Realtime AI News
NVIDIA launches open-source Alpamayo 2 Super model for autonomous vehicles
NVIDIA has announced Alpamayo 2 Super, an open-source model for autonomous vehicles that is now available for commercial use under an open commercial license, offering leading multi-task autonomous-driving reasoning capabilities. Chinese intelligent-driving concept stocks rallied on the news on August 11, with Dazhong Transportation hitting the daily limit.
NVIDIA has announced the open-source Alpamayo 2 Super model for autonomous vehicles, now available for commercial use under an open commercial license and offering leading multi-task autonomous-driving reasoning capabilities. Sina Finance reported the news, citing market research outlet Caiwen.
Alpamayo 2 Super is the latest addition to NVIDIA's Alpamayo family, an open reasoning vision-language-action (VLA) model line for autonomous driving. The first-generation Alpamayo 1 debuted at CES in January as a large teacher model that developers can fine-tune, distill and test.
The announcement sent Chinese intelligent-driving concept stocks higher on August 11: Dazhong Transportation hit the daily limit, while Beilu Zhikong, Zhejiang Shibao, Jinjiang Online, Wanji Technology and Suoling shares all rallied.
The market read the open-source autonomous-driving model as a positive signal for the supply chain. An open commercial license means automakers, Tier 1 suppliers and research teams can directly build end-to-end intelligent driving solutions on top of the model, lowering development barriers.
Technically, the Alpamayo line focuses on reasoning-based autonomous driving, using chain-of-thought reasoning to handle rare long-tail scenarios that traditional perception-planning architectures struggle to cover. Open commercial licensing lets developers fine-tune, distill and test on the model to accelerate safe, reliable reasoning-based ADAS deployment.
What to watch: NVIDIA has yet to disclose Alpamayo 2 Super's parameter scale and release channels. The first-generation Alpamayo 1 shipped with companion simulation tools and datasets, and whether the company repeats that playbook is worth watching.
For China's intelligent-driving supply chain, the push toward open-source autonomous-driving models will shape the competitive landscape of end-to-end solutions, which helps explain the swift market reaction. The next question is whether the model reaches production vehicles and how much it shifts the technical paths of domestic players.
Why it matters
Opening a commercially licensed autonomous-driving reasoning model could accelerate end-to-end intelligent driving adoption and reshape the technology roadmaps of Chinese players. The immediate stock-market surge shows the signal was priced in fast, making production deployment the key variable to watch.
Nearby Updates
All08/11, 23:31
AI startup Manus to resume independent operations as deal with Meta unwinds
Reuters reports that AI startup Manus will resume independent operations as its deal with Meta unwinds. The split returns the agent-focused startup to standalone status and leaves it free to set its own product and partnership strategy.
08/12, 00:25
An unreleased Anthropic model made surprising progress on the Riemann hypothesis, TechCrunch reports
TechCrunch reports that an unreleased Anthropic model made more progress than expected on the Riemann hypothesis, one of mathematics' biggest unsolved problems, though Anthropic has not solved it. The report signals that frontier mathematical reasoning is still improving quickly, and raises questions about when the model will ship.
08/11, 21:57
Ant Group leads strategic round in Daimon Robotics, first tactile bet as startup unveils world-first ‘physical interaction brain’
Daimon Robotics said on August 11 it closed a strategic funding round worth hundreds of millions of yuan led by Ant Group, marking Ant's first move into the tactile sensing layer of embodied AI. The company also released Daimon-TWM, a 10B-parameter tactile-grounded world model it calls the world's first “physical interaction brain.”
08/12, 00:45
Alibaba unveils Wan3.0 with 30-second video generation and multimodal inputs
Alibaba has unveiled Wan3.0, its video generation model that supports up to 30 seconds of output and multimodal inputs, according to Pokde.Net. The release extends Alibaba's push into long-form, controllable video generation and intensifies competition in the field.