At the 2025 Consumer Electronics Show (CES), NVIDIA launched the brand-new Cosmos platform, designed to accelerate the development of physical artificial intelligence (AI) systems, particularly for autonomous vehicles and robots. The Cosmos platform integrates generative world foundation models (WFM), video annotators, safety mechanisms, and an accelerated data processing pipeline, enabling developers to create and optimize AI models with reduced reliance on real-world data.
The Cosmos platform will be available in an open model license format in the Hugging Face and NVIDIA NGC catalogs, with optimized NVIDIA NIM microservices to follow, providing enterprise support through the NVIDIA AI Enterprise Software Platform.
NVIDIA CEO Jensen Huang stated at the event, "Robotics is about to experience a pivotal moment similar to that of ChatGPT. Just like large language models, world foundation models are essential for advancing robotics and autonomous vehicles, but not all developers have the capability and resources to train their own models. We created Cosmos to democratize the development of physical AI, allowing every developer to access general robotic technologies."
The Cosmos model can generate physics-based high-definition videos from text, images, and sensor data, making it suitable for applications such as video search, synthetic data generation, and reinforcement learning. Developers can customize the model to simulate industrial environments, driving scenarios, and other specific use cases. Additionally, NVIDIA has introduced NeMo Curator, an accelerated video processing pipeline capable of processing 20 million hours of video data in just 14 days, as well as the Cosmos Tokeniser, a visual data compression tool.
Pras Velagapudi, CTO of Agility Robotics, pointed out, "Data scarcity and variability are key challenges for successful learning in robotic environments. Cosmos's ability to transform text, images, and video into world scenarios enables us to generate and enhance scenes for various tasks, allowing us to train models without needing excessive costly real data capture."
Several major robotics and transportation companies, including Agile Robots, XPENG, Waabi, and Uber, have begun adopting Cosmos for AI development. Uber CEO Dara Khosrowshahi stated, "Generative AI will drive the future of mobility, requiring both rich data and robust computing power. Through our collaboration with NVIDIA, we are confident in helping to accelerate the process of safe, scalable autonomous driving solutions."
In addition to Cosmos, NVIDIA also introduced the Llama Nemotron large language model and the Cosmos Nemotron visual language model, developed specifically for enterprise use in industries such as healthcare, finance, and manufacturing.
Official blog: https://nvidianews.nvidia.com/news/nvidia-launches-cosmos-world-foundation-model-platform-to-accelerate-physical-ai-development
Key Points:
🌍 The Cosmos platform aims to accelerate the development of autonomous vehicles and robots while reducing reliance on real data.
🚀 Developers can customize models based on their needs to generate video data for various applications.
🤖 Several robotics and transportation companies have begun using Cosmos to accelerate the practical application of AI technology.
