Ai Merges With Machines: A Slow But Revolutionary Evolution

The Age of Physical AI: A Reality Check

Recent years have seen significant transformations in the artificial intelligence (AI) landscape, with language models like ChatGPT taking center stage. However, as industry stakeholders gather to discuss the future of AI, a more nuanced understanding emerges. Physical AI, which integrates AI with robots, vehicles, and other devices, is not poised for a single breakthrough or viral demo. Instead, it is expected to be a gradual process, shaped by the complexities of embodied intelligence.

At the AWS re:Invent 2025 conference in Las Vegas, panelists Ryu Jung-hee, founder and CEO of South Korean robotics startup RLWRLD, Kevin Peterson, CTO of Bedrock Robotics, and Sri Elaprolu, director of AWS Generative AI, shared their insights on the future of Physical AI. While they acknowledged the excitement surrounding ChatGPT-like language models, they emphasized that Physical AI is constrained by realities that digital AI never faced.

One of the key limitations of digital AI is its inability to interact with the physical world. Unlike robots and autonomous systems, which can perceive and respond to their environment, digital AI exists solely in the virtual realm. This distinction has significant implications for the development of Physical AI. According to Jung-hee, “Physical AI is not just about throwing more computing power at a problem; it’s about creating systems that can truly interact with the world around them.”

The importance of robotics and embodied intelligence cannot be overstated. As NVIDIA CEO Jensen Huang noted during the conference, “The future of AI is physical AI, embodied by robots, including self-driving cars.” However, this vision is not without its challenges. The development of humanoid robots, for instance, requires a deep understanding of human physiology, neuroscience, and psychology.

In San Francisco’s Silicon Valley, where the artificial intelligence revolution is underway, robotics investment and innovation are on the rise. Companies like Waymo are already making waves in the autonomous vehicle sector, with their driverless cars becoming a common sight in the city. However, despite this progress, experts caution that the development of Physical AI will be a slow and iterative process.

The complexities of predicting technological advancements were highlighted at the PRSA annual Silicon Valley Media Predicts event. Don Clark of The New York Times noted that his newspaper has never found a technology it can’t find the dark side of. He referenced an early 1953 article warning that robots might someday pose a threat to humanity, mirroring concerns raised by modern experts about the ethics and governance of Physical AI.

In this context, panelists emphasized the need for human capital, expertise in robotics, and advanced manufacturing capabilities to drive the development of Physical AI. Brett Adcock, Figure AI CEO, highlighted the importance of hiring talented engineers and technicians to tackle challenges like bipedal dynamic control, deep learning, and more.

The Bay Area’s top universities are fueling Silicon Valley’s growth by training engineers and technicians who can address these complex problems. The region is home to a diverse group of robotics professionals from leading companies, professors, and industry executives, all working together on training and education initiatives.

These efforts aim to develop hardware platforms that enable students to learn the basics of upper-body versus lower-body dynamics, manual dexterity, and degrees of freedom for the neck and hip. The focus is not only on theory but also on practical applications, with experts seeking to create systems that can interact with the physical world in a meaningful way.

As NVIDIA announced a significant reduction in the price of its Orin Nano AI chipset, marking a major milestone in edge intelligence, experts highlight the significance of agentic AI. These robots will have reasoning capabilities, enabling them to retrieve knowledge, reason over long time horizons, and physically respond with agility. While this vision is still in its infancy, it represents a crucial step towards creating systems that can truly interact with the world around them.

The San Francisco Bay Area’s role as a hub for AI development and innovation cannot be overstated. As experts caution that this vision will not emerge overnight, the development of Physical AI will require time, patience, and significant investments in research, education, and manufacturing capabilities.

While Physical AI holds tremendous promise for transforming industries and revolutionizing the way we interact with technology, its emergence will be a gradual process, shaped by the complexities of embodied intelligence. Industry stakeholders must acknowledge both the challenges and opportunities that lie ahead.