⚫ In memoriam — Joshua Baer, founder of Capital Factory (1975–2026). Read the tribute →
WholeTechPicks|
RobotNewsTodayThe Machine Age · Daily Dispatch

wire active · Austin TX · est. 2024
// Analysis · the robot economy

The Decade of the Robot Has Started — Here's What the Money Is Actually Betting On

By the Robot News Today desk · Austin, TX · analysis · July 12, 2026

Across humanoid robots, factory automation, and the chips underneath both, an enormous amount of money is now chasing the transformation of human work — and it is arriving faster than the results. That gap is the whole story of the robot boom in 2026. Here is how to read it without either dismissing it or overpaying for it.

Wall Street has started sizing a $200 billion market

The clearest sign the robot decade has begun is that serious analysts are now putting serious numbers on it. Barclays projects the humanoid-robot market could grow from roughly $2–3 billion today to about $200 billion by 2035. Wedbush's Dan Ives goes further, calling humanoids the single biggest opportunity in physical AI — machines that do their work in the real world rather than on a screen — and forecasting a market worth trillions over the decade.

The demand is demographic, not hype. Populations are aging across the developed world and fewer people want physically demanding jobs, so the labor shortages already exist with no human solution in sight. Expect humanoids to scale first in manufacturing, logistics, and construction, with healthcare, elder care, and education following after 2030.

China is winning the deployment race

The uncomfortable fact for American investors is that, on the ground, China is ahead. Barclays estimates China accounted for about 85% of humanoid-robot installations last year and builds machines at roughly half the Western cost — on the order of $50,000 each. In industrial robots broadly, China installs about 300,000 a year against roughly 34,000 in the US, and its robot density has grown some 600% since 2016.

ChinaUnited States
EdgeVolume, price, real-world dataFrontier AI, paid deployments
Humanoid installs (2025)~85% of world totalSingle digits of share
Industrial robots / yr~300,000~34,000
Unit cost~$50kPremium / contract-priced
Proof pointNvidia picked Unitree's H2 for its first humanoid platformFigure 03 paid at BMW

Part of the lead is manufacturing muscle and part is data: JD.com is reported to be mobilizing as many as 500,000 workers to collect the movement data that trains robots. Volume compounds — every machine installed generates data, drives down cost, and funds the next generation. For the fuller picture, see our China vs the US scoreboard.

The awkward part: AI can cost more than the workers it replaces

The same wave has a cost problem that rarely makes the headline. Uber reportedly burned through its entire 2026 AI coding budget in four months; Microsoft told some of its own engineers to stop using an AI coding assistant after the bills climbed. An MIT study found automating a task with AI is economically worthwhile in only about 23% of jobs — for the other 77%, humans are still cheaper. And today's prices are not the real prices: OpenAI is reported to spend close to $2 for every $1 it earns serving models, and enterprise AI bills are expected to rise another 30–50% as usage-based pricing bites.

Why this matters for robots: it is exactly why the first durable robot deployments are narrow and boring — a fixed job, a fixed cost per robot-hour, a payback you can calculate — rather than a general-purpose machine that can supposedly do anything. The automation that survives is the automation you can measure.

Michael Burry is shorting the chips underneath it all

Michael Burry — who bet against the 2008 housing bubble — has taken bearish positions against Nvidia, Micron, Applied Materials, and the semiconductor ETF, calling the AI story a “mass addiction.” His point bears directly on robotics, because the humanoid supply chain and the AI-compute supply chain are largely the same chain: the GPUs, memory, and power that train models also give robots their perception and control. Memory names — Micron, SK Hynix, Samsung — have already fallen more than 20% from recent peaks even with strong demand.

The honest caveat is that Burry has been early before, and early feels identical to wrong for a long time. Treat it as a reason to check your own exposure, not a signal to sell everything.

How to read the robot decade

History says two things happen from here and they sound contradictory: the technology wins, and many investors lose. The railroads transformed America and bankrupted a generation of railroad investors; the internet changed everything and still erased trillions in 2000. Transformation and bubbles travel together.

So the grounded posture is to own the trend as a basket rather than a lottery ticket — broad exposure to the supply chain that gets built no matter which robot wins — to cap any single robotics bet at a slice you could lose without changing your life, and to be patient for the full decade the technology will actually take. The people who got rich from the internet were rarely trading it in 1999; they were still holding it in 2015. Watch unit counts and signed contracts, not demo videos.

Recommended reading & kits

Big pictureThe Coming WaveSuleyman on AI & robotics' next decade.Check price on Amazon →AutomationRise of the RobotsMartin Ford on automation and the future of work.Check price on Amazon →US vs ChinaAI SuperpowersKai-Fu Lee on where the race is heading.Check price on Amazon →

Keep reading

Analysis by RobotNewsToday.com, synthesized from reporting compiled by TheFinanceNewsletter.com and the primary sources named inline (Barclays, Wedbush, MIT, company statements). Independent and not affiliated with the companies or publications covered. Some links are affiliate links; as an Amazon Associate we earn from qualifying purchases.

MODE
◐ Theme