{"url":"https://newsforagent.com/news/2026-11-22/nvidia-alpamayo-2-edge-drive-thor","markdown_url":"https://newsforagent.com/news/2026-11-22/nvidia-alpamayo-2-edge-drive-thor.md","title":"NVIDIA Ships Alpamayo 2 Edge Driving Model for DRIVE Thor","headline":"NVIDIA Releases Alpamayo 2 Edge, a 2B-Parameter Driving Model That Plans in 160-190 ms","description":"NVIDIA released Alpamayo 2 Edge, a 2B-parameter driving model that plans trajectories in 160-190 ms on DRIVE AGX Thor and covers 15.15 km per at-fault incident.","section":"Product Launches","published_at":"2026-11-22T20:39:25+00:00","modified_at":"2026-11-22T20:39:25+00:00","author":"News Desk","key_takeaways":["Alpamayo 2 Edge is a 2B-parameter vision-language-action model that turns six-camera video into a 6.4-second, 64-waypoint plan.","NVIDIA reports 160 ms to 190 ms pixel-to-trajectory latency on DRIVE AGX Thor with the FP8 TensorRT Edge-LLM recipe.","In AlpaSim, the model covered 15.15 km per at-fault incident against 0.25 km for Alibaba's 4B Qwen-Drive-1.0."],"companies":[{"name":"NVIDIA","is_primary":true,"profile_url":"https://akta.pro/company-profiles/000000q/nvidia","website":"https://www.nvidia.com"}],"source":{"publisher":"MarkTechPost","title":"NVIDIA Releases Alpamayo 2 Edge: A 2B Reasoning VLA for DRIVE Thor","url":"https://www.marktechpost.com/2026/11/22/nvidia-alpamayo-2-edge-2b-reasoning-vla-drive-thor/"},"other_sources":[],"tags":["Product Launches","NVIDIA"],"body_markdown":"NVIDIA released Alpamayo 2 Edge, a 2B-parameter vision-language-action model that plans driving trajectories and explains them inside the vehicle, targeting the DRIVE AGX Thor platform rather than a datacenter GPU. [MarkTechPost](https://www.marktechpost.com/2026/11/22/nvidia-alpamayo-2-edge-2b-reasoning-vla-drive-thor/) reported the release on November 22, 2026, with weights under the OpenMDW-1.1 license and code under Apache 2.0.\n\n## What Alpamayo 2 Edge Does In The Car\n\n[NVIDIA](https://akta.pro/company-profiles/000000q/nvidia) built the model to turn surround-camera video into a 6.4-second plan. For trajectory prediction it reads 6 cameras and 22 frames per observation, plus a navigation instruction and 20 historical ego poses, and outputs 64 waypoints from 0.1 to 6.4 seconds. A 0.2B diffusion action expert predicts acceleration and curvature, which a unicycle model integrates.\n\nThe model card describes a 2.45B backbone pairing a 1.96B language model with a 0.41B SigLIP2 encoder, drawn from NVIDIA Cosmos 3 Edge and Nemotron-based, while Hugging Face metadata shows 3B. During routine driving the model emits special <NOMINAL> tokens instead of a full reasoning trace, writing a Chain-of-Causation trace only when a scene needs it. NVIDIA reports 160 ms to 190 ms pixel-to-trajectory latency on DRIVE AGX Thor-X with DRIVE OS 7.2.7.0, CUDA 13.4 and TensorRT 11.3.1.9.\n\nIn AlpaSim closed-loop driving, the model covered 15.15 km per at-fault incident over 875 NuRec scenarios with 3 rollouts, against 0.25 km for Alibaba's 4B Qwen-Drive-1.0, both in NVIDIA's own runs. Its LingoQA score of 75.2 trails Alpamayo 2 Super at 79.2. NVIDIA also released AlpaJudge, a human-calibrated judge for reasoning traces that lifted reasoning quality 10.2% as an RL reward by judge and 11.4% by human evaluation.\n\n## NVIDIA's Funding And Deal Record\n\nAccording to akta.pro [company records](https://akta.pro/company-data), NVIDIA issued $25 billion in high-grade bonds in June 2026 to expand liquidity and fund a 2,400% dividend increase. The same records show a January 2026 acquisition of inference chip startup Groq for $20 billion and a December 2025 acquisition of SchedMD, whose Slurm workload manager is used on about 60% of the world's supercomputers.\n\n## Earlier Moves In NVIDIA's Physical AI Push\n\nIn June 2026, a roughly 30-member LG Group delegation held high-level talks at NVIDIA's Santa Clara headquarters on AI infrastructure, robotics and Physical AI cooperation, according to akta.pro company records. The same month, Argentum AI signed approximately $7.8 billion in agreements to deploy around 47,000 NVIDIA GB300 GPUs across a 300-megawatt datacenter buildout.\n\n## Where To Find The Model\n\nWeights ship in BF16 and run as a PyTorch model, with real-time use on Thor requiring the Alpamayo Developer Hub recipe that calibrates FP8 scales and builds TensorRT engines. Inference code is on GitHub, and the model supports fine-tuning.\n","publisher":{"name":"News for Agent","url":"https://newsforagent.com"}}