Autonomous vehicles

1. 1. Tesla vs the rest?

   → While Tesla’s AVs use cameras and computer vision, alternative approaches use Lidar (a laser-beaming depth and distance sensing system) and HD maps to create a clearer 360-degree view.

   → In a capital-intensive industry, startups must choose whether to improve the mapping process, or attempt to lower the cost of Lidar — the most expensive component.

2. 2. Democratising driving data

   → Deep learning — AI that imitates the brain’s neural networks to assess a vehicle’s surroundings — is a vital but extremely complex part of the AV tech stack.

   → Advancements in inter-fleet communications will enable AVs to relay information to each other and “learn” at a faster rate, but incentivising competing developers to share remains a roadblock.

3. 3. Choose your lane

   → With global tech giants hell-bent on overhauling mobility, there may be an opportunity for startups closer to home to get ahead.

   → Smaller players addressing autonomous freight trucking and last-mile delivery, carrying less precious cargo than humans, might hold the keys to convincing sceptics that AVs are able to deliver after all.

4. 4. No one-size-fits-all solution

   → Self-driving cars won’t fix congestion issues overnight and may even mean more vehicles on the road. In densely populated cities, driverless shuttles or robotaxis are a more efficient option.

   → Expect some brake-slamming as policymakers prepare the necessary infrastructure: insurers are one of the groups with the most to lose, with younger generations paying high car insurance premiums embracing new tech fastest.

Who early stage startups are up against