Humanoid robots aren't science fiction. They're real, and they're coming to a home or business near you, whether or not you've cleared a space. There's a handful of serious players, and the list keeps growing. Tesla says Optimus is in "early production." Agility Robotics is a week away from a November 2 vote to go public at a $2.5 billion equity value.

Let's dive into what they are, how they work and what you can expect next.

Hashi's take: Humanoids are overhyped in both directions. They aren't taking every job next year, and they aren't a party trick either. It's inevitable they'll be part of our lives, and soon. So what we should be talking about is who controls and governs them, and how we keep people safe around them.

Sales signal real demand. Some folks aren't waiting around. They're hedging their bets while everyone else argues about it on LinkedIn.

Forward this to the colleague who still thinks humanoids are a Netflix plot. And to the one who's already budgeted for ten. Share your personal link and earn rewards when they join.

A few things landed at once:

A robot built roughly like you: torso, two arms, hands, a head stuffed with cameras, usually two legs. No butler. No Terminator. The arm bolted to your line does one job until someone unbolts it. A humanoid is meant to walk over to the job, and do a different one tomorrow.

Why the human shape? The world was built for people: stairs, doors, shelves, totes at waist height. That's the pitch, and the catch. If a wheeled cart can do the job, it's cheaper. Legs look lovely on a keynote stage. On a flat warehouse floor, they're optional.

The part of the robot ad that got cut: when a robot gets stuck, a human often steers it remotely. 1X sells a home humanoid called NEO for $20,000, and its own order page says that for tasks NEO doesn't know, "an Expert from 1X can remotely supervise." Your robot butler comes with a butler.

Tesla gets its own paragraph, because Elon would insist. Musk says Optimus starts on "dangerous repetitive tasks, things that people really don't want to do," around the clock, and reaches homes once production scales. The hands still need work. Musk deadlines tend to arrive fashionably late. Use pencil.

Physical AI is where the paradigm shifts next. Once humanoids are commercially available, they'll be part of our everyday lives, which means learning to coexist with them, safely.

Expect the call: can a humanoid replace a line worker? Agility's illustrative model is $8,500 a month (about $102,000 a year) plus roughly $25,000 to deploy. The average US factory production worker earns $30.21 an hour, about $66,000 a year. The vendor will still have plenty of reasons for you to say yes.

So where do we draw the line? Who decides how it works around people, who can shut it down, and who governs it after the demo team leaves? We're still building governance for chatbots and agents. Physical AI needs it too, ideally before the robot bumps into someone. Or picks a fight.

We spend a fortune making robots clever and pennies deciding where they're allowed to stand.

In August I wrote about factory equipment learning to talk to AI. Now the machines are getting legs and losing the fence.

I've said it before: AI doesn't fix every business problem. Put it on top of a broken process and it amplifies the mess, quickly. Most plants, especially smaller ones, aren't ready for humanoids yet. Labor savings will matter. But when everyone has humanoids, the company with the broken process still loses. A robot doing the wrong job faster is just a more expensive mistake.

Fix the process before you buy the robot. And ignore the somersaults.

Keep reading!

- Hashi