Intrinsic, a robotics software and AI company now part of Google, is aiming to make industrial robots more accessible to smaller manufacturers by open-sourcing its basic software. Wendy Tan White, CEO of Intrinsic, explained the company's mission is to provide a platform for robotics and AI that is analogous to "Android for robotics and AI."
Intrinsic, initially part of Google's "Other Bets" group, was integrated into Google in February and now works closely with DeepMind, utilizing Google's Gemini models and cloud services.
Tan White, who has a background as an entrepreneur, investor, and operator, co-founded the U.K. website-building platform Moonfruit, worked with growth capital funds, and was a general partner at Entrepreneur First. She then joined Alphabet's X, the moonshot factory, where she launched Intrinsic in 2021.
The Evolution of Manufacturing
Tan White highlighted that traditional manufacturing processes, largely unchanged since the Industrial Revolution, require high volume and product similarity due to cost constraints. This model often prevents customization at an affordable price. Intrinsic's goal is to enable more flexible production, allowing a single robot cell to perform multiple tasks and handle diverse products. This flexibility aims to reduce inventory waste and make factories more adaptable.
The company also aims to address the difficult, dangerous, and repetitive tasks often performed by humans in manufacturing, such as deboning chicken or working on assembly lines. The integration of robotics and AI, Tan White believes, can lead to more intelligent and flexible production, potentially bringing manufacturing back locally and reducing reliance on complex, long-distance supply chains.
Physical AI and Google's Integration
Intrinsic's return to Google signifies a focus on "physical AI," extending the intelligence of generative AI and agentic tools into the physical world through vision, touch, and mobility. Tan White views Intrinsic as a core component of Google's next-generation software for physical AI.
The integration has provided Intrinsic with greater certainty and focus on its core mission without the immediate need for external investment. While Intrinsic maintains its separate identity and CEO, its alignment with Google allows for faster decision-making and a clearer line of sight to senior leadership.
Initial Applications and Partnerships
Intrinsic is initially focusing on manufacturing environments, which are semi-constrained and have established safety protocols, rather than more complex settings like homes. One key application is machine tending in smaller machine workshops that produce components for devices like phones and servers. These workshops often face skilled labor shortages and high demand, which Intrinsic's flexible robot cells can help address.
The company is also working with partners like Foxconn, the world's largest electronics manufacturer, and Nvidia. The Nvidia partnership involves simulation through Nvidia Omniverse and exploring grasping technologies. Intrinsic's platform is designed to be agnostic, integrating skills from various sources, including Nvidia, Google's Gemini, and its own vision models. This allows for the creation of modular robotic systems that can perform tasks like machine tending, inspection, polishing, and even assembly.
Future Outlook and Workforce Skills
Tan White emphasized Intrinsic's focus on democratizing access to robotics for small and medium-sized manufacturers globally. She also noted the potential for sophisticated tooling and AI to attract younger generations to manufacturing careers, which are currently facing a significant skilled labor shortage.
Regarding the future workforce, Tan White advises young people to focus on their passions and unique strengths, particularly in areas like design thinking, curation, and breakthrough creative thinking, which she believes will become increasingly valuable as AI handles more routine tasks. She also expressed optimism about women's potential in AI, suggesting their often multi-hybrid skill sets are well-suited for integrating and utilizing AI tools collaboratively.