrial partners Boston Dynamics is working with, specifically whether the four-finger hand performs reliably across the range of tasks Atlas is assigned and whether any partners push back on the design choice in practice.