Mark Cuban: The Humanoid Robot Wave Will Fail in Five to Ten Years
Investor Mark Cuban stated that the humanoid robot wave will fail within five to ten years, with homes redesigned for specialized task robots rather than humanoid forms.
He mentioned on the live show TBPN that everyone is pushing for humanoid machines, which may only last five years, at most ten, before they fail. Homes do not have to be designed for humanoid forms: storage, refrigerators, and laundry can be hidden in the back, with transport robots in the form of spiders or ants interfacing, leaving living space for humans. "You design houses for robots, and you design robots for houses." Amazon warehouses have already taken this path, where box-moving is not done by humanoids.
He later added that there are technical shortcomings: latency, computing power, dexterity, and the need for continuous upgrades. Humanoid machines may temporarily be a novelty for fetching beer and folding clothes, but they are not optimal for most tasks. Specialized robots are faster, cheaper, and more efficient. He predicts that within twenty-five years, new buildings will be designed for application-specific robots, with logistics already paving the way.
Benchmark targets include Tesla's Optimus and financing narratives from Figure, Agility, and others. Cuban does not deny human-robot coexistence; what he denies is the notion that "houses are made for humans, so robots must look like humans." He claims that the assumption of humanoids replacing humans "has no chance."
Specialized machines rewrite the environment as a constraint, while humanoid machines treat the environment as an unchangeable legacy. Battery work hours, grasping, and gait remain public shortcomings. Valuations are still based on the replacement of general labor.
In market mechanisms, this is a narrative collision, not an order cancellation. Buyers are logistics capital needing warehouse throughput and humanoid startups seeking valuation stories; sellers are robot suppliers of different forms. Currently, funds still flow towards humanoid forms that can be showcased, while specialized machines benefit from verified warehousing returns. Beneficiaries include real estate and logistics automation that can modify building codes, while projects that define legs as necessary functions are under pressure. A prediction does not change production line scheduling.
Source: Public Information
ABAB AI Insight
Cuban extrapolates what has happened in warehouses to residential settings: first modify shelves, then modify vehicles, never first create a humanoid that can walk on shelves. The financing logic of humanoid machines is "the world is built for humans, so a general body is the most convenient." He reverses this: once the reconstruction cost is lower than mimicking the human body, the optimal form will no longer be human. Spiders and ants are metaphors for multi-point contact and narrow operations, not actual insect pets.
Capital pathways are diverging. Humanoid projects sell platform options, a round body that can handle chores, factories, and care. Specialized machines sell single-point investment returns: one transport system per floor. Amazon has already proven that throughput can be achieved without legs using non-humanoid machines. If real estate begins to reserve pipelines for back-end tracks, the humanoid machine's advantage of "can be used without modifying houses" will disappear.
Analogous to personal computers versus mainframes, and electric scooters versus humanoid vehicles: general forms first win imagination, while specialized forms win work hours. The industry phase is concept validation versus infrastructure rewriting. Tesla has framed Optimus as a business larger than cars, while Cuban has framed the same time window as the lifespan of concept stocks.
Structural judgment belongs to technological substitution. What is being replaced is not labor itself, but the assumption that "must look like a human to fit into existing spaces." The mechanism is: when the environment can be modified, forms follow tasks; when the environment is treated as unchangeable, investment will flow towards legs and fingers. Five to ten years is the window he leaves for concept premiums, not for motor failures.