On April 26, during the Beijing International Auto Show, Momenta held a themed sharing session titled "Momenta R7: The Prelude to Physical AI." Four partners appeared on stage together to officially announce the first mass production launch of the Momenta R7 reinforcement learning world model. The Momenta R7 reinforcement learning world model comprehends the laws of motion and interaction logic of the physical world, rather than relying on scene memory and rule matching.
Momenta announced that it has successfully delivered over 70 mass-production vehicle models and has received orders for more than 200 models in total. Its mass-production deployments cover more than ten countries and regions. The number of mass-produced vehicles equipped with Momenta's systems has exceeded 800,000 units, with the company capable of delivering 100,000 units in under 40 days. At this year's Beijing Auto Show, more than 20 brands and over 60 vehicle models featured Momenta's solutions, including newly launched models from Mercedes-Benz, Audi, and BMW.
Cao Xudong, Partner and CEO of Momenta, stated that prediction is the cornerstone of intelligence evolution. He explained that large language models rely on Next Token Prediction to compress common sense about the digital world, granting AI the ability to understand text and natural language. Conversely, world models use World Model Prediction to forecast future states and interaction logic within the physical world, enabling an understanding of objects' physical properties, cause-and-effect relationships in motion, and potential possibilities for interaction.
"Autonomous evolution is the key driving force for the continuous iteration of intelligence. In an environment, AI constantly receives feedback, undergoes trial-and-error iteration, and autonomously optimizes based on set goals. This evolutionary logic is precisely the core essence of reinforcement learning," summarized Cao Xudong. "Therefore, world models and reinforcement learning together form the two core pillars of Physical AI."
Reflecting on the original motivation for starting the company, Cao Xudong expressed emotion, recalling an encounter a decade ago on "Fairchild Drive," a street in Silicon Valley named after Fairchild Semiconductor. It was this very street, he noted, that nurtured the origins of the global semiconductor industry and ultimately shaped the legendary foundation of Silicon Valley. "In that moment, the fire within me was instantly ignited," he said. "And today, we hope to join hands with all Chinese AI companies to collectively write the Silicon Valley legend of the East."
Comments