The concept of Artificial Intelligence (AI) has existed for decades. The field was formally established in 1956 at the Dartmouth Summer Research Project on Artificial Intelligence, where John McCarthy coined the term. Yet, the groundwork began earlier, with ideas about thinking machines and machine learning surfacing in the 1940s and early 1950s.
Still, it wasn’t until late-2022 that AI truly entered the mainstream consciousness—thanks to the release of ChatGPT. For many, the capabilities of Generative AI (GenAI) felt like a mix of technical wizardry and science fiction. GenAI gave anyone with a computer the power to create something—an image, essay, video, or code snippet—that hadn’t existed just moments earlier. Suddenly, creative and analytical work could be executed in seconds from a single prompt.
This breakthrough triggered anxiety and anticipation across industries. Professionals in graphic design, marketing, consulting, software development, and data entry realized that their jobs might soon be supplemented—or replaced—by AI.
The next leap came with Agentic AI: systems that not only generate content but take autonomous actions in digital environments. These agents now assist with coding, customer service, workflow automation, and other enterprise tasks, lowering costs and streamlining workflows across industries.
And yet, while Generative and Agentic AI continue to dominate headlines, a quieter but potentially more transformative movement is underway: AI is becoming physical.
In his keynote at CES 2025, NVIDIA CEO Jensen Huang declared, “The next frontier of AI is physical AI.” The shift he described reflects a fundamental change: instead of AI operating solely in software, we are now seeing intelligent systems embedded in machines that can perceive, decide, and act in the real world. If you haven’t taken a trip to San Francisco recently, you might be surprised to witness how many Waymo’s (formerly known as the Google Self-Driving Car Project) are shuttling people around from A to B without a human behind the wheel.
So—is the future of AI physical? Let’s explore.
To understand why Physical AI is different, it's helpful to look at how AI has evolved in stages:
Generative AI focuses on content creation—text, images, video, code. Tools like GPT-4 and DALL·E are examples. They are impressive but fundamentally passive: they respond to prompts but don’t take initiative or interact with the world beyond the screen.
Agentic AI introduces autonomy in digital spaces. These systems can initiate actions, adapt to real-time feedback, and coordinate complex tasks—such as managing workflows or handling customer service tickets. Think of these agents as “digital employees” that operate within the bounds of software.
Physical AI, by contrast, brings intelligence into the physical world. It powers machines that can move, manipulate, and make decisions in real environments—robots, autonomous vehicles, drones, and more. These systems sense their surroundings using cameras, radar, LIDAR, and tactile sensors, and act on them using motors, actuators, and robotic limbs.
As Huang described, we’re on the cusp of a “ChatGPT moment for robotics.” That is, a leap in general-purpose robotic intelligence that makes machines useful, flexible, and context-aware in everyday settings.
In essence:
This progression reflects AI’s transition from the virtual to the tangible—a convergence of machine learning, robotics, sensor tech, and embedded systems. And it’s opening doors to new applications far beyond what software alone can accomplish.
Source: Recreated from Jensen’s Huang Keynote at CES 2025
Physical AI is beginning to garner more attention because it represents AI’s final leap from cyberspace into the physical world. By moving AI into physical processes, enormous economic and societal value is unlocked that purely digital AI cannot reach.
Several factors explain why many see Physical AI as the next frontier in AI:
1. The Fusion of Digital and Physical Worlds
After decades of computing confined to screens, we are entering a new phase where AI systems are embedded in our environment. Advances in spatial computing, computer vision, and sensor integration make it possible for AI to perceive the world and interact with it.
Major tech companies are investing accordingly. OpenAI has filed trademarks for humanoid robots, AR glasses, and AI-powered hardware—signaling plans to bring AI into the physical realm. The World Economic Forum summarized it best: “AI’s next great leap will be powered by hardware.”
As AI increasingly merges with robotics, wearables, and autonomous devices, it begins to play an active role in how we live and work.
2. From Passive Outputs to Active Agents
Generative AI stunned the world with its creativity, but it remains passive. It can answer questions or generate content, but it cannot take real-world action.
Agentic AI systems bridge that gap in digital spaces—and increasingly, they’re crossing into the physical world. Huang noted that the shift from suggestion engines to autonomous agents is what enables Physical AI to flourish. Systems that can perceive, decide, and physically act—grasp objects, navigate roads, care for humans—represent a qualitative leap in intelligence and utility.
3. Trillions in Economic Potential
Physical AI is poised to transform vast industries—transportation, manufacturing, logistics, healthcare, domestic labor—all of which involve real-world activity that software alone can’t replace.
Huang predicts that autonomous machines could create “the first multi-trillion-dollar robotics industry.” Vinod Khosla believes humanoid robots could eventually surpass the automotive sector in value.
This isn’t just hype: NVIDIA’s autonomous systems category is already worth over $4 billion in revenue, with $5 billion projected for 2025. Billions in venture capital are flowing into robotics and AI-driven automation, in pursuit of the next platform-shifting companies.
4. Solving Labor and Demographic Challenges
Many developed economies face serious labor shortages in jobs involving physical work—warehouse operators, truck drivers, caregivers, construction workers.
Physical AI can fill these gaps. Robots can perform dangerous or monotonous tasks, while autonomous vehicles and assistive machines can augment human labor in ways that are cost-effective and scalable.
Rather than displacing humans, Physical AI has the potential to enhance human productivity, freeing people to focus on creative, interpersonal, or supervisory work.
5. Technology Is Finally Ready
Several enabling technologies have reached maturity:
Hardware is catching up too. NVIDIA’s Thor chip, designed for autonomous vehicles, delivers 20× the processing power of its predecessor. These advances mark a turning point—making Physical AI not only possible, but increasingly viable at scale.
Physical AI represents more than a technological leap. It’s a new paradigm—where intelligence isn't confined to servers and screens but infused into the infrastructure around us.
As a World Economic Forum report put it, “The era of software-only AI is coming to a close. The next chapter belongs to physical computing, where intelligent systems interact with and respond to the world around us.”
It’s still early, but the momentum is undeniable. Robots are showing up in warehouses, autonomous cars are hitting city streets, and AI assistants are appearing in homes and hospitals. The industries that adopt Physical AI fastest will likely gain a competitive edge—and the companies that build its core technologies may define the next tech era.
So—is the future of AI physical?
Increasingly, yes. We are moving toward a world where AI doesn’t just think or talk—it moves, builds, drives, and cares. That future is being built now.
DISCLOSURE: Securities highlighted or discussed in this communication are mentioned for illustrative purposes only and are not a recommendation for these securities. Evergreen actively manages client portfolios and securities discussed in this communication may or may not be held in such portfolios at any given time. This material has been prepared or is distributed solely for informational purposes only and is not a solicitation or an offer to buy any security or instrument or to participate in any trading strategy. Any opinions, recommendations, and assumptions included in this presentation are based upon current market conditions, reflect our judgment as of the date of this presentation, and are subject to change. Past performance is no guarantee of future results. All investments involve risk including the loss of principal. All material presented is compiled from sources believed to be reliable, but accuracy cannot be guaranteed and Evergreen makes no representation as to its accuracy or completeness.