For NVIDIA, an "open model" is more than just open weights. Their approach provides a complete toolkit: the pre-trained model itself, the training framework required to fine-tune it with custom data, and even the datasets used in its creation, enabling true customizability and innovation.
Traditional simulators are rule-based, programmed with physics equations. World models pioneer "neural simulation," which learns the physics of the world implicitly from massive datasets of visual observations. It's a pattern-recognition approach to predicting outcomes, rather than one based on pre-defined rules.
NVIDIA's open model development is a strategic R&D effort. By building models firsthand, they gain deep insights into computational demands, which directly guides the design of next-generation hardware like GPUs, ensuring their products meet future ecosystem needs.
Unlike digital applications, every physical AI device like a robot has a unique hardware setup with different sensors. Closed, monolithic models cannot cater to this variety. Open models are essential as they provide a foundational base that developers can customize and fine-tune for their specific physical embodiment.
Instead of creating separate models for understanding the past (e.g., explaining a video) and simulating the future, NVIDIA is merging these capabilities. Their Cosmos model shares a core world representation, creating a unified "omni model" that handles diverse inputs (text, video, action) and outputs.
Cloud AI can batch user requests for efficiency. Physical AI devices like robots operate in real-time on a single stream of data ("batch size one"). This fundamental difference necessitates different, more efficient model architectures that don't rely on aggregation for performance.
A key application of world models is to simulate physical environments. This allows developers to test AI "policies" (e.g., a self-driving car's decision-making model) virtually, drastically speeding up development cycles and reducing reliance on expensive, slow, and risky real-world testing.