Embedded AI will reach a share of 50%

Markt&Technik: Mr. Gromov, how do you define embedded AI?

Viacheslav Gromov: At Aitad, we differentiate between cloud AI – which is abstract and generates continuous costs, but can be scaled differently – and edge AI, which allows partial pre-processing of data on site and further processing, for example using algorithms in the cloud. We ourselves go one step further: even though central AI (in this case: cloud AI) was the origin of the overall technology and edge AI has been in use for around a dozen years, we draw another sharp line between it and embedded AI. In contrast to edge AI, which is to be understood more as a hybrid system with partial processing on site, we see embedded AI as an almost completely decoupled system that does not require a back end or external dependencies – all data processing therefore takes place on site, usually within the sensor component.

It is important to mention that we do not view the interaction of AI technologies as opposing, but as a symbiosis. Just like in nature: in humans, for example, not everything happens centrally in the head, but also decentrally. Consequently, there are tasks that need to be real-time capable and decoupled under enormous amounts of data, while abstract systems are also needed for broader analyses and decisions.

How do you assess the current market for embedded AI?

Embedded AI has been usable in practice for a good three years and is currently in the early adopter phase – in other words, it is the latest trend in the field of AI solutions. More and more – but so far only a fraction – of manufacturing companies are starting to look into using it. This is due, among other things, to the recent changes in the general conditions caused by global competition and the need for transformation. The media presence, starting with ChatGPT almost two years ago, is also ensuring that artificial intelligence and its use are increasingly coming into focus.

In an exclusive interview with Markt & Technik, CEO Viacheslav Gromov talks openly about the development opportunities of embedded AI and highlights a wide range of use cases in industry, automotive, smart home and energy efficiency.

Read the entire interview at elektroniknet.de (in german)