aiMotive

17/06/2026

Neural rendering. Dynamic actors. Real-time sensors.

Next week at AVTE 2026 (Stuttgart, June 23–25), we're showing exactly what this looks like in practice – live, at booth A204.

aiSim 6 will debut at the event, and we will also bring a live closed-loop HiL demo.

Stop by to see the latest ADAS & AD technologies up close.

11/06/2026

A simulator is only as good as the world it runs in.

5,000+ assets. 50+ maps. 200+ pedestrians. All built in-house – for aiSim.

Here's how we build the environments that make validation credible: https://www.linkedin.com/pulse/building-world-inside-aisim-aim-aisim-kwfif

How we create the 3D environments that make simulation credible A simulator is only as useful as the world it runs in. Sensor models, physics engines, and scenario logic all depend on one thing being right first: the environment.

04/06/2026

Finding the right data is harder than collecting it.

The aiData platform is built to solve exactly that - using AI-powered data curation tools to surface the data that actually matters, fast and efficiently. To demonstrate its capabilities, we curated a training dataset for a roadwork classifier from millions of unlabeled clips, using text, image, map, and tag-based search. We assembled a 10,000-frame dataset, spanning diverse geographies and weather conditions.

And we're just getting started – scenario-level search is coming in 2026.
Think: “bicycle overtaking a car from the right at an intersection” - retrieved from a hand-drawn sketch or a video input.

Read the full case study: https://aimotive.com/w/finding-data-that-matters-ai-powered-dataset-curation-with-the-aidata-platform

01/06/2026

We're proud to share that two aiMotive research papers have been accepted to CVPR 2026 workshops.

1) Hybrid Rendering for Multimodal Autonomous Driving: Merging Neural and Physics-Based Simulation

Accepted to the Simulation for Autonomous Driving workshop. This work explores how neural rendering and physics-based simulation can be combined to create more realistic and scalable multimodal simulation environments for autonomous driving.

Workshop: https://agents4ad.github.io/
Paper: https://openreview.net/forum?id=la73FGCkWO
Poster presentations:
• Room 102/104 — 3:15 PM–4:15 PM
• Room 607 — 4:30 PM–5:30 PM

2) Zero-shot 3D General Obstacle Detection via Multimodal Foundation Models and Geometry

Accepted to the AUTOPILOT workshop. This work investigates how multimodal foundation models and geometric reasoning can enable zero-shot detection of previously unseen obstacles in 3D autonomous driving scenarios.

Workshop: https://www.autopilot-cvpr.net/
Paper: https://bitly.cx/kSZAa
Poster presentations: Hall 3A

A big thank you to everyone on the team and all collaborators involved. If you're at CVPR 2026 in Denver, we'd love to connect.

28/05/2026

"Real-world driving data alone is no longer sufficient to scale validation for next-generation ADAS systems."

aiMotive Professional Services together with STRADVISION built a pipeline that closes that gap – combining SVNet's perception-driven scenario understanding with aiSim's ISO 26262 ASIL-D certified neural simulation and World Extractor's 3D Gaussian Splatting reconstruction.

Fleet recordings in. Certified simulation results out.

Read the full announcement: https://aimotive.com/w/stradvision-and-aimotive-combine-perception-driven-scenario-understanding-and-scalable-simulation-for-adas-validation

21/05/2026

We’re excited to share this next step in our journey.
Looking forward to what’s ahead – more details soon.

https://www.linkedin.com/posts/aimotive-news-stellantis-share-7463213965817499649-F5q9?utm_source=share&utm_medium=member_desktop&rcm=ACoAABMB6J4BfjWCEBHUYkDq-Ycy_l8Vl4hrHWg

14/05/2026

"From day one, we have dedicated ourselves to automotive-native inference. While others adapt general-purpose designs, our focus on the specific rigors of the vehicle environment allows us to provide an unrivaled, safety-certified, and future-proof NPU IP solution."
– Márton Fehér, SVP Semiconductor Engineering

aiWare5 is now available for general licensing.

https://aimotive.com/w/aimotive-announces-aiware5-delivering-unrivalled-flexibility-and-scalability-for-l2-to-l4-automotive-ai-workloads

07/05/2026

We’re happy to share that our paper “CCLSTM: Coupled Convolutional Long-Short Term Memory Network for Occupancy Flow Forecasting” was accepted to CVPR 2026, Workshop on Autonomous Driving (WAD).

Occupancy flow forecasting supports safe autonomous driving by predicting how the space around the vehicle will evolve, enabling the motion planner to anticipate conflicts early, account for uncertainty, and choose trajectories that remain collision-aware in dynamic traffic.

CCLSTM is a lightweight, fully convolutional recurrent model (no transformers, no vectorized inputs) designed for efficient onboard deployment with BEV perception. As of today, it ranks 1st in all metrics on the 2024 Waymo Occupancy and Flow Prediction Challenge leaderboard.

Learn more:
Paper (arXiv): https://arxiv.org/abs/2506.06128
Project page: https://aimotive.com/occupancy-forecasting
WOMD leaderboard: https://waymo.com/open/challenges/2024/occupancy-flow-prediction/
WAD: https://cvpr2026.wad.vision/

If you’re at CVPR 2026 in Denver, we’d love to chat in Room 603 on June 3.

04/05/2026

Excited to be joining Socionext Europe as an ecosystem partner at VECS 2026.

Socionext will showcase a complete autonomous driving closed-loop demo featuring the full aiMotive stack – aiDrive, aiSim, aiData, and aiWare – running on an optimized automotive SoC. It's a strong proof point for what integrated hardware-software co-design looks like in practice.

See you at booth C04:32 in Gothenburg, May 5–6.

28/04/2026

When all our expertise comes together… aiMotive introduces the world’s first closed-loop, interactive neural simulation toolchain and environment for Hardware-in-the-Loop testing.

Use recorded test data and virtualize them overnight to test your AD system in challenging dynamic scenarios with complete multi-sensor simulation and vehicle emulation using real hardware and system components.

An end-to-end portfolio for end-to-end autonomous vehicles.