September 30, 2026
Of the two headlines in front of us today, exactly one touches humanoid robotics in any real way, and even that one is more career profile than news event. The other -- a philosophical essay asking whether 'robotics' and 'automation' mean different things -- is interesting reading for people who write conference abstracts, but it isn't a story about machines with arms and legs doing new things in the world. So rather than stretch a semantics debate into a humanoid trend piece, I want to spend today's column on the one story that actually qualifies, because it says something useful about where this industry's money and attention really are.
Robots Guide published a profile of Charlie Kemp, cofounder and CTO of Hello Robot, tracing his path from AI research at MIT to building Stretch, a mobile manipulator designed to help people with everyday tasks and support independent living. There's no funding round here, no new product launch, no benchmark being smashed. What there is, is a founder explaining -- in his own words, across a career -- that the hard part of humanoid-adjacent robotics was never the walking or the gripping. It was understanding the person the robot is supposed to help.
That's worth sitting with, because it cuts against the dominant humanoid narrative of the last two years, which has been almost entirely about warehouses, factories, and investor demo videos. Kemp's work, and Stretch specifically, has always been aimed at home and care settings -- a much smaller, much harder market, with users who are older, less tech-fluent, and have zero tolerance for a robot that fails at the one task they actually needed done. Businesses watching the humanoid space for automation potential tend to fixate on throughput and ROI in industrial settings. Kemp's career is a reminder that the assistive and eldercare use case is a real, parallel track -- slower to commercialize, but arguably closer to the original promise of 'a robot that helps a person,' rather than 'a robot that replaces a worker.'
For a business reader, the practical takeaway isn't 'go buy a Stretch.' It's that the humanoid and human-assistive robotics fields are still bifurcated between two very different design philosophies -- optimize for the task, or optimize for the person -- and vendors in this space will increasingly need to say which one they're building for. If you're evaluating any automation vendor's roadmap, that question is a useful filter: are they solving for a warehouse's throughput or a person's actual daily friction? The two are not the same engineering problem, and conflating them is how expensive pilots quietly fail. If you're mapping out where automation fits in your own operations, it's the same discipline we'd apply to operations solutions generally -- start with the specific friction point, not the shiny hardware.
Robohub's essay on what 'robotics' and 'automation' actually mean, and whether the distinction matters, is a fair question for the field to keep asking itself -- terminology shapes funding calls, conference tracks, and eventually procurement budgets. But it's not evidence of anything happening with humanoid hardware this week, and I'd rather tell you that plainly than dress up a definitions debate as a trend. If your team is trying to figure out where a given vendor's pitch actually falls on that spectrum, that's less a robotics question than a workflow automation question, and it's worth asking it before the invoice, not after.
So: another quiet day for humanoids specifically, with one genuinely good story about why the assistive side of this field deserves more attention than it gets. Which matters more to your business right now -- a robot that can lift a pallet faster, or one that can help a person live independently a little longer?
Sources