AuraML Unveils Multimodal World Simulation Model Built on NVIDIA AI Infrastructure

AuraML, a pioneer in generative AI for physical systems, today announced the launch of its groundbreaking Multimodal World Simulation Model- AuraSim. This industry-first platform will generate physics-ready, high-fidelity simulation environments for robots using diverse inputs such as text descriptions, floorplans, videos, depth maps, and point clouds.

As the first multimodal world simulation model built leveraging NVIDIA Omniverse and NVIDIA Cosmos, AuraML is positioned to become a critical infrastructure layer for the global robotics industry, significantly reducing the time and cost required to develop robust physical AI.

To accelerate the adoption and expertise around its new World Simulation Model, AuraML in collaboration with NVIDIA Inception is also launching a dedicated training program: The Global Physical AI & Robotics Cohort.

Bridging the Gap Between Digital and Physical Reality

Training autonomous systems demands large-scale, high-quality simulation data.  AuraML addresses this by automatically  transforming multimodal inputs into "sim-ready" digital twins that respect the laws of physics.

Key Features of the AuraML World Model:

●Multimodal Flexibility: Generates complex 3D environments from 2D floorplans, and converts text  descriptions to  raw sensor data.

●Physics-Ready Outputs: AuraML produces environments with accurate collision boundaries and physical properties, ready for deployment in high-fidelity engines.

●Agentic Automation: The model integrates an automated simulation pipeline powered by AI agents, capable of running thousands of scenario variations without human intervention.

●Rapid Iteration: Developers can quickly "stress-test" Physical AI by generating edge-case scenarios by simply modifying a text prompt.

Use Case: Orchestrating the "Dark Factory"

The power of AuraML is most evident in complex factory automation. Using the platform, users can generate an entire industrial facility from a single architectural floor plan and a few video walkthroughs. This allows for the simultaneous simulation of diverse robotic fleets working in sync:

●Humanoid Robots navigating narrow corridors and performing dexterous assembly.

●Autonomous Mobile Robots (AMRs) managing logistics and material transport across the floor.

●Stationary Robotic Arms performing high-speed precision tasks.

Additionally, AuraML enables developers to test how these different agents interact on the same floor, ensuring collision-free synchronization and optimized workflows before a single robot is ever deployed in the real world.

Supercharged by NVIDIA Technology

To deliver  industry-leading performance and physical accuracy, AuraML worked closely with NVIDIA to build its platform on NVIDIA Omniverse, and NVIDIA Isaac Sim - a collection of libraries and microservices for developing physical AI,  digital twins, and robotics. Powered by NVIDIA Cosmos world foundation models and accelerated computing, AuraML creates real-time, photorealistic environments that enable  seamless data transfer between the digital and physical worlds.

AuraML’s Global Physical AI & Robotics Cohort

 To accelerate the adoption and expertise around its new World Simulation Model, AuraML in collaboration with NVIDIA is launching The Global Physical AI & Robotics Cohort,. An intensive, 8-week program designed for robotics engineers, AI researchers, and developers.  With limited seats, the cohort offers  a high-touch, mentorship-driven experience focused on building, stimulating, and deploying real-world physical AI systems.

 Powering the Future of Robotics Foundation Models

The launch marks a significant milestone in the evolution of "Physical AI", and by  providing an end-to-end agentic workflow, AuraML enables robotics companies to move from concept to deployment at unprecedented speed. 

"Our integrated workflows provide the critical infrastructure needed to make Physical AI more robust," said Ayush Sharma, co-founder of AuraML. "By launching the first multimodal world simulation model from India, powered by NVIDIA's world-class simulation technology, we are providing a global toolset to accelerate the development of advanced robotics foundation models everywhere,” he added."

“India’s AI startup ecosystem is primed for acceleration, driven by exceptional technical talent and global ambition,” said Tobias Halloran, Director of EMEAI Startups and Venture Capital at NVIDIA. “NVIDIA is accelerating this momentum by giving founders direct access to accelerated computing, scalable AI infrastructure, and programs like NVIDIA Inception and the Inception VC Alliance - helping startups scale faster and build for global markets.”

The announcements align with the broader momentum around the India–AI Impact Summit 2026, a flagship gathering hosted by the Government of India under the IndiaAI Mission, taking place in New Delhi.