Summary
Physical AI is attracting growing interest from technology companies and investors, but its potential extends beyond humanoid robots and automated machines. Amulya Vishwanath, a former NVIDIA robotics platforms marketing executive and founder of Techable VC, believes the field is still developing its fundamental technologies and has opportunities across robotics, materials science, and life sciences.
Vishwanath describes the current stage of robotics as its “Lego era,” in which essential components, software, robotic hands, and safety standards are still evolving. Her investment approach focuses on businesses that use AI to address practical problems in manufacturing, healthcare, and other industries.
What Is Physical AI?
Physical AI refers to artificial intelligence that interacts with or helps control the physical world. While many AI applications work with text, images, or digital information, physical AI connects intelligent software with machines, materials, biological systems, and real-world processes.
Robots are one example, but Vishwanath takes a broader view of the technology. She sees potential applications in molecules, new materials, medicines, and machines operating on factory production lines.
This broader definition means physical AI could influence several industries, including industrial manufacturing, healthcare, scientific research, and materials development.
For example, AI-powered systems could help factories automate repetitive operations, assist researchers in discovering new materials, or support the development of tools used in medical environments.
The technology is not limited to humanoid robots that perform visible physical tasks. Some of its most important applications may emerge in specialized industrial and scientific settings.
Why Robotics Is Still in Its Early Stages
Vishwanath spent seven years leading product marketing for NVIDIA’s robotics platforms. Based on her experience, she believes robotics has not yet reached the stage where a single breakthrough will make most existing development approaches unnecessary.
She compares the current state of the industry to a “Lego era,” in which developers are still improving the basic building blocks required to create useful robotic systems.
These building blocks include actuators, robotic hands, software platforms, and safety standards. Improvements in these areas can help robots perform tasks more reliably and operate in environments designed for humans.
The comparison also highlights a challenge facing the industry: developing a capable AI model is only one part of building a practical robot. A machine must also sense its surroundings, manipulate objects, respond to unexpected conditions, and operate safely.
Progress therefore depends on advances in both AI software and the physical hardware that allows machines to interact with the world.
Will Robotics Have a ChatGPT Moment?
Some technology leaders believe a general-purpose AI foundation model could transform robotics in the same way that ChatGPT changed public awareness of generative AI in late 2022.
Under this view, a broadly capable model could perform many tasks that currently require specialized robotic systems, reducing the need to develop separate solutions for individual applications.
Vishwanath’s perspective is more cautious. Her “Lego era” description suggests that important hardware, software, and safety challenges remain, even as AI models become more capable.
The debate reflects two different expectations for the industry’s development. One emphasizes the possibility of a major breakthrough from general-purpose models, while the other highlights the need to improve the underlying technologies and engineering systems.
Both software intelligence and physical capabilities will matter when evaluating whether robots can move from controlled demonstrations into dependable, real-world use.
Techable VC’s Investment Focus
Vishwanath is the founder and managing partner of Techable VC, an early-stage investment firm focused on physical AI.
Her experience at NVIDIA provides a background in robotics technology and connections across the technology industry. According to the discussion featured on TechCrunch’s Equity podcast, she uses her industry network to help evaluate companies and connect founders with potential customers.
Her investment interests extend across robotics, materials science, and life sciences rather than concentrating exclusively on humanoid robots.
This approach reflects the possibility that physical AI opportunities could emerge in less visible applications, including factory automation, scientific discovery, and technologies that improve how physical products are developed or manufactured.
For early-stage startups, access to technical expertise, industry relationships, and potential customers can be important when turning a promising technology into a commercial product.
Solving Practical Problems Through AI
Vishwanath’s investment philosophy emphasizes addressing meaningful problems rather than developing technology simply because it attracts attention.
One example is the challenge facing family-run factories that struggle to find enough workers. Automation could help these businesses maintain production, reduce repetitive manual work, and improve operational efficiency.
Healthcare is another potential area of application. Medical professionals, including surgeons, face demanding workloads, creating opportunities for technologies that could assist with physical tasks or improve workflows.
However, the usefulness of these systems will depend on their reliability, cost, safety, and ability to work alongside people. In healthcare, particularly, AI-enabled equipment must meet appropriate clinical and regulatory requirements.
The wider opportunity lies in identifying situations where physical AI can solve an existing problem more effectively, rather than introducing automation without a clear benefit.
What Does Conscious Automation Mean?
Another concept highlighted in Vishwanath’s discussion is “conscious automation.”
The phrase points toward an approach in which automation is designed around the needs of people and the environments in which machines operate, rather than simply replacing human involvement wherever possible.
In practical terms, this could mean developing systems that support workers, improve difficult or repetitive tasks, and help organizations address labor shortages.
For manufacturers, this may involve automating selected production activities while employees supervise operations and handle more complex work. In healthcare, technology could assist professionals without removing the need for human expertise and accountability.
The exact implementation will vary by industry, but the underlying question is how automation can deliver practical benefits while accounting for the people affected by it.
The Role of the United States and China
The physical AI industry is also shaped by competition between major technology markets, particularly the United States and China.
Both countries have companies working on robotics, AI models, manufacturing technologies, and related hardware. Their respective strengths may influence how quickly different applications develop and reach commercial scale.
Vishwanath discussed the relative positions of the two countries during the Equity podcast. Her investment perspective includes evaluating founders, technical capabilities, industry networks, and opportunities for commercial adoption.
For investors, the competitive landscape matters because physical AI businesses often depend on more than software development. Access to manufacturing capabilities, specialized components, engineering talent, and customers can all affect a startup’s ability to grow.
The long-term competitive position of individual companies will depend on how effectively they combine these resources with useful products and reliable technology.
Humanoid Robots Need Practical Applications
Humanoid robots often attract public attention through demonstrations involving dancing, walking, or performing other visually impressive movements.
However, commercial adoption requires machines to deliver measurable value in real environments.
Factories, warehouses, and other workplaces could provide opportunities for robots that can handle useful physical tasks. To succeed, these machines must be able to perform consistently, operate safely, and justify their cost.
Vishwanath’s broader perspective on physical AI reinforces the importance of looking beyond demonstrations. Investors and businesses need to consider what a robot can accomplish, which problem it solves, and whether it can operate reliably outside controlled settings.
The same principle applies to other forms of physical AI, including systems designed for scientific research, materials development, and healthcare.
Conclusion
Amulya Vishwanath’s perspective highlights a broader view of physical AI at a time when humanoid robots and general-purpose AI models are receiving considerable attention.
Her “Lego era” description suggests that robotics still needs progress in hardware, software, and safety before many applications can become widely practical. Through Techable VC, she is looking at opportunities across robotics, materials science, and life sciences, with an emphasis on solving real-world problems.
As the industry develops, progress will depend not only on more capable AI models but also on reliable physical systems, useful applications, and sustainable business models. The most significant opportunities may come from technologies that improve everyday industrial and scientific work, rather than from robots designed primarily to impress audiences.