Archetype AI Team
The world may be chaotic but it isn’t random. There is cause and effect, action and reaction. From the basic laws of physics to how people interact with environments, there are hidden patterns and behaviors all around us. But these patterns are too complex and fast-moving for humans to perceive.
Imagine a world where technology could help us make sense of the world's hidden patterns, understand the root causes of problems, and identify solutions.
At Archetype AI, we believe such a world is possible. We’re unveiling a new form of artificial intelligence that takes us a step closer to this reality.
Today we are introducing Newton™ – a first-of-its-kind foundation model for understanding the physical world. Instead of analyzing online text and images like the common generative AI models that exist today, it fuses real-time sensor data – such as from radars, cameras, accelerometers, temperature sensors, and more – with natural language, so you can ask open-ended questions about the world around you, and take informed action to make it better.
The world’s biggest problems are physical, not digital. Our mission is to use the power of AI to solve these real-world problems. With our early customers and partners across industries, we are already building solutions with Newton to do things like:
- Improve construction safety and reduce waste at massive construction sites
- Provide vehicles with situational awareness
- Enhance home safety and energy efficiency
- Optimize logistics and supply chains
- Predict maintenance for industrial equipment
The Era of Physical AI
With Newton, we believe we’re at the dawn of the physical AI era. This new form of artificial intelligence is capable of perceiving, understanding, and reasoning about the world in ways that were previously impossible.
We are now at a unique moment in time where the development of several critical technologies is converging to make physical AI possible:
- Sensor adoption – the IoT era (early 2010s) led to the proliferation of sensors. However, this first era of smart devices was built with simple machine learning techniques that are only capable of learning individual skills for bespoke, verticalized solutions. Today we are at the cusp of the trillion sensor economy.
- Cloud compute – massive adoption of cloud computing infrastructure at scale over the last decade, which has made it possible to create sensor networks and large AI models.
- Shift from ML to AI – the advent of large transformer-based models able to generalize to tasks they haven’t seen before and learn new things from small numbers of examples.
Previously, our biggest challenges have been solved with one-off bespoke solutions, which are costly and slow to scale. In the era of physical AI, we believe that we will see the rapid proliferation of AI solutions that can be easily adapted and scaled across use cases.
A Human-centric Perspective
While the opportunities for physical AI seem nearly limitless, we believe in building products with a human-centric perspective where AI augments and accelerates people, instead of trying to replace them. This principle is one that our team has carried with us throughout our careers.
New interaction models need to be created for physical AI. We bring deep and diverse experience building and scaling some of the most widely used technology products in the world. From leading Google’s Advanced Technology and Projects group to our time at companies like Sony, NASA, Disney Imagineering, Samsung, MIT, and Stanford, we have worked on core technologies that have been shipped in millions of devices, from the world’s smallest radar to computational woven materials, to distributed learning for private AI models.
Throughout each of these experiences, we have sought to create solutions that can disappear into the background and feel truly integrated with our lives. Now with Archetype AI, we have the opportunity to build purposeful products that are in service of humanity.
This is Just the Beginning
The endeavor ahead of us is an ambitious one. Our ultimate vision is to encode the entire physical world. This isn’t something that is possible to achieve alone.
To that end, we are proud to have an incredible set of partners joining us to make this vision a reality. We’ve raised $13 million in seed funding led by Venrock, with participation from Amazon Industrial Innovation Fund, Hitachi Ventures, Buckley Ventures, Plug and Play Ventures, Earl Grey Capital, Strange Ventures, and several angels. You can read the official press release here, as well as Steven Levy’s story about us in Wired.
– Ivan Poupyrev, Brandon Barbello, Leonardo Giusti, Jaime Lien, and Nicholas Gillian
