First University-designed 7-nanometer AI Chip

The Technical University of Munich (TUM) has unveiled the EU's first AI chip using modern 7-nanometer technology. The neuromorphic chip was designed by Prof. Hussam Amrouch in accordance with the industry standards set by world-leading chip manufacturer TSMC. The professor of AI processor design and his research group now plan to produce at least three new designs per year, to be manufactured from 2028 onward by the Dresden-based European Semiconductor Manufacturing Company (ESMC).

The COVID-19 pandemic, the war in Ukraine and broader geopolitical shifts have exposed weaknesses in global supply chains. 'That is why it will be crucial in the future to cover all stages in Germany and Europe - from training AI chip specialists to technological development and chip manufacturing,' says Prof. Amrouch. Earlier this year he established the MACHT-AI chip research and training center at TUM, funded by the Bavarian Ministries of Science and Economic Affairs. Looking ahead, he envisions students in Germany learning to design AI chips and develop algorithms directly with companies, with production also taking place domestically.

AI chip enables local data processing

In contrast to cloud-based chips such as those manufactured by NVIDIA, Prof. Amrouch's approach focuses on technology that processes data locally on the device itself rather than sending it to the cloud. 'This is a fundamental solution for protecting the privacy of our citizens,' says Amrouch.

The AI chips are based on the open-source RISC-V architecture and can be customized for specific applications - from recording and analyzing heartbeats or brain signals in healthcare to implementing language models. This specialization is more efficient than using a general-purpose AI chip. Amrouch: 'You can buy a Ferrari, but that doesn't necessarily make you faster in the city. An e-bike is more efficient here.' Another critical point is that the technology is also intended for use in control electronics in quantum computing in the medium term.

Data remains private and secure

Local data processing enhances cybersecurity and helps protect technologies against misuse by third parties. 'Those who design and build the chip know exactly what is inside it and can guarantee how it will function,' says Amrouch, who believes it is essential for companies to be able to assure their customers that their data will remain private.

Trust is a critical 'currency' for industries such as the automotive sector, where Europe currently lacks sovereignty in high-performance chips. The defense industry also places the most stringent demands on security, for example when it comes to the use of chips for drones. Potential built-in vulnerabilities such as Trojans pose incalculable risks here, especially if chips are developed and manufactured outside Germany and the EU.

Markus Blume, Bavaria's Minister of Science, said: "This is a big breakthrough for AI research at TUM. The European Union's first AI chip using 7-nanometre technology comes from Bavaria and combines performance, energy efficiency and security. With MACHT-AI, TUM plays a central role in Bavaria's semiconductor ecosystem. To remain competitive and independent while retaining technological sovereignty, we are developing key technologies and training talent here."

Bavarian Minister of Economic Affairs Hubert Aiwanger is convinced: 'Artificial intelligence has undergone tremendous development in recent years. As an economy, we must be in a position not only to keep pace with this development, but also to help shape it. The development of TUM's own AI chip is an impressive demonstration of how well the Bavarian semiconductor ecosystem is developing.'

TUM President Thomas Hofmann emphasizes: "Geopolitical risks have increased dramatically in recent years. That is why cutting-edge technology must be developed and produced in Germany and Europe. With his AI chips, Prof. Hussam Amrouch impressively demonstrates how dependence on Asia and the USA can be successfully reduced, and Europe's technological self-determination strengthened. And, incidentally, with a very sustainable concept."