If I told you that in 2026, as AI finally steps into the physical world, only by understanding these four companies can you stay ahead in future tech investing—would you believe it?
If you still think the AI main theme in 2026 is just those boring chatbots, then I have to say: buddy, you're out of date!
Wall Street is now really focused on Physical AI—AI that can walk into factories, drive on roads, enter operating rooms, and live inside homes. It is not just talking anymore; it is actually acting. This may be a wealth shift ten times bigger than large language models.
Over the past three years, we watched large language models go from toys to productivity tools, but they were still trapped on screens. Starting in 2026, AI is finally stepping out of the screen and into the real world.
It will drive, deliver food, perform surgery, and manage entire factories... which means AI is no longer just saving your brain power; it is replacing human hands.
Why does this matter so much for ordinary investors?
Because it will directly reshape tens of trillions of dollars in the real economy: manufacturing, logistics, healthcare, warehousing, food service... every industry with high labor costs and repetitive work is heading into an epic efficiency revolution.
But opportunity always comes with risk.
Many companies tell beautiful stories and then die burning cash; others look ordinary but quietly eat the juiciest part of the entire supply chain.
Today I won't waste your time—I’ll break down four real companies directly:
-
The first represents explosive upside, is the most controversial, and has the wildest beta;
-
The second represents profitability and is already the closest thing to a money printer in robotics;
-
The third provides the eyes of robots and is often overlooked, but impossible to avoid;
-
The fourth is the nervous system of the smart factory—the classic "shovel seller."
After this episode, you'll not only understand how big the Physical AI theme really is, but also clearly see which companies are just pitching concepts and which are truly bringing AI into the real world. Most importantly, you'll get a clear framework for avoiding traps and structuring exposure—so you won't miss the upside, but you also won't blindly chase highs and become exit liquidity.
Want to know which company could become the next ten-bagger? Which one can truly survive cycles?
Keep listening—I’ll break down the logic, earnings, risks, and opportunities for you in detail.
Invest smart, avoid the brick wall, and let wealth speak for itself. Hello everyone, welcome to the show.
If the hottest AI story over the past three years was large language models, then starting in 2026, the next main theme Wall Street is watching is no longer the tiny chatbot on your screen that only plays question-and-answer games. It is Physical AI—the kind that can actually walk into factories, drive on roads, enter hospitals, warehouses, restaurants, and homes.
Why is this so important? Because the AI of the past was more like a super-smart brain. It could answer questions, write code, generate images, and help organize materials, but most of the time it still stayed on the screen and its ability to change the real world was limited.
Physical AI is completely different. It does not just speak; it starts to act. It can see the environment, understand tasks, make judgments, and then actually execute—driving, moving goods, delivering meals, inspecting, performing surgery, managing warehouses, and eventually entering manufacturing and home services.
That means AI is moving from the digital world into the physical world. In the past we talked about compute, models, and data centers. In the future, we may talk more about robots, sensors, machine vision, automation factories, and the data produced by every motion in the physical world.
That is also why NVIDIA CEO Jensen Huang keeps emphasizing Physical AI. Especially at the end of May 2026 at GTC Taipei, NVIDIA unveiled a humanoid-robot reference design based on the Isaac GR00T platform, integrating the robot body, dexterous hands, the Jetson Thor edge-computing platform, simulation training, robot models, and deployment tools into a full stack.
The signal is very clear: NVIDIA does not just want to keep selling AI chips—it also wants to push AI from the data center into the real world.
You can think of it like this: in the past, AI was responsible for thinking; in the future, it must also be responsible for acting. In the past it was a brain; in the future it must grow eyes, hands, legs, and a complete nervous system.
From an investment perspective, the imagination here is huge. Manufacturing, logistics, healthcare, agriculture, construction, warehousing, and restaurant automation together are not a small niche—they are tens of trillions of dollars of real-economy market.
More importantly, these industries all share the same traits: high labor costs, repetitive workflows, and lots of room for efficiency gains. So what Physical AI is really replacing is not necessarily the human brain—it is human hands.
Of course, don't get too excited and rush in. At the start of every new technology wave, markets get extremely excited, but the companies that survive to the end are usually not the loudest storytellers—they are the ones with the most viable business model, the strongest financials, and the hardest-to-replace position in the chain.
Today we are going to break down four companies around this Physical AI theme. The first represents explosive upside, the second profitability, the third machine vision—the eyes of robots—and the fourth industrial automation infrastructure, the nervous system of the future smart factory.
If you want to understand how Physical AI becomes an industrial-chain opportunity step by step instead of just hype, this episode is a must-listen.
Company 1: Serve Robotics (SERV) — explosive upside with controversy
Let's start with the first company—and the most controversial one right now—Serve Robotics, ticker SERV.
This company makes autonomous delivery robots. If you've been on U.S. college campuses, business districts, or neighborhoods and seen those small driverless delivery carts slowly rolling by the curb, delivering takeout, coffee, or groceries, that's probably their product.
A lot of people may think first: is this really reliable? I saw videos online of robots bumping walls, getting lost, getting stuck by the road, and even being watched by curious crowds—it looks kind of funny.
But capital markets are interesting: many things that end up changing the world look clumsy, slow, or even ridiculous when they first appear. Early EVs looked like that. Early ChatGPT was dismissed by some as a toy. Early autonomous driving was heavily mocked too.
What matters is not whether it looks clumsy today. What matters is whether it is accumulating real-world data, iterating continuously, and moving from the lab into commercialization. The hardest part of robotics is not how pretty it looks in a demo video; it is handling the messiness of the real street—pedestrians suddenly crossing, road construction, pets running out, fading light, rain, slopes, obstacles, unstable signals... every real-world scenario becomes valuable training data.
Whoever gets robots into the real world first gets the real-world data first; whoever has more data may evolve faster. That is one reason I started paying serious attention to SERV.
A lot of people think SERV's biggest value is the robot itself, but I care more about the real-world operating data it is accumulating and the ability to operate robots in actual environments.
Of course, a story alone is not enough. The market ultimately wants the numbers. Serve Robotics' first-quarter 2026 results at least show it is moving from concept toward commercialization. Q1 revenue was about $3 million, up 238% sequentially and 578% year over year. The absolute number is still small, but the growth rate is astonishing.
Even more important is the revenue mix: software services already contributed about one-third of revenue, and nearly half of total revenue has recurring characteristics. In simple terms, it is no longer just selling one robot once and making money once; it is moving toward robot services, software services, and recurring operations revenue. The valuation logic for those models is completely different.
As of Q1, the company had deployed about 2,000 robots across 44 U.S. cities in 14 states, with an average of 812 robots active per day and over 10,000 service hours delivered daily. It also acquired Diligent Robotics, expanding from outdoor delivery into indoor hospital service robots.
That is the most exciting part of SERV: it is not just a little delivery cart company; it is trying to extend the same automation capabilities into more real-world scenarios.
But the risks also need to be clear. SERV is still deeply unprofitable. Q1 net loss was about $49 million, full-year revenue is expected to be about $26 million, while non-GAAP operating expense guidance is as high as $160 million to $170 million. It is still a classic high-investment, high-burn, high-volatility company.
So in my view, investing in SERV is not buying stable cash flow; it is buying a possibility. If autonomous delivery, hospital service robots, and multi-scenario operations really work, it could become an early high-beta Physical AI name. But if the business model cannot improve efficiency and revenue cannot keep up with burn, it may keep losing money for many years.
That is SERV's essence: high risk, high imagination, high volatility. If you're a conservative investor, it may not deserve a big position. If you are willing to track a small position for a new-tech inflection point, you can put it on the watchlist—but you must keep size under control and not buy it as if it were already mature.
Company 2: Intuitive Surgical (ISRG) — the strongest profit machine
Next, let's talk about the second company. It is the opposite of SERV. SERV is a bet on the future; this company is already making money steadily. It is Intuitive Surgical, ticker ISRG.
When many people hear robotics, they first think of factory arms or Tesla's humanoid robot. But in reality, one of the most profitable and moat-heavy robotics use cases has never been the factory—it is the operating room.
Why? Because healthcare has two special traits: customers can pay, and switching costs are extremely high. Once a hospital buys a core surgical robot system, trains doctors, builds workflows, and ties together consumables and service, it is hard to switch.
Intuitive Surgical's core product is the da Vinci surgical robot. In Q1 2026, global da Vinci procedures grew about 16% year over year, and Ion lung robotic procedures grew about 39%. The company installed 431 da Vinci systems in the quarter, of which 232 were the latest da Vinci 5 systems, along with 52 Ion systems.
What I value most is not how many machines it sold, but the business model behind them. Intuitive Surgical is increasingly becoming a subscription-like medical robotics platform. Q1 revenue reached $2.77 billion, up 23% year over year, with instruments and accessories revenue at $1.69 billion, also up 23%. Every surgery consumes related instruments and disposable supplies; the more procedures there are, the more stable the consumables and service revenue becomes.
That creates a powerful flywheel: the more installed base, the more skilled doctors become, the more surgeries are performed, the more stable consumables and service revenue becomes, the more money the company has to develop the next generation, and the next generation further increases doctors' willingness to use it.
Compared with SERV, ISRG's advantage is obvious: it has proven the model, it has revenue, profit, installed base, consumables revenue, a doctor-training ecosystem, and long-term stickiness. It may not be the most exciting name, but it is one of the closest things to a stable money machine in the Physical AI theme.
If you find this kind of company-by-company, chain-by-chain analysis useful rather than just staring at stock price swings, please like, save, and subscribe. I'll keep tracking the latest moves in Physical AI, AI chips, robotics, and automation themes.
Company 3: Cognex (CGNX) — the eyes of robots
Now let's talk about the third company. If a robot has a brain and limbs, what is the next most important thing? It is being able to see the world.
No matter how smart a robot is, if it cannot see its environment, it is just a pile of expensive metal and chips.Machine vision is one of the most critical links in bringing Physical AI to life.
Cognex, ticker CGNX, sells the eyes of robots. Its position in the Physical AI chain is crucial. In Q1 2026, revenue was about $268 million, up 24% year over year, while adjusted EBITDA and EPS both grew strongly across multiple regions and industries.
Cognex is also pushing hard into AI vision products, launching the NVIDIA-powered In-Sight 6900 and the Qualcomm-powered In-Sight 3900 embedded AI vision system, targeting edge AI vision so machines can recognize, judge, and act directly on site.
It may not be the sexiest company, but it may be one of the deeper-moat links in the supply chain. No matter who builds the robot in the future, if it enters the real world, it must first see, identify, and judge things.
Company 4: Rockwell Automation (ROK) — the nervous system of the smart factory
Finally, let's talk about the fourth company, Rockwell Automation, ticker ROK.
This may be the least flashy company among the four today, but it is also one of my favorite types of business. It is more like a shovel seller in the Physical AI era. If robots really enter factories on a large scale, the hardest problem is not whether a single robot can move—it is how thousands of devices connect, coordinate, run safely, and make the entire factory operate like a living system.
Rockwell Automation is the nervous system of the smart factory. It connects equipment, controls processes, and manages automation systems. No matter which robot company wins in the end, as long as factory automation keeps advancing and data centers, semiconductors, warehouses, and energy keep increasing capex, Rockwell has a chance to benefit.
In fiscal Q2 2026, both its software and control business and its intelligent devices business performed well, and the company raised full-year guidance. That shows the underlying automation cycle is recovering.
Summary and investment view
Let's briefly summarize the four companies today:
-
SERV represents explosive upside—high risk, high volatility, high imagination. The key question is whether real-world data and multi-scenario operations can work.
-
ISRG represents profitability and has already proven the business model. It is the most mature sample in Physical AI.
-
CGNX represents machine vision, the key link that lets automation systems understand the real world.
-
ROK represents industrial automation infrastructure, betting on the broader smart-factory and automation trend with stronger certainty.
Physical AI is not one single theme—it is an entire industrial chain. Smart investors do not just chase the hottest concept. They look for who can win orders, who can build data moats, who can generate recurring revenue, and who can survive cycles.
