Mark Cuban Says Humanoid Robots Fail in 10 Years. He's Probably Right.
Mark Cuban predicts the humanoid robot category collapses within a decade, arguing homes will get redesigned around task-built machines instead. The venture math says he's fighting a $38 billion narrative. The engineering math says he wins.
Will there be a humanoid robot folding your laundry by 2035? Mark Cuban says no. The billionaire investor thinks the entire humanoid category fails within five to 10 years, and his reasoning is less contrarian than it sounds.
He's not betting against robots. He's betting against the shape.
The Numbers Don't Support the Shape
Here's what matters: the money is enormous. Goldman Sachs Research pegs the humanoid robot market at $38 billion by 2035, up from roughly $6 billion today. Figure AI raised capital at a $39 billion valuation. Tesla's Optimus owns the headlines. Unitree sells a G1 for about $16,000, which sounds cheap until you watch it struggle with a load of laundry.
Cuban's point is simpler than the pitch decks suggest. Every dollar is chasing a form factor, not a function.
The numbers tell the story. A robot with two arms, two legs, and a head has roughly 40 joints to coordinate. Most of them spend their day fighting gravity. A dishwasher has zero joints, costs $700, never falls over, and does one job perfectly for 15 years.
Which one would you actually buy?
Form Follows Function
The home appliance market already answered this question. iRobot has sold more than 40 million Roomba units since 2002, and not one of them looks like a person. Washing machines, microwaves, and dishwashers all won by being ugly and single-minded.
So the real shift Cuban is describing isn't robotic at all. It's architectural. Kitchens and laundry rooms get rebuilt around machines that only do one thing, and the machines never need to climb stairs or open a door.
That's a much smaller technical problem. It's also a much smaller story to sell to a venture fund.
Why the Smart Money Disagrees
Ask a robotics founder and they'll tell you a general-purpose robot is a $1 trillion market. A robot that only does dishes is a $5 billion company. That gap is the entire reason humanoids keep getting funded.
According to people who write these checks, the bet is on software catching up to hardware. Vision models and reinforcement learning have improved fast. But hands are still the bottleneck. Nobody has shipped a cost-effective robotic hand with enough dexterity to handle a wine glass and a sock in the same afternoon.
From a risk perspective, that's the trade. You're paying a general-purpose valuation for a task-specific reality, and you're paying it five years early.
What to Watch
Four things will settle this argument. Tesla's Optimus production guidance for 2026. Figure's deployment scale at BMW's Spartanburg plant. Whether Chinese makers push humanoid pricing under $10,000. And whether any humanoid ships more than 100,000 units to real paying customers.
But watch the appliance makers too. Samsung's Bot Handy and LG's Q9 are quiet, unglamorous, and aimed at the exact jobs Cuban says will win. If one of them ships a $2,000 kitchen machine that loads the dishwasher on its own, humanoid valuations get repriced in a week.
Frankly, that's the more interesting bet. And it doesn't need a robot that looks like you.
