NVIDIA DRIVE
@NVIDIADRIVE
Followers
16K
Following
991
Media
1K
Statuses
2K
Transforming transportation with #AI. Follow us for the latest in #autonomousvehicle development and news from our ecosystem.
Santa Clara, CA
Joined August 2016
0
0
0
Just 3 days until IAA Mobility 2025.๐ฆ. Weโll be on stage sharing how:.๐ Cloud-to-car AI is redefining safety + autonomy.๐ญ Industrial AI + digital twins are transforming automotive manufacturing.๐ค Generative + agentic AI are shaping the future of automotive innovation. Donโt
2
0
2
๐ Smarter AV validation starts with better scenarios. See how @nvidiaomniverse NuRec + NVIDIA Cosmos on @ForetellixHQ Foretify turn raw driving logs into realistic, scalable digital twins for safer AV training.
๐ ๐
๐ซ๐จ๐ฆ ๐๐ซ๐ข๐ฏ๐ ๐ฅ๐จ๐ ๐ฌ ๐ญ๐จ ๐ฌ๐ข๐ฆ๐ฎ๐ฅ๐๐ญ๐ข๐จ๐ง-๐ซ๐๐๐๐ฒ ๐๐ข๐ ๐ข๐ญ๐๐ฅ ๐ญ๐ฐ๐ข๐ง๐ฌ, ๐ง๐๐ฎ๐ซ๐๐ฅ ๐ซ๐๐๐จ๐ง๐ฌ๐ญ๐ซ๐ฎ๐๐ญ๐ข๐จ๐ง ๐ข๐ฌ ๐๐ก๐๐ง๐ ๐ข๐ง๐ ๐ก๐จ๐ฐ ๐ญ๐ก๐ ๐ข๐ง๐๐ฎ๐ฌ๐ญ๐ซ๐ฒ ๐๐ฎ๐ข๐ฅ๐๐ฌ ๐๐ ๐ฌ๐๐๐ง๐๐ซ๐ข๐จ๐ฌ. Join Foretellixโs webinar with Gianmarco Macaro to
1
1
5
@BotAutoAV As @BotAutoAV CEO Xiaodi Hou puts it:.โNVIDIA DGX B200 is the cornerstone of our AI infrastructure. Its performance enables us to train large-scale foundation models with high efficiency and rapid iteration, a critical advantage in developing real-time autonomy systems."". ๐.
0
1
3
๐ Want to see how @BotAutoAV is using NVIDIA DGX B200 to power real-time autonomy?. ๐๏ธ Training on DGX B200 delivered 70% faster performance vs. H100 from a cloud vendor. ๐ Scaling multimodal data โ from camera to LiDAR โ with NVLink + RDMA. โก Deploying faster with end-to-end
4
9
40
๐ AV dev just got a serious upgrade. The DRIVE AGX Thor Developer Kit is powered by Blackwell GPUs, next-gen Arm CPUs, and the advanced DriveOS 7 software stack. What it enables:.โ
Faster prototyping, validation, and deployment of AV apps.โ
In-vehicle LLMs + VLAs with
0
8
33
๐ Session: Generative and Agentic AI: Driving the Future of Automotive Innovation.Sri Subramanian | Global Head of Generative AI, Automotive | Sept. 9 @ 10:30 a.m. CEST. ๐
Save the date:
0
2
7
๐ Session: Industrial AI in Auto: Digital Twins, Simulation, and Robotics in Automotive.Norm Marks | VP, Automotive Enterprise | Sept. 9 @ 1:15 p.m. CEST. ๐
Save the date:
1
3
6
Join us at IAA Mobility 2025 (Sept 9โ14) to see how we are accelerating the future of safe, intelligent mobility. ๐๐ค. ๐ Keynote: From Cloud to Car: Driving the Future of Safe, AI-Defined Mobility.Ali Kani | VP, Automotive | Sept. 9 @ 12:30 p.m. CEST . ๐
Save the date:
3
4
18
RT @WeRide_ai: Congrats @nvidia on the @NVIDIADRIVE AGX Thor Developer Kit launch! Our #HPC 3.0 platform built with @Lenovo & dual DRIVE AGโฆ.
0
5
0
Neural recon = real world in sim. Check out how NVIDIA Omniverse NuRec turns sensor data into photorealistic 3D scenes for safer, smarter AVs. ๐โจ. ๐.
What's neural reconstruction got to do with physical AI? ๐ค. NVIDIA Omniverse NuRec libraries bring the real world into interactive simulation, transforming sensor data into 3D Gaussian-based scenes that can be replayed and altered, accelerating testing and validation. Watch the
0
2
16
NVIDIA DRIVE AGX Thor DevKit is here. ๐. With the new DRIVE AGX Thor developer kit + DriveOS 7.0, developers can build safer, smarter, and more secure AVs built on NVIDIA Blackwell architecture. Available for pre-order today ๐
7
26
129
โก Petabytes? No problem. Beamrโs content-adaptive technology, accelerated by NVIDIA, cuts up to half the size of AV video data while protecting every detail AI needs for safe driving. Less storage. Less cost. Same accuracy.
Here's what we validated with NVIDIA's AV team: . ๐ 20-50% video data compression improvement .๐ Zero impact on AI model accuracy. Here's why this matters: if you're managing massive video repositories .>>
1
8
19
Tensor unveiled the world's first personal Robocar, accelerated by NVIDIA. Tensorโs Robocar is built on NVIDIAโs cloud-to-car AV platform, trained with NVIDIA datacenter GPUs and running on eight DRIVE AGX Thor SoCs with Blackwell GPU architecture, delivering 8,000 TOPS of AI
8
37
125
Render real world drives in interactive simulation with @nvidiaomniverse NuRec, now integrated with @carlasimulator.
Bring the real world into simulationโinstantly. ๐. The newly announced NVIDIA Omniverse NuRec libraries let you reconstruct photorealistic 3D scenes from sensor data in minutes and directly into #NVIDIAIsaac Sim or @carlasimulator. ๐. Powered by #GaussianSplatting and
0
2
21
High-quality data is critical for safe, scalable AV development. By integrating @nvidiaomniverse NuRec with @Voxel51's data engine, AV developers can streamline dataset ingestion, validation, and preparationโcatching costly errors early and accelerating simulation workflows.
@nvidiaomniverse NuRec and @Voxel51 FiftyOne integration tackles a major bottleneck in AV development when working with multi-sensor datasets. As AV systems move from R&D to deployment, issues like misaligned sensor calibrations, drifting ego-poses, and timestamps often introduce
0
2
8
Transform real-world AV logs into simulation-ready, sensor-accurate 3D environments at scale with the Foretellix Foretify toolchain, integrated with @nvidiaomniverse NuRec neural reconstruction libraries for faster, more robust AI model training and validation.
Foretellix's Gianmarco Macaro shares a how-to for transforming raw drive logs into simulation-ready scenarios with neural reconstruction via two integration options from either NVIDIA or Parallel Domain. Learn how to:.1. Turn raw AV logs into fully rendered 3D environments with
0
2
8
NVIDIA Omniverse NuRec with #GaussianSplatting is bringing the real world into simulation for safe, scalable autonomous vehicle development. NuRec libraries are integrated into leading AV toolchains from @ForetellixHQ and @Voxel51, and NuRec rendering is available to try today.
.@nvidia technologies are taking robotics to the next level. ๐ฆพ. New #NVIDIAOmniverse libraries, #NVIDIACosmos world foundation models, and #NVIDIARobotics simulation frameworks let developers: . โ๏ธ Build physically accurate digital twins.โ๏ธ Capture the real world in simulation
2
0
4