Best of LinkedIn: SDV & AI-Defined Vehicles CW 45/46 2026

Every carmaker now talks about AI-defined vehicles, and the chips are shipping and the strategies are written. The hard parts are still unsolved: proof, funding, and keeping cars current and compliant after the sale. Chinese OEMs lead the readiness race while established players and many suppliers struggle to turn strategy into position. A software-defined car is judged twice: when it leaves the factory, and every time an update reaches it.
Date
November 19, 2026
Software-Defined & AI-Defined Vehicles
Thomas Allgeyer

Methodology: Every two weeks we collect most relevant posts on LinkedIn for selected topics and create an overall summary only based on these posts. If you´re interested in the single posts behind, you can find them here: https://linktr.ee/thomasallgeyer. Have a great read!

Listen to our podcast

If you prefer listening, check out our podcast summarizing the most relevant insights from SDV & AI-Defined Vehicles CW 45/ 46:

China Leads Readiness Race

  • XPENG and Xiaomi sit at 97% in CAM's Connected Car Readiness ranking, with Tesla, NIO and Li Auto at 93%.
  • Seven of the top ten manufacturers in the 2026 readiness ranking originate from China.
  • Volkswagen, the 2021 to 2025 innovation leader, lands at 47% in the readiness ranking, while BMW Group ranks 10th.
  • Volvo's third-quarter China revenue dropped by nearly 50%, as buyers weigh smart-cockpit software and fast development cycles over foreign badge prestige.
  • Global sales from 2015 to 2025 show BYD at 4.6 million vehicles, nearly reaching Ford's global volume, while Toyota stays above 11 million.

Safety Proof Gets Funded

  • Proving physical AI safe is where many timelines slip, and Qvyon raised $8M to turn recorded sensor data into 3D test scenarios.
  • RAND estimated an autonomous car would need to drive about 11 billion miles to prove it is safer than a human driver.
  • Rare moments such as a car cutting in or a blocked sensor are too dangerous to stage, so simulation fills the evidence gap.
  • Every new operating domain, regulation, market and software release multiplies validation effort, which pushes the case for continuous homologation over manual scenario testing.
  • Accountability frameworks for driverless mistakes remain unsettled, leaving harmed pedestrians and regulators without clear remedies as autonomy scales.

AI Code Needs Verification

  • AI writes code fast, but proving it safe is still slow, and verified evidence matters more than code volume.
  • One written rule in the repo took agent success from 0 of 120 to 117 of 120 when new tasks conflicted with existing code.
  • Written intent next to the code and external gates on the result keep engineers in charge of final decisions.
  • Vehicle autonomy levels offer a model for labeling how much work AI agents can take over.
  • Software complexity delays vehicle launches by 12 to 24 months, so governance has to start at the requirements.

Supplier and Silicon Shifts

  • Qualcomm's automotive revenue grew 89% while handsets fell 26%, as weak smartphone demand pushes chipmakers toward automotive and AI infrastructure.
  • Over 1 in 3 of the mapped SDV suppliers sit within 0.3 points of a quadrant line, which leaves their strategic positions unclear.
  • AI-defined vehicles add another investment cycle for suppliers that have not yet built a sustainable software business.
  • Infineon's AURIX RC1 brings the open RISC-V architecture into real-time vehicle control for SDV and physical AI.
  • Tying software tightly to specific silicon produces SoC-defined vehicles instead of truly hardware-independent ones.

Edge AI and Virtual Sensors

  • Silicon is now powerful enough to run AI where vehicle data is created, which changes how much hardware a car needs.
  • Virtual sensors can replace physical ones, as in virtual headlight leveling and virtual tire pressure monitoring, lowering cost and weight.
  • Quantizing a perception model cut inference latency from 29.0 to 13.7 ms at a cost of 1.7 mAP points.
  • As architectures centralize, reducing complexity and simplifying integration matters more than adding semiconductor performance.
  • Edge AI in vehicles needs deterministic and safe orchestration, not only deployment.

Cars Outlive Their Software

  • Cars built to last 15 to 20 years now run on digital features that age in a few.
  • Volkswagen's ID. Software 3.9.1 brings battery preconditioning and thermal management updates to older ID. models already on the road.
  • A premium car bought in 2026 may still drive well in 2036 while its apps, AI assistant and connected services fade.
  • Features such as Android Auto video depend on each car's infotainment system, so older cars may need extra testing and certification work.
  • Over-the-air updates split into FOTA for ECU firmware and SOTA for applications, and each carries distinct validation and failure risks.

Cockpit and Voice AI

  • Mercedes-Benz splits voice AI between in-vehicle processing and the cloud to balance latency, privacy and offline reliability.
  • The automotive digital cockpit market is forecast to reach $46.26B by 2032, driven by centralized computing and AI-enabled functions.
  • Next-generation in-car AI is expected to understand context inside and outside the car and assist occupants proactively.
  • A visually polished interface still fails if information placement and interaction design do not suit real driving use.
  • One Qualcomm-linked demo showed five screens, digital cockpit, vision and AI detection running from a single unit.

Privacy and Cybersecurity Rules

  • GM's $220 million privacy-related settlement drew a quick response from Ford in the form of its Privacy Promise.
  • ISO/SAE TR 8477:2026 gives practical guidance on how to run cybersecurity verification and validation across the vehicle lifecycle.
  • The connected car security market is described as a $6.99 Bn opportunity, driven by V2X, regulation and centralized SDV architectures.
  • Regulations such as UNECE WP.29 R155 and R156 require continuous monitoring, pushing OEMs toward end-to-end cybersecurity.
  • Cybersecurity thinking is shifting from protecting individual components to protecting the trusted behavior of the whole vehicle.
Subscribe to newsletter

Subscribe to receive the latest blog posts to your inbox every week.

Please confirm your GDPR consent to join our mailing list.
By subscribing you agree to with our Privacy Policy.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Frequently asked questions

What is the big shift from software-defined to AI-defined vehicles?

The shift is from building the technology to proving it and keeping it current after the sale. Chips are shipping and strategies are written, but safety proof, funding and compliance are unsolved. Chinese OEMs lead readiness, with XPENG and Xiaomi at 97% while Volkswagen sits at 47%.

How are companies proving that physical AI and autonomous systems are safe?

Teams are turning recorded sensor data into simulated test scenarios because road testing alone cannot supply enough evidence. RAND estimated about 11 billion miles would be needed to prove an autonomous car safer than a human driver. Qvyon raised $8M to build this kind of testing.

Why is the software in cars becoming a problem after the sale?

Cars are built to last 15 to 20 years, but digital features age in a few. A car bought in 2026 may still drive well in 2036 while its apps and AI assistant lose support. Volkswagen's ID. Software 3.9.1 update for older ID. models shows one response.

Want to see the posts voices behind this summary?

This week’s roundup (CW 45/ 46) brings you the Best of LinkedIn on SDV & AI-Defined Vehicles:

→ 70 handpicked posts that cut through the noise

→ 39 fresh voices worth following

→ 1 deep dive you don’t want to miss