Outrageous Predictions
Carry trade unwind brings USD/JPY to 100 and Japan’s next asset bubble
Charu Chanana
Chief Investment Strategist
Unitree’s blockbuster IPO is putting humanoid robots back in the spotlight. Its IPO was more than 8,000 times oversubscribed by retail investors, highlighting just how much enthusiasm has built around the theme.
Its humanoid robots first reached a mass audience after performing a synchronised dance at China's widely watched Spring Festival Gala, while subsequent videos have shown Unitree machines running, performing backflips and martial-arts moves.
Just before its stock-market debut, the company added to the spectacle by unveiling its new “SuperMan” humanoid, designed to run and jump at speeds and heights beyond typical human capability.
But investors do not necessarily need to own the robot maker itself. The bigger opportunity may span the entire robotics value chain — the chips, sensors, motors and automation systems needed to give AI a body.
The excitement is considerable, but commercialisation remains the big test. Many humanoids can perform impressive demonstrations, but the industry still needs to prove that robots can complete economically useful tasks reliably and with limited human supervision.
Other risks include valuations running ahead of adoption, intense Chinese competition and price pressure, rapid technological obsolescence, cyclical factory capex and geopolitical/export restrictions. Robotics ETFs also vary considerably in what they actually own, so a “robotics” label does not necessarily mean pure humanoid exposure.
Nvidia is already at the centre of generative AI, but its ambitions increasingly extend into Physical AI.
Its Jetson computing platform, Isaac robotics software and GR00T foundation models are designed to help developers train and deploy intelligent machines.
The attraction is simple: investors do not have to predict which robot manufacturer wins if Nvidia can supply the computing architecture used across the industry.
Risk: Nvidia is far from a pure robotics play, and expectations around its wider AI franchise remain extremely high.
Robots need to understand the environment around them before they can move safely through it.
Hong Kong-listed RoboSense develops LiDAR and perception systems that allow machines to map and interpret their surroundings. The company is increasingly expanding beyond autonomous vehicles into robotics and Physical AI.
That makes RoboSense an interesting picks-and-shovels exposure to the proliferation of intelligent machines.
Risk: Chinese sensor markets are highly competitive, and rapidly falling component prices could pressure margins even as volumes increase.
AI may provide the intelligence, but a robot still needs to move.
Japan's Yaskawa is one of the world's major producers of servo motors, motion-control systems and industrial robots. These technologies allow robotic arms and machines to make extremely precise movements.
If humanoid and industrial robot production scales, demand for sophisticated motion systems could grow with it.
Risk: Yaskawa remains exposed to the global manufacturing and factory-investment cycle.
FANUC is one of the best-known names in industrial automation, with robots already operating across factories globally.
The next opportunity is integrating more intelligence into those machines. FANUC has been working with major AI technology providers as manufacturers seek robots that can adapt and learn rather than simply repeat pre-programmed tasks.
It offers exposure to the transition from traditional automation toward intelligent automation.
Risk: Factory automation can be cyclical, particularly when manufacturing capex slows.
Most investors know Teradyne for semiconductor testing equipment. But it also owns Universal Robots and Mobile Industrial Robots.
Universal Robots is a leading player in collaborative robots, or cobots: smaller robots designed to work alongside humans rather than inside fenced-off industrial areas.
That could make Teradyne an interesting bridge between today's automation industry and tomorrow's more flexible robotic workforce.
Risk: robotics still represents only part of Teradyne's overall business, while semiconductor testing remains an important earnings driver.
For investors attracted specifically to the Unitree story, Hong Kong-listed UBTECH is one of the more direct listed humanoid exposures.
Its Walker series of robots is being developed for industrial use cases including manufacturing, logistics, inspection and material handling.
That gives investors relatively direct exposure to the question at the centre of the Physical AI boom: can humanoids become economically useful workers?
Risk: this is also one of the highest-risk exposures on the list. Humanoid deployment remains early, profitability is uncertain and competition in China is intense.
Tesla increasingly describes itself as an AI and robotics company rather than simply an electric-vehicle manufacturer.
Its Optimus humanoid project aims to leverage many of the technologies Tesla has developed for autonomous driving—computer vision, AI models, batteries and manufacturing.
If Tesla can eventually manufacture robots at scale, Optimus could create a significant new addressable market.
Risk: commercial expectations for Optimus remain well ahead of proven revenues today, while Tesla's investment case continues to depend heavily on its automotive business.
Japan's Keyence makes sensors, machine-vision systems, measuring equipment and automation technologies used across manufacturing.
As factories become increasingly automated, machines need more sensors and vision systems to detect objects, measure distances and inspect products.
Keyence therefore provides a less glamorous—but potentially important—way to participate in the infrastructure behind increasingly autonomous factories.
Risk: premium valuations and exposure to industrial capex can amplify downside during manufacturing slowdowns.
ABB sits at the intersection of robotics, electrification and factory automation.
Its robotic arms and automation systems are already used across manufacturing, while its broader industrial presence gives it exposure to companies seeking to automate production and improve efficiency.
ABB therefore offers a more diversified route into the robotics theme than the emerging humanoid players.
Risk: diversification also means robotics is only one part of the investment case, so successful humanoid adoption would not necessarily translate directly into earnings.
Rockwell is less about building the robot and more about building the environment in which increasingly intelligent machines operate.
Its factory-automation software, control systems and industrial technologies help manufacturers connect machines, production lines and data.
If Physical AI ultimately transforms industrial production, the opportunity could extend well beyond individual robots toward entire factories becoming more autonomous.
Risk: Rockwell remains highly exposed to corporate manufacturing investment, which can weaken during periods of economic uncertainty.
Investors who do not want to choose the eventual robotics winners can also consider diversified ETFs.
Global X Robotics & Artificial Intelligence ETF (BOTZ) provides exposure across robotics, automation and AI companies. It can be a simpler route for investors looking for diversified thematic exposure.
ROBO Global Robotics & Automation ETF (ROBO) spreads exposure across the wider robotics ecosystem, including components, industrial automation and robotics companies.
For investors looking for a UCITS structure, the iShares Automation & Robotics UCITS ETF (RBOT) provides another diversified route into global automation and robotics.