Skild AI went from $0 to $30M in revenue in 2025 building a universal robot brain.

That's not just fast growth. It means companies are already paying to avoid building custom automation.

I think the build-vs-buy window just closed 🧵

SAP just ran humanoids in a customer warehouse connected directly to their existing management software.

No custom middleware. No integration project.

I think the enterprise adoption timeline is about to collapse 🧵

PNDbotics claims their humanoid learns new tasks in hours, not months.

They ship with 10,000+ scenarios pre-loaded.

But the real question isn't whether it works in their lab—it's whether "plug and play" humanoids actually exist yet 🧵

1X is shipping NEO humanoids to homes at $20,000 this year.

But here's what the promo videos don't show: humans in VR headsets are still controlling these robots when they get stuck.

This isn't full autonomy. It's expensive teleoperation 🧵

Agibot just claimed the #1 spot in global humanoid shipments.

Not prototypes. Not demos. Actual units shipped and deployed.

I think this is the moment China's "ship fast" strategy lapped Western "perfect then ship" thinking 🧵

LEM Surgical is performing commercial spinal surgeries in Las Vegas with a three-armed robot system.

It's already FDA-cleared. Already generating patient outcomes.

But what caught my attention isn't the tech—it's the timing 🧵

Mobileye just spent $900M to buy a humanoid startup.

But the interesting part isn't the price tag—it's the timeline.

They're committing to commercial deployments in 2026 and series production by 2028.

That's not R&D. That's a product roadmap 🧵

Boston Dynamics just gave Atlas a Google DeepMind brain.

But the real story from CES 2026: Hyundai just committed to manufacturing 30,000 units annually by 2028.

That's not a pilot. That's a production bet. 🧵

Qualcomm just entered the humanoid chip race.

But I think the story isn't the chip—it's the timing.

When smartphone giants pivot serious resources toward robotics, it signals something changed. Here's what I see 🧵

A robot just learned 1,000 tasks in 24 hours from one demo each.

But the real story isn't the speed—it's that this flips the economics of flexible automation.

Here's why that matters 🧵

Midea just showed a 6-armed wheeled robot and called it a "super humanoid."

It's not humanoid at all—and that's exactly the point.

China's 2026 strategy isn't about making robots look human. It's about winning on price and function 🧵

A surgeon in Shanghai just performed cancer surgery on patients in Mumbai—5,000km away.

This wasn't a demo. It was actual patient care with regulatory approval.

But I think the real story is what didn't happen 🧵

A robot that feels pain and pulls away reflexively—without waiting for its brain to process the signal.

City University of Hong Kong just published this, and I think it matters more than the headlines suggest.

Here's why 🧵

Boston Dynamics is targeting brownfield factories with electric Atlas.

Not new factories. Not labs. Existing plants that weren't designed for robots.

I think this is the smartest humanoid strategy I've seen. Here's why 🧵

Everyone's talking about humanoid robots entering factories.

But the bigger shift is factories being redesigned for robots—eliminating lights, climate control, and 30-40% of floor space built for humans.

I think we're underestimating how fast this happens 🧵

China issued 7,705 humanoid patents in five years.

The US? 1,561.

But the real story isn't the 5-to-1 ratio—it's what it does to your unit economics if you're building humanoids outside China.

🧵

CATL just deployed humanoid robots at scale on battery production lines.

99% success rate on high-voltage testing. Not a demo. Not a pilot.

This is the deployment everyone's been waiting for—but I think the tech choices matter more than the milestone 🧵