All Electric Society – from pure electricity supply to energy orchestration

In an increasingly electrified world, from e-mobility through renewable energies to data centres and industrial drives, wide-bandgap power electronics enable higher efficiency, smaller systems and therefore lower material and energy consumption across the entire lifecycle.

The most important trends include the broad commercialisation of silicon carbide (SiC) and gallium nitride (GaN) in inverters, chargers and DC/DC converters, the integration of WBG components into modular power electronics platforms, and the coupling with digital control and AI optimisation methods. In parallel, requirements for reliability, standardisation and lifecycle-oriented environmental assessments are rising, for example in the context of PV systems, wind power or railway power systems.

In semiconductor technology, the current focus is above all on scaling SiC substrates to larger wafer diameters, improving GaN-on-Si processes for cost-effective mass production, and developing ultra-wide-bandgap materials such as gallium oxide. In addition, vertical GaN power devices, improved gate driver architectures and thermally optimised package concepts are gaining importance in order to exploit the advantages of WBG semiconductors fully in high-power applications too.

In the All Electric Society, edge computing enablesdata processing directly at decentralised nodes such as charging stations, smart meters or microgrid controls. By processing real-time data locally, edge computing reduces latency to milliseconds, relieves central cloud infrastructures and increases the resilience of critical energy applications. Current technology trends include the integration of AI models at the network edge for predictive maintenance, the networking of edge devices via 5G networks, and hybrid edge-cloud architectures that combine data protection and scalability.

In the field of edge AI in particular, the decentralised execution of AI models is gaining importance, because it enables real-time decisions without cloud dependency. For applications such as load forecasting in smart grids, adaptive charging strategies for electric vehicles or fault detection in energy infrastructures, edge AI is indispensable in order to ensure autonomy and speed of response.

Energy-efficient inference chips form the technological backbone of this development. These specialised semiconductors optimise energy consumption when running pre-trained AI models and thus enable edge-AI systems to operate even in resource-constrained environments. In the All Electric Society, where every watt-hour counts, they reduce the ecological footprint of decentralised computing power and are what make edge AI economically scalable in the first place.

Sensors are the central sensory organs of an increasingly electrified world: they supply the data with which power grids, buildings, industrial plants and mobility are controlled, monitored and secured efficiently. Without precise sensor technology, efficiency potential remains unused, faults are hard to detect and complex systems can only be managed to a limited extent. Sensor development is therefore currently moving to the centre of numerous applications around energy, automation and digital infrastructure.

The most important technology trends include higher measurement accuracy, miniaturisation and sensor fusion, that is, the intelligent combination of several measured variables. Added to this are connected sensor systems that not only capture data but pre-process it directly and thus make it usable faster. New material systems are also gaining importance: the focus here is above all on flexible, thin and partly organic or inorganic semiconductor materials. They enable more sensitive, more robust or more flexible sensors and can be adapted better to demanding environments.

Also in demand are low-power designs, for example architectures with short activation, local signal processing and as little data traffic as possible, in order to reduce energy consumption and costs. Particularly relevant is also the integration of several functions in a small space: sensors increasingly handle not only measurement, but also pre-processing, diagnostics and interface tasks.

Cybersecurity is not a peripheral topic in the All Electric Society, but the prerequisite for connected energy, industrial and building systems to function reliably and for failures, manipulation or data leaks not to disrupt operation. Current trends are above all zero-trust architectures, AI-supported attack detection, secure remote access, continuous asset and vulnerability management, and the protection of IT and OT networks along the entire supply chain. For energy applications, segmented networks, robust identity and permission concepts, and resilient restart plans are also gaining weight.

The EU Cyber Resilience Act significantly increases the pressure on manufacturers: products with digital elements will in future have to be developed securely, documented and supplied with updates across their entire lifecycle. For operators and suppliers, this means more evidence, reporting and governance obligations as well as a stronger focus on security by design and security by default. Particularly for connected devices in energy grids, charging infrastructure and building automation, cybersecurity is thus becoming a competitive and approval issue.

In semiconductor technology, hardware root of trust, secure elements, TrustZone approaches, cryptographic accelerators and protected firmware update paths are therefore moving to the centre. Better protection mechanisms against manipulation and side-channel attacks, as well as more transparency across the entire supply chain, are also relevant. In the end, what is decisive is a chip that is not only powerful, but demonstrably trustworthy.

Non-volatile memories such as MRAM and eNVM are gaining importance in the All Electric Society, because they make controls, sensors and edge systems immediately operational again even after a power failure, while also reducing standby consumption. Current trends range from embedded NVM in microcontrollers through MRAM as a robust alternative to flash to eNVM for safety-critical industrial, energy and automotive applications.

The EU Cyber Resilience Act increases the pressure for hardware-based security functions, secure update capability and resilient data retention across the entire product lifecycle. This speaks in favour of MRAM and eNVM, because they preserve configuration data, logs and key material without an additional power supply and thus make security by design in connected devices easier.

From the point of view of semiconductor technology, integration into advanced process nodes, lower write voltages, higher endurance and better scalability of embedded solutions are currently particularly relevant. In parallel, the focus is shifting to energy-efficient architectures for edge AI and industrial IoT, where MRAM and other eNVM approaches are convincing less as a mass replacement for flash than as specialised devices for robust, low-power electronics.

Advances in power management ICs and integrated voltage regulators are what make electrification efficient, compact and system-capable. Current trends range from higher integration density through finely scalable multi-rail supply to better efficiencies, smaller designs and more closely coupled protection and diagnostic functions. Also particularly relevant are the closer integration with GaN and SiC power paths and the growing role of PMICs in AI edge devices, industrial electronics, energy storage and connected infrastructures. At the same time, the EU Cyber Resilience Act puts manufacturers under pressure to consider security by design, update capability and vulnerability management at chip and firmware level. This shifts PMICs and voltage regulators from pure supply components to security- and compliance-critical system building blocks. In semiconductor technology, system-in-package approaches, heterointegration, digital control and the coupling of energy efficiency with functional safety currently dominate.