\u201cAny button in the car that controls something has at least one chip and a significant amount of software behind it. The amount of software is continuously increasing. Four years ago, we first mentioned this industry transformation and the shift from distributed architectures toward centralized architectures.<\/p>\n\n\n\n
Every year, we make progress in this chip industry. The industry has evolved so much that it is now very clear the transition has already been made toward those centralized architectures, and we are seeing a series of improvements.<\/p>\n\n\n\n
Together with our customers\u2014who may be OEMs or Tier 1 suppliers\u2014we contribute to the development of these architectures. Companies like ours no longer produce just chips, but integrated solutions,\u201d said Simona Almajan, Country Manager, NXP Semiconductors Romania, during the Automotive Forum 2025 organized by Automotive Today and The Diplomat-Bucharest.<\/em><\/strong><\/p>\n\n\n\n Key statements:<\/strong><\/p>\n\n\n\n \u2022 We implement relevant use cases coming from OEMs, helping with integration. We are talking about simplification, about abstracting the complexity underlying these modern architectures. And, of course, a very important aspect here is the trade-off between hardware and software: what we put in software and what we put in hardware so we can work faster and have a more efficient and modular architecture.<\/p>\n\n\n\n \u2022 We continue to improve battery performance systems, for example. We use AI for better data collection and better data classification, so that at the battery level we can determine its state-of-health.<\/p>\n\n\n\n \u2022 On the radar side, I also want to make the connection with AI, because it is not just a buzzword\u2014it is something that is happening. We all know that a car is surrounded by sensors and makes a series of decisions based on what it \u201csees,\u201d right?<\/p>\n\n\n\n \u2022 We must determine how accurate the data is so the car can make the right decision. And this is not only about data collection; it is about data classification, about the accuracy of the models used, so we can rely on them.<\/p>\n\n\n\n \u2022 Functional safety today is not what we knew 3\u20134 years ago; we are now talking about new challenges. It is about functional safety in the presence of AI and how much trust we can place in those predictive models based on AI.<\/p>\n\n\n\n \u2022 We continue developing increasingly intelligent systems and investing more in preventing failures and detecting them. I would say the biggest evolution is not in functional safety, but in cybersecurity. Why cybersecurity? The more connected the car becomes and the more interfaces it has, the larger the attack surface. A larger attack surface means a higher likelihood of attack.<\/p>\n\n\n\n