Autonomous Driving

Today's Stories

1. 1

   Toyota to launch Level 2++ automated driving by 2028

   Toyota announced on August 27 that it will bring to market passenger cars with an AI-based advanced automated driving system around 2028. The system is expected to enable 'Level 2++' hands-off driving in urban areas. Current automated driving is mostly Level 2, which still requires driver supervision. Toyota aims to use AI learning to get closer to Level 3, where the driver is not needed in certain conditions, though liability for accidents remains with the driver as in Level 2.

   Toyota plans to expand the system to more models from 2030 onward. In commercial vehicles, it aims to begin nationwide rollout of Level 4, where no human is involved under specific conditions, also after 2030.

2. 2

   Ford adopts Apple Maps data for next-gen smart driving

   Ford Motor Company will adopt Apple's MapKit for Automotive software development kit (SDK) for its universal EV platform, and feed Apple Maps data to its Latitude AI team, which works on the next-gen BlueCruise advanced driver assistance system. The move pairs Apple's mapping data with Ford's BlueCruise development, suggesting a deeper tie-up between the two companies in smart driving technology. The collaboration could shape how Ford integrates navigation and autonomous features across its EV lineup.

   How Ford's Latitude AI team uses Apple Maps data to enhance BlueCruise features, and whether this signals broader adoption of Apple's automotive SDK across more Ford models.

3. 3

   Kodiak AI, AMD partner on driverless truck compute

   Kodiak AI, a provider of AI-powered autonomous driving tech, has partnered with chipmaker AMD to use AMD EPYC Series processors in its seventh-generation autonomous truck platform. Kodiak is the first company to deploy these processors in a driverless trucking hardware platform. Driverless trucks need high-performance central processing units (CPUs) to quickly aggregate and preprocess sensor data from LiDARs, cameras, and radars, and to support path planning. The AMD EPYC CPUs offer higher clock speeds than the previous generation, enabling faster data processing at lower power consumption.

   Kodiak is working toward launching driverless trucks on public highways later this year. Its fleet in the Permian Basin already operates with no humans in the cab today.

4. 4

   Tesla swaps Grok for ByteDance's Doubao LLM in China

   Tesla is setting aside its Grok model in China and will use ByteDance's localized large language model (LLM), Doubao, in its smart cockpits. Regulatory reality in China forced Tesla to pivot from Grok, even as it pushes into Robotaxis and Optimus humanoid robots and builds its own cloud service provider (CSP) capabilities.

   The move underscores how local AI models are becoming essential for automakers operating in China, though terms and rollout details were not disclosed.

5. 5

   Tesla's AI5 chip claims 2-3x Nvidia efficiency

   Tesla CEO Elon Musk said the company's upcoming AI5 inference chip could deliver two to three times the performance per watt of Nvidia products in autonomous vehicles and Optimus humanoid robots, at around 10% of the cost. This suggests Tesla is doubling down on custom AI chips to reduce reliance on Nvidia, potentially lowering costs and improving efficiency for its AI-driven products, though the claims are as yet unverified.

   Whether the AI5 chip's performance and cost advantages materialize when deployed, and how Nvidia responds in the competitive AI chip market.

6. 6

   Lockheed Martin expands into AI airspace monitoring and sea launch

   Lockheed Martin (LMT) is expanding into AI airspace monitoring and sea launch operations. This move broadens the defense contractor's portfolio into two emerging areas, potentially opening new revenue streams and strategic capabilities.

   Details on the scale, timeline, and specific technologies involved have not been disclosed yet.