Indian EVs were blowing up because of cheap laptop chips in battery: NXP tech chief Lars Reger

Lars Reger, chief technology officer of NXP

When you think about the connection between a European high-tech company and India, the intuitive assumption would the vast market potential offered by India's status as the fastest-growing major economy.

However, this assumption doesn't hold true for NXP, a Dutch semiconductor chip design company. Spun off from electronics giant Philips in 2006, NXP has instead embarked on an acquisition spree in India to bolster its engineering division.

“We are very research heavy in India with presence in four sites across Delhi, Bengaluru, Pune and Hyderabad. We have in the range of 3,500-4,000 people here in India, and that's mainly in engineering. One third of our engineering population is here,” NXP chief technology officer Lars Reger said.

About 55 percent of NXP's revenue comes from automotive chips, around 25 percent from industrial use-cases and the rest is due to segments like payments and mobile.

In an interview with Moneycontrol, the chief technology officer of NXP talked about India's growing heft in chip design, how electric vehicles blow up because of using the wrong chips in their batteries, the difficulties of making autonomous vehicles and how Wall Street financial analysts hyped them up over the past few years.

Edited excerpts:

India is said to be the biggest supplier of chip designers to the world. Yet, there is a criticism from some quarters that it is a back office and not an innovator. What do you feel?

For the Western world, say 15 years ago, India was mainly seen as an extended workbench for research and development. That has of course changed big time. Today, all our most advanced products come out of India. Our most advanced chip for automotive networks, which is the size of a thumbnail and has 4 billion transistors, is being developed in Noida. It is the world's first automotive chip with a node size of 5 nanometres.

Earlier, India was known mainly for good software programming. The products were designed somewhere in the western world and then when you needed to change a few things for a derivative product, you gave it to India so that it could be done at a lower cost. That perception has totally changed in the last couple of years as the most advanced stuff is not doable anymore without India.

One of the biggest issues in electric vehicles in India has been that of scooters catching fire because of faulty battery management systems (BMS). As you are making BMS chips, what is your diagnosis of the problem?

What these companies have done is they have taken cheap solutions from laptops into complicated battery management systems. My laptop is here in this room. But it never gets monsoon rain, the New Delhi sun, or minus 20 degrees temperature in northern Germany winters. Therefore, do not use consumer electronics in battery management systems.

Did that really happen?

They used the cheapest solutions. And, it was the same failure each time. The India market has to be super, super cheap, and they cannot afford a little bit more expensive battery management chips. Now, the cost of the battery management chips, as compared to the entire scooter or car that you are building, is a joke. It's a fraction. Even if you double the price of these chips, you will not really recognise it on the sales price of the entire vehicle. A lot of people made the beginner's mistakes of taking cheap consumer solutions.

Apart from EVs, the other big futuristic shift that is expected in the automotive industry is self-driving or autonomous vehicles. How far are we from that in reality?

In self-driving or driver replacement, there are five levels. Level zero is manual driving and level five is when there are no steering wheels, when my kids can go to the kindergarten without me.

The focus today is on level three which is the highway pilot mode. You have to drive manually in a busy city, but the robot can take over on a highway where traffic is less complex. And these cars are reaching the market in the meantime. The BMW seven series and the Mercedes S Class are already level three.

Apple has recently shut down its autonomous car division which led to a view that self-driving cars are still far away from reality. Do you think the tech industry has over promised and under delivered in this regard?

I don't think that is the conclusion. Of course, I can't speak on behalf of them, but there is certainly an overall learning in the market.

In 2018, journalists were asking me how long is NXP going to survive. And I was like, 'What exactly do you mean?' They said, 'Well, we have all these articles out that by 2020, half of the cars will be self driving.' Now, if you ask me about how the first crazy story came into play, the answer simply is the stock market. If I tell you that I can revolutionise the entire market because I can build a taxi company and remove the drivers, then you take your pencil and say okay, drivers are 30 percent of the cost of the entire operation.

If you can make Uber 30 percent cheaper with robo taxis, then Wall Street throws money at you until it sticks. So the financial analysts had the dream that is going to make them filthy rich. And then, suddenly, reality kicked in. That is not how robots are built. It is complicated and takes longer. The story was mainly driven by the truck and taxi fleet companies.

What more developments in artificial intelligence (AI) are required for a fully autonomous car?

A car is as much a brain on wheels as you are a brain on shoes. There is an entire body in between various different functional layers. In other words, as a biological robot, you have a spine that helps you with reflexes, when you stumble and quickly straighten your legs. This is what real time electronics can do. That has nothing to do with AI and machine learning.

For breathing, heart rate and temperature control in your body, you have the cerebellum. And only on top of that, you have a creative center that can write love letters, paint pictures, sing songs, play piano, and so on. That is our main brain. For transportation, you need the brainstem and the cerebellum. Look at insects. An ant can do transportation without falling off the table without a main brain. My car doesn't need to write me a love letter. But it has to have certain levels of creativity if it gets stuck.

You don't need AI for rules based driving such as keep on the left side, don't overspeed and stop at traffic lights. But you need creativity to tell you that waiting three hours behind a parked car is a bad idea.

Are you using generative AI to write code or design chips at NXP?

The answer clearly is that we are taking the first steps. And only the first steps because you have to be careful, like hell, that you are not sharing your intellectual property with the rest of the world. So, having an understanding of how these tools work and how the database is suddenly visible to everyone is important.