The Technical University of Munich (TUM) has witnessed a groundbreaking development in artificial intelligence (AI) chip technology, thanks to the achievements of Professor Hussam Amrouh. His innovative architecture, utilizing ferroelectric field-effect transistors (FeFET), demonstrates twice the power of other existing approaches to in-memory processing. This advancement has the potential to significantly enhance generative AI, deep learning algorithms, and robotics applications.

A key aspect of Amrouh's innovation is the integration of data storage within the transistors themselves. Unlike conventional circuits that rely on separate locations for processing and storing data, the new solution allows for time and energy savings, thereby increasing chip efficiency.

The transistors employed in this architecture have a size of only 28 nanometers, with millions of them included in each chip. The increased performance is particularly crucial for power-demanding applications, such as real-time drone flights.

The chip's efficiency is measured by the parameter TOPS/W (tera operations per second per watt), which determines the number of trillion operations a processor can perform with one watt of power. Amrouh's AI chip, developed in collaboration with Bosch and Fraunhofer IMPS with support from GlobalFoundries, achieves an impressive 885 TOPS/W, making it twice as powerful as similar AI chips available on the market.

Amrouh's chip architecture is inspired by the human brain, where neurons process signals, and synapses store and memorize information. Through the use of ferroelectric transistors (FeFET) that retain information even when disconnected from the power source, the chip can simultaneously process and store data.

This breakthrough allows for the creation of efficient chips applicable in areas such as deep learning, generative AI, and robotics. Amrukh states, "We can now create high-performance chipsets that can be used in applications requiring on-the-spot data processing."

Amrouh's AI chip represents a significant leap forward in the field of AI computing. Its increased power and efficiency open up new possibilities for energy-efficient applications. As technology advances, AI chips like those created by Amrouh will play a pivotal role in shaping the future of artificial intelligence.