If there’s one thing that BEYOND EXPO in 2026 has shown us, it’s that physical AI is the next frontier.
Embodied or physical AI is a cutting-edge field of artificial intelligence where algorithms are integrated into physical or virtual bodies. Think robots, smart drones, autonomous vehicles.
Instead of just processing data on a screen, these AI systems can perceive, learn from, and interact with the physical world in real-time.
The challenge is that the precision with physical AI needs to be perfect to prevent any mishaps. Also, it’s harder to amass data for embodied AI, compared to digital data that LLMs can study from.
For now, though, that frontier is still some distance away. About three to five years away, actually, as some experts say.
Which, if you think about it, means now is the perfect time for you to study a degree related to this burgeoning field.
What to study if you want to work with physical AI
Mechatronics engineering
Perhaps one of the broader paths that can lead into embodied AI, mechatronics fuses mechanical engineering, electronics, and computer science into a single curriculum.
In simple language: You’ll learn how motors, sensors, microcontrollers, and control systems talk to each other.
If you want to be the person who makes a robot’s arm actually move reliably, this is your degree. Universities in Germany, Japan, and increasingly Southeast Asia have some of the strongest programmes in the world.
Physical AI doesn’t have to be humanoid. Source: Kindel Media via Pexels
Robotics engineering
Where mechatronics gives you the building blocks, a dedicated robotics degree gives you the full system. You’ll go deep on kinematics (how bodies move through space), path planning, simultaneous localisation and mapping (SLAM), and robot operating systems like ROS 2.
Many programmes now embed machine learning directly into the curriculum, training you to bridge the gap between a neural network’s decision and a physical actuator’s response.
Both mechatronics and robotics blend mechanical, electrical, and software engineering. The core difference is scope: mechatronics is a broad field focusing on electromechanical systems (like CNC machines), while robotics is a specialised subset focusing on machine motion, perception, and autonomy (like AI-driven drones or robotic arms).
Electrical and electronic engineering (EEE)
Physical AI still runs on circuits, and EEE gives you the ability to design the low-level hardware that everything else depends on — power management, PCB design, embedded firmware, and signal processing.
Companies building their own AI chips (think NVIDIA, Qualcomm, or any serious robotics startup) almost always need EEE graduates who understand how silicon and software meet.
Computer science (Robotics or AI specialisation)
A computer science degree leads into many careers, but with a robotics or AI specialisation specifically, you’re set for a pretty direct passage into a career dealing with embodied AI.
This degree would be particularly interesting to those who want to software side but still want to work in the physical world. The researchers teaching robots to learn from human demonstration, or training autonomous cars on simulation data, are largely coming from this background.
A strong CS foundation also gives you the flexibility to pivot as the field evolves.
Mechanical engineering
When building a physical embodiment of AI, be it a humanoid rbot or a vehicle, someone has to design the body.
An education in mechanical engineering would allow you to do this. Mechanical engineering is the branch of engineering that designs, builds, and tests physical machines, devices, and thermal systems.
It applies principles of physics, mathematics, and materials science to create and maintain everything from microscopic sensors to massive aerospace engines. And of course, robots or other kinds of embodied AI.
Cognitive science
This one might be a less obvious choice, but speaks to those who aren’t as savvy with the technical side of things. Plus, it’s an increasingly important one in the world of AI.
Physical AI is moving toward systems that interpret human intent, adapt to unpredictable environments, and work alongside people rather than in isolation.
Researchers with backgrounds in how biological systems process sensory information, make decisions, and learn motor skills are shaping the next generation of brain-computer interfaces, prosthetics, and human-robot interaction. Pair it with programming and you have a genuinely distinctive profile.
Hardware vs software: Are jobs in AI safe from AI?
As digital AI moves into the physical world, are software jobs becoming less in demand compared to those related to hardware components?
That’s a concern that Matt White, the Global CTO of AI at The Linux Foundation, shared during his fireside chat at BEYOND EXPO.
“Even people within AI are worried about their jobs because now you’ve got recursive research and AI models that can build themselves,” White said on stage.
But of course, that doesn’t mean that computer science students are out of jobs. In fact, there are still new jobs being created now. It just means that they have to be a bit more intentional and choosing on focus on the right stuff.
“If you are a student today and you’re taking computer science, I would strive to be a full-stack engineer,” he says. “Understand how to build with AI, how to build a model and how to fine tune it, how to create an application with it.”