Surprisingly, despite Elon Musk's involvement, this may turn out to be a scam.

Like the Metaverse, like Musk’s humanoid robots and brain implants and eugenics and Mars colonization, you realize these guys aren’t creative thinkers so much as they’re mega-dorks who want to make their favorite sci-fi real. The creative ethos of Ready Player One, with a trillion dollars behind it

— Will Stancil (@whstancil.bsky.social) May 1, 2025 at 7:59 AM

 

Folks, we've been keeping an eye on hype and bubbles for a long time here at the blog, and we've accumulated a pretty good track record. So take it from me: we are looking at a metaverse-level event with humanoid robots. When the dust settles, hundreds of billions of dollars will have been wasted on an embarrassing techno-optimist hash of bad engineering, flawed economic reasoning, and hackneyed sci-fi tropes.

I just went through the Morgan Stanley rainbows and unicorns report on our robot future. I'll have more to say about that in the coming days (spoiler alert: it's bad). For now, here are a couple of threads from an actual engineer who does a good job laying out the absurdity of it all.

UPS should give me a call. I can guarantee them that I can design a non-humanoid robot to do this that will be substantially lower in total lifecycle costs. Another feature of these humanoid robotics companies? The “exact functions” always remain “unclear”. That’s the tell in the robotics space.

[image or embed]

— Adam Cook (@motorcityadam.com) April 29, 2025 at 8:42 AM

Again, recognize that one is adding a TON of complexity when one reaches for a humanoid robot. And that means that complexity has to be justified. So you better damn be sure that you have a process design intent and is tight as a drum. And seldom do any of these companies offer one.

— Adam Cook (@motorcityadam.com) April 29, 2025 at 8:48 AM

🧵 Yes, sort of lol. Humanoid robotics are *still* very much confined to the lab in many cases. And manufacturing use cases are the most ill-suited for humanoid robotics in general. One would REALLY have to justify the added complexity there - since the physical process can be changed.

[image or embed]

— Adam Cook (@motorcityadam.com) May 1, 2025 at 6:18 AM

And that’s the thing with manufacturing that is lost on many that are not familiar with that work. Industrial automation is a SYSTEMS problem - we don’t just throw catalog robots at an application. The product can be and is designed to mesh with the automation - creating a highly-optimized SYSTEM.

— Adam Cook (@motorcityadam.com) May 1, 2025 at 6:18 AM

Perhaps. But Agility has a well-defined design intent and a relatively simplistic robot - key differentiators to many other humanoid robots firms that cannot (or refuse to) offer a coherent design intent. The complexity reduction counts for a lot when talking lifecycle costs.

[image or embed]

— Adam Cook (@motorcityadam.com) May 1, 2025 at 6:18 AM

At least Wired did not use the Adam Jonas (Morgan Stanley) report. The key problem with these projections is that, especially for manufacturing, you are fighting against traditional automation that we have decades of experience with and, in many ways, has been commoditized in terms of upfront cost.

[image or embed]

— Adam Cook (@motorcityadam.com) May 1, 2025 at 6:18 AM

Any AI or electromechanical advancements made by the humanoid robotics realm is immediately available for use with EXISTING traditional robotics of far less complexity and risk - which further provides a competitive barrier for humanoid robotics companies.

— Adam Cook (@motorcityadam.com) May 1, 2025 at 6:18 AM