»If you take a breather – you have already lost!«

»We have to be the ones who understand end-systems in detail«

To what extent is Infineon affected by the effects of Big Data and AI, to what extent will Infineon react with its own developments?

We are participating, for example, in the European High-Performance Computing Joint Undertaking (EuroHPC JU) where we are responsible for the embedded part. For us, autonomous driving is the entry-level topic for the whole thing. At Level 3, the challenges can still be solved with conventional architectures. At levels 4 and 5 we have to go beyond that.

But there are more AI applications than only autonomous driving.

Actually a lot is going on in this arena – for example in the sensor field, in speech recognition and gas analysis. We are therefore working at full speed on the development of new types of sensors. This initially concerns our Power Management & Multimarket business unit, PMM for short. A very important subject in this environment is sensor fusion. The cooperation with XMOS should also be seen in this context. We have combined radar and microphone to improve speech recognition. Using beamforming, the microphones focus on people who are now recognized as such. It could be a thing of the past for language assistants to accidentally invite unwanted people to a party or for the television set in the background to order hundreds of dollhouses.

In the field of AI we are developing many new methods of learning programming. Our engineers in the Innovation Lab in Silicon Valley and in the new development center we are setting up in Dresden are also working on this. What matters here is to approach the subject in an explorative way, like an expedition breaking new ground.

Will this also affect organizational structures? Up to now there were four segments: Automotive, Industrial Power Control, Power Management & Multimarket and Digital Security Solutions.

It all depends on the system concept. We can no longer just sell components that users then somehow integrate into their systems. We have to be the ones who understand end-systems in detail and develop concrete solutions based on the resulting customer requirements. This is only possible if we work together beyond segment boundaries. This is why we have defined lead applications. In robotics, for example, products from our four segments are used in parallel, from MOSFETs to power ICs, safety ICs, sensors and controllers. Especially in the area of emerging robot companies, we have noticed that they prefer to concentrate their own efforts on software and algorithms and are looking for a partner in the hardware area. For each lead application one segment assumes responsibility. For robotics it is IPC, other lead applications include car security, smart homes and drones.

What do you hope to gain from this approach?

Within these lead applications, all the parties involved gain new experience and jointly explore how the different chips and software influence each other and lead towards new system approaches for networked devices, IoT and AI. Here we also work closely with Start-ups, which again provides interesting impulses for both sides. To expand on this example: A robot Start-Up deals with programming the systems and providing them with their own algorithms. This results in a competitive advantage, not from the assembly of components.

For time reasons alone, Start-ups cannot concern themselves with our complex ICs in detail nor are they experts in power electronics. Working together with Start-ups is completely differently to working with our typical large and established customers - we basically deliver instant solutions, stir and it’s done!

Matchmaker+ Anbieter zum Thema

zu Matchmaker+

1. »If you take a breather – you have already lost!«
2. »We have to be the ones who understand end-systems in detail«
3. Aurix 3 generation
4. »We need people who can think in superordinate contexts«