The Most Important Part of Unitree’s IPO Wasn’t the Robot
A 460% debut shows capital flooding into physical AI – and the supporting hardware stack may offer the broader opportunity
Listen to the audio version of this article (generated by AI).
Editor’s note: “The Most Important Part of Unitree’s IPO Wasn’t the Robot” was previously published in July 2026 with the title “The Physical AI Proof Points Are Suddenly Everywhere.” It has since been updated to include the most relevant information available.
When trading began on August 19, Unitree Robotics was a roughly $9 billion company. By the closing bell, investors had turned it into a $50 billion one.
Shares of the Chinese humanoid-robot maker finished their Shanghai debut 460% above the IPO price after climbing as much as 629% during the session.
That is probably too much, too fast.
Many humanoid robots still spend more time performing demonstrations than doing valuable work inside homes and factories. One blockbuster IPO does not settle the commercial case for the entire industry.
But the market’s message is hard to miss.
Physical AI has moved from the demo reel into the public markets.
For the first phase of the AI boom, intelligence mostly lived behind a screen. You typed a prompt. A model answered. Maybe it wrote code, summarized a document, generated an image, or helped draft an email.
Now AI is moving into machines that can see, hear, navigate, and act in the world around us.
AI is getting a body.
And public markets just put a price on it.
Unitree Is Only One Signal in the Physical AI Boom
Unitree’s debut is the loudest signal yet, but it is not happening in a vacuum.
Across the Physical AI market, products are shipping, developer tools are opening up, and companies are building the supply chains required to scale.
- Microsoft’s (MSFT) new Surface AI laptops, powered by Qualcomm’s (QCOM) Snapdragon X2 chips, are now shipping. On-device AI has moved from the roadmap into real products and enterprise budgets.
- Nvidia (NVDA) and Hugging Face are bringing robotics models, teleoperation tools, datasets, and workflows into LeRobot, giving developers a more open toolkit for building Physical AI.
- 1X unveiled a new hand for its NEO humanoid robot that is designed to grip, adjust, and manipulate objects more like a human hand. Useful robots need more than the ability to walk. They need to handle real-world objects reliably.
- Applied Materials (AMAT) and EssilorLuxottica are developing intelligent optical systems for AR and AI-powered eyewear, showing that the smart-glasses supply chain is beginning to industrialize.
- Mobileye (MBLY) is moving from autonomy supplier to robotaxi operator, with a planned U.S. launch in 2027 and a target of roughly 17,000 vehicles over the following five years.
These developments are starting to reinforce one another. Better tools speed up development, while real product launches give suppliers a reason to build for scale.
Physical AI is beginning to move like a real market.
What Physical AI Is – and Why It Needs a Different Hardware Stack
Physical AI changes where intelligence has to run.
Cloud models can draw on vast data centers and send an answer back over the internet. A robot, car, wearable, or factory system has to make many decisions on the spot – often in milliseconds, on a tight power budget, and sometimes without a reliable connection.
That creates a very different hardware problem.
Your headphones need to filter background noise before you notice it.
A warehouse robot has to identify the right box and decide how to grip it.
An autonomous vehicle has to recognize a pedestrian and react immediately.
Those systems need fast local chips, sensors, memory, optics, power management, and connectivity working together.
That difference runs through the entire supply chain.
Six Types of Physical AI Stocks to Watch
Think of Physical AI not as a single industry but as six distinct hardware categories that all need to scale simultaneously.
1. Edge AI Chips: Running Intelligence Locally
This is the foundation. Every physical AI device needs a chip that can run inference locally – fast, cool, and cheap. Qualcomm’s Snapdragon X2 is now shipping inside Microsoft’s Surface lineup. On-device AI has moved from the roadmap into products businesses can actually buy.
Arm‘s (ARM) architectures dominate mobile computing and increasingly sit inside processors designed for edge AI. Nvidia (NVDA) is pushing into embedded inference with its Jetson platform. AMD (AMD) and Intel (INTC) are fighting for their share of the AI PC market. The edge silicon war is just beginning, and the winners here get paid on every device that ships.
Key names: QCOM, ARM, NVDA, AMD, INTC
2. Sensors and Machine Vision: Giving AI Eyes and Ears
Image sensors, depth cameras, radar, lidar, microphones – these are the eyes and ears of every robot, wearable, and autonomous vehicle.
Apple (AAPL) is reportedly exploring camera-equipped AirPods. Though the timeline remains fluid, the larger direction is clear: wearables are gaining the ability to see and interpret the world around us. That creates a new demand cycle for smaller cameras, microphones, depth sensors, and related components.
Key names: Ambarella (AMBA), ON Semiconductor (ON), STMicroelectronics (STM), Sony (SONY), Cognex (CGNX)
3. Advanced Optics: The Interface for AI Glasses
AR glasses and AI eyewear aren’t a consumer curiosity anymore – they’re a hardware category. And the bottleneck? Optics.
Waveguides, photonic displays, specialty glass, and laser projection systems are what separate a pair of glasses from a heads-up display. Corning (GLW) and Coherent (COHR) are two of the most underappreciated Physical AI plays in the market for precisely this reason. Applied Materials’ pivot into intelligent optics manufacturing signals how seriously the semiconductor equipment industry is taking this category.
Key names: AMAT, GLW, Lumentum (LITE), COHR
4. Robotics and Automation: Where Physical AI Does the Work
Unitree proved investors are ready to pay for the humanoid-robot story. The next wave will belong to robots that can earn their keep.
Genesis AI’s Eno is built to adjust to changing tasks and environments. Traditional automation follows a script. Eno is designed to respond when the real world refuses to follow it.
Companies like Symbotic (SYM), Teradyne (TER), Rockwell Automation (ROK), and Honeywell (HON) are already deploying AI-driven automation in factories and warehouses at scale. Tesla‘s (TSLA) Optimus will keep drawing the headlines. Much of the near-term revenue, however, is already coming from less glamorous automation systems working inside factories and distribution centers.
Key names: SYM, TER, ROK, HON, TSLA
5. Memory, Storage, and Power: Feeding Edge AI
On-device AI needs more memory and storage than ordinary electronics. A wearable, robot, or AI PC has to hold models locally, process large streams of sensor data, and deliver sudden bursts of computing power without draining the battery or overheating.
That supports demand for next-generation low-power memory, larger storage systems, power-management chips, and analog components that translate signals from the physical world.
Micron (MU) is already winning here with its LPCAMM modules for AI PCs. The storage plays – Seagate (STX), Western Digital (WDC), SanDisk (SNDK) – get a demand tailwind as every edge device needs local model storage.
Key names: MU, STX, WDC, SNDK, Monolithic Power (MPWR), Analog Devices (ADI), Texas Instruments (TXN).
6. Connectivity: Keeping Physical AI Connected to the Cloud
Even edge AI needs the cloud. Local inference handles the latency-sensitive tasks; cloud AI handles the heavy lifting – model updates, data sync, fleet coordination for robotaxis, telemetry from billions of wearables.
That means the optical networking and connectivity layer is a direct beneficiary of Physical AI scaling. Robotaxis syncing to the cloud. AR glasses streaming map data. Industrial robots phoning home with diagnostic telemetry. Broadcom (AVGO), Marvell (MRVL), Arista (ANET), Ciena (CIEN), Credo (CRDO), and Corning are all toll roads on that data highway.
Key names: AVGO, MRVL, ANET, CRDO, CIEN, GLW
How to Invest in Physical AI Stocks Without Chasing One Robot
Unitree’s debut will make the robot maker the obvious stock people want to chase.
The broader Physical AI opportunity is much larger than any one robot company.
Every humanoid, wearable, AI PC, smart-glasses platform, autonomous vehicle, and industrial machine pulls demand through the same hardware layers: chips, sensors, optics, memory, storage, power management, and connectivity.
The winning device may change. The need for supporting hardware does not.
That creates an enormous opportunity across public markets. It also highlights an increasingly difficult task.
As this boom stretches across edge-chip suppliers, optics companies, robotics specialists, memory manufacturers, power-management firms, and networking stocks, investors have to decide which names deserve capital, how much weight each one should carry, and where different holdings depend on the same underlying trend.
Louis Navellier, Eric Fry, and I have gone back through more than 200 AI recommendations and narrowed the field to roughly 20 stocks we believe deserve capital now, each with a recommended portfolio weight.
Unitree’s debut shows how quickly an exciting AI theme can attract capital – and how easy it is to chase the company making the loudest headline.
See where Louis, Eric, and I are focusing our attention now.
