Mistral Launches First Robotics AI Model in Major Push Into Physical AI

French AI outfit Mistral has ventured into a hot new sector with its first robotics AI model. The startup’s new model is intended to provide robots with a better understanding and awareness of the world they occupy, moving into areas beyond large language models and more towards what is called physical AI.

The models will empower robots to understand, reason with and interact with physical reality, potentially unlocking new applications in manufacturing, research, service industries, and logistics. Mistral is joining a wave of other major tech companies pouring funds into robotics.

Physical AI refers to the applications of artificial intelligence to physical tasks, requiring systems to perceive and react to the physical world with sensors, cameras and manipulators in addition to performing traditional cognitive tasks like learning and reasoning. In contrast, a traditional language model deals solely with symbolic or image information and generates either a textual or graphical output; robots deal with real-world objects and events.

“The model is aimed at helping developers build and deploy the autonomous robots of tomorrow for applications across manufacturing, logistics, household chores and scientific research,” Mistral said.

This is in line with broader trends in the AI space: while large language models and generative AI have seen a rapid evolution and adoption, many companies are now looking to imbue machines with AI for the real world. Robotics that can process sensory data, plan movements, and perform physical tasks is being positioned as the next growth market for the artificial intelligence sector.

“Physical AI will probably be one of the main AI trends over the next decade,” says industry expert Nick van Hoogstraten. “You can see a lot of activity here from various players.”

The space is already quite crowded. Nvidia, Google DeepMind, OpenAI and Tesla are some of the prominent players along with various robotics startups that are betting on building AI models for physical tasks.

Mistral is no stranger to developing highly competitive AI models, already gaining acclaim for its open weight language models and solutions aimed at enterprise use cases. Entering the robotics segment is seen as a natural progression to broaden its business into another burgeoning market, especially in Europe, a growing hot-spot for AI innovation.

There are several reasons for this sudden interest in the field. Industries such as manufacturing are expected to continue needing more advanced robotics due to labour shortages and increasing pressure for high productivity. This demand for robotics that can perform more complex and delicate tasks with less programming is where this new category of AI models comes in, as the goal is to make these robots more autonomous and capable of learning.

Physical AI requires machines to interact and operate safely in real-world environments that are constantly changing and not entirely predictable, which is a challenge. For robots, in addition to being able to intelligently perceive and understand the physical world, they also need to have a certain amount of knowledge about real-world constraints.

The demand for robotics as a solution for numerous applications from manufacturing to home and service is on the rise and companies are keen to find solutions for problems that are currently faced due to human limitation or for repetitive or hazardous tasks.

There is no shortage of potential applications for AI powered robotics, with applications including warehouse automation to industrial manufacturing, personal and household assistance and autonomous vehicle technology being among those expected to benefit the most. As an increasing number of businesses look to increase productivity, improve supply chains and enable employees to focus on higher value work, demand for robotics technologies that combine the ability to work in complex, unstructured environments with improved processing speeds and smarter learning capabilities continues to grow.