Drivers now use AI every day on their phones. So why is it still so hard to bring the same experience into the car, especially in vehicles already in production?

In this session at ELIV 2026, Sriram G explains why relying on the cloud or adding more powerful AI hardware isn’t always the answer. He also presents a third pathway developed at KPIT, where AI runs entirely in the vehicle by making better use of the compute that’s already there.

Watch the session to see

  • Why a more powerful AI chip doesn’t always mean better performance, and what KPIT’s own experiments found
  • How the phone in your pocket can become part of the car’s AI system
  • A live demo of a voice assistant on a series-production head unit, with the language model running on a phone and no connectivity required
  • How critical and non-critical AI workloads are prioritised within the vehicle
  • Why this approach matters for emerging markets like India, where cost is a key factor

Your feedback form has been submitted successfully!