Automating Scenario Variation in Simulation at Scale
Nov 06, 2025

AMD Silo AI is initiating a collaboration withAILiveSim, a pioneer in high-fidelity multi-sensor simulation for industries including maritime, defense, aerospace, and automotive, to advance AI-driven scenario generation and sensor simulation workflows on AMD Instinct™ and Radeon™ Pro GPUs.
This collaboration unites AMD’s scalable compute platforms with AILiveSim’s photorealistic and physics-accurate simulation engine. Together, they enable industrial developers to create, test, and validate AI systems across multisensors —visual, infrared, radar, lidar, and inertial sensors — with unprecedented realism, control, and efficiency.
“Working with AMD Silo AI allows us to scale our platform to meet the growing demand for high-performance, on-premise simulation,” saidJérôme Leudet, CEO of AILiveSim. “By leveraging AMD’s GPU technology and the ROCm™ software stack, we can deliver richer and faster digital-twin environments to our customers.”
The collaboration will focus on:
Optimizing AILiveSim’s software stack for AMD hardware and ROCm™, ensuring seamless performance across diverse compute configurations.
Developing AI-assisted scenario generation pipelines that extract semantic intrinsic signatures from real-world data to retrieve, edit, and refine simulation assets from AILiveSim’s catalogue—minimizing the sim-to-real domain shift.
Exploring generative AI and scene reconstruction techniques to automate the creation and adaptation of synthetic environments, improving fidelity and diversity for training and validation.
AILiveSim’s engine provides full programmatic control over geometry, materials, lighting, and sensor configurations—producing highly detailed, physics-grounded environments ideal for testing perception and autonomy stacks in both open- and closed-loop scenarios.
“Together with AILiveSim and our growing ecosystem of simulation software providers, we are narrowing the gap between real and simulated data,” saidNiko Vuokko, Sr. Director at AMD Silo AI. AILiveSim’s state-of-the-art simulation technology and the leading-edge compute infrastructure of AMD, in combination, shorten customer development cycles while accelerating innovation”.
This collaboration reinforces AMD Silo AI’s commitment to building an open and scalable simulation ecosystem—leveraging the full AMD compute stack from edge and digital-twin infrastructure to large-scale AI cloud compute, and building on AMD Silo AI’s broader simulation and world-modeling efforts, following recent initiatives withParallel Domain andRobotec.ai.
The joint work will also include open-source software releases, reinforcing both companies’ commitment to transparency, interoperability, and responsible AI innovation.
As a leader in end-to-end compute and AI R&D, AMD Silo AI continues to expand global access to cutting-edge AI capabilities, helping industries harness simulation, generative AI, and high-performance computing to build safer, smarter, and more sustainable systems.
About AILiveSim
AILiveSim is an AI-powered, multi-sensor simulation platform that enables companies to safely and efficiently develop autonomous systems across air, sea, and land.By generating intelligent, adaptive virtual environments, AILiveSim empowers engineers to test thousands of scenarios at scale, reducing development costs, accelerating innovation, and enhancing safety.Trusted by leading aerospace and maritime organizations in both civil and defense sectors, AILiveSim is shaping the future of autonomy with technology that transforms real-world complexity into smarter, faster, and more reliable solutions.

Related Blogs
Fine-Tune LLMs for Proteins with AMD Enterprise AI Suite — ROCm Blogs
Fine-tune Llama 3.1 8B with ROCm for advanced protein sequence insights in bioinformatics
November 26, 2025
Exploring Gameplay Video Generation with Hunyuan-GameCraft — ROCm Blogs
Learn to generate dynamic, action-controllable gameplay videos from single images using Hunyuan-GameCraft on AMD Instinct MI300X GPUs with ROCm.
November 26, 2025
Now Available: AMD Spartan™ UltraScale+™ FPGA SCU35 Evaluation Kit–An Affordable Platform for Every Developer
Experience flexible connectivity, power efficiency and accelerated FPGA development with the affordable AMD Spartan UltraScale+ FPGA SCU35 Evaluation Kit.
November 25, 2025
Breaking Barriers in Genomics: Slorado Brings Open-Source Genomic Basecalling to AMD GPUs
Slorado enables real-time, open-source nanopore DNA basecalling on AMD GPUs, advancing global genomic analysis with high performance and flexibility.
November 24, 2025
Using Reinforcement Learning to Fix Text in AI-Generated Videos — ROCm Blogs
Demonstrates how Flow-GRPO can be used to fine-tune Wan models to better generate text in videos by covering background, set up and some examples of training runs.
November 24, 2025
The vLLM MoE Playbook: A Practical Guide to TP, DP, PP and Expert Parallelism — ROCm Blogs
Learn how to combine TP, DP, PP, and EP for MoE models. Discover proven strategies to maximize performance on your vLLM deployments.
November 23, 2025
Zyphra Demonstrates Large Scale Training on AMD with ZAYA1
Zyphra successfully trained ZAYA1-base, the first large-scale Mixture-of-Experts (MoE) foundation model trained entirely on an AMD cluster comprised of AMD Instinct GPUs and AMD Pensando Pollara AI NICs.
November 22, 2025
Accelerating AI-Driven Crystalline Materials Design with MatterGen on AMD Instinct MI300X — ROCm Blogs
Learn how to run a generative model for inorganic material design with MatterGen on MI300X.
November 20, 2025



