Digit 5 can't run — Agility pins a design trade-off to the work-robot pivot
From the Robots digest for 2026-10-19
What happened: Agility Robotics on X, October 19 (~4:54 PM ET): "Our co-founder's earlier robots could run. Digit 5 can't" — with the post pointing to cofounder Jonathan Hurst's IEEE Spectrum essay for the reasoning behind the redesign. The tweet turns the cofounder essay published the same day at 9:00 AM ET into a stated hardware decision: the running ability of Agility's lab-era lineage is deliberately absent from the machine it intends to sell.
What's genuinely new: A development of the Digit 5 arc within the reality-check war — the essay item earlier in the day carried the launch claims (23-kilogram bin lift, 90-minute run on a 9-minute charge, "tens of thousands of humanoid workers") but not this explicit capability removal. The trade-off lands at the heart of the war's core argument: the motion repertoire that draws viral spectacle is excluded from the commercial product in favor of the work posture, echoing the Google DeepMind calibration already on record from October 3 — locomotion as a "low-value sideshow," reliability the binding constraint. The design logic remains vendor-framed until the essay's full reasoning is in hand.
Watch: Hurst's full reasoning in the essay, which the tweet capture truncates; whether "can't run" becomes a deliberate marketing point ahead of the November 2 SPAC gate; and whether independent testing ever rates Digit 5's work performance against the running robots of the demo era.
Links: Agility Robotics on X · IEEE Spectrum