MLG Innotek is the company that builds the camera module in your iPhone. Now it's selling eyes to humanoid robots.
CThat's the quiet part of the robot boom—a real supplier with real volume already shifting over.
MGreg on X, who follows humanoid investing, posted the case. LG Innotek already packs high-end optics into tiny high-volume parts for Apple.
CRight. A humanoid needs many cameras and 3D sensors, and it needs them built at scale.
MAnd this is not a render. Greg says LG Innotek already has supply work with Figure AI and Boston Dynamics.
CThose are two of the biggest names in humanoid robots. That's a legitimate customer list.
MFor vision components, yes. The camera modules and sensors that let the machines see.
CI'll say it—the humanoid supply chain is becoming a real business. LG doesn't chase science projects.
MLet's not get carried away. Shipping development kits to Figure is not ten million robot eyes.
CThat's how every hardware ramp starts. Apple didn't ship a hundred million iPhones in year one either.
MOh come on, Apple was selling to consumers. There is no consumer market for a robot that costs more than a car.
CNot yet. But warehouses and factories are already testing them.
MTesting, Chris—pilot programs that can die quietly.
CYou said the same thing about electric cars, and look what happened.
MI said electric cars were late and overpriced. I was right about the timeline.
CFine! But suppliers still built the battery plants first. That's exactly what's happening here.
MSupplier investment can also mean they're hedging, not that the market is real.
CIt's a socket. You need the socket before the product line.
MIt's a press release and two engineering samples. Call me when the factory line runs.
CYou'll be the last one to buy a robot.
MI'll be the one who doesn't get fired for buying a robot that costs more than a car.
CThere it is. The market.
MDifferent corner of the same race: a clip is going around of Musk on why robot hands are so hard.
MRobotics.biz on X says he called hands "shockingly difficult"—precision and force in one motion, touch sensors that survive wear, skin made to be replaced.
CI don't need the exact quote to see the point. The Optimus hand has over a hundred small parts, many still assembled by hand.
MThat's the actual story: the most advanced robot hand still needs human fingers to put it together. That's a bottleneck.
CIt shows why the hand is the hardest part. If they solve that, everything else gets easier.
MOr it shows we're years away from robots building robots.
CYou're moving the goalposts. The claim was hands are difficult, not finished.
MFine, the exact quote is unverified, but the difficulty is not.
MThat's the thing about these clips. The quote gets cut, shared, translated, and suddenly it's gospel.
CThe underlying engineering problem is real though.
MSure. But the hype machine is also real.
CAnd the rumor mill has Boston Dynamics aiming much bigger.
CThe Embodied Era on X points to a Korea Economic Daily report: Atlas suppliers were told to prepare for ten thousand robots by the end of 2027.
MUnnamed industry sources. So take it for what it is.
CEven half that number would be a massive leap from where Atlas is today.
MAtlas does backflips in a lab. That's not ten thousand units in a factory.
CHyundai owns Boston Dynamics. They have real assembly lines. If anyone can scale it, they can.
MMaybe. But "reportedly" and "could" are doing a lot of work in that sentence.
CAll I'm saying is the robot supply chain is no longer a science fair.
MNo, it's a supply chain audition. And the buyers are still kicking the tires.
CKicking tires is a step, Matilda. Last year they were just drawings.
MFair. Drawings to tires is something.
CAnd tires to ten thousand units—that's the jump we're debating.
MAnd I'll believe it when the first invoice ships.