Wireless communication has evolved dramatically over the past decades. Early longwave radio signals (150–300 kHz) traveled great distances but slowly and carried limited amounts of data. The introduction of FM radio (87.5–108 MHz) and analog television marked a major leap forward, using shorter wavelengths able to deliver clearer sound and higher-quality images.
Today, wireless data transmission is increasing exponentially. In 2024 alone, an estimated 2000 million photos are shared every month — a surge made possible by even higher-frequency waves. Technologies like 4G (700 MHz–2.6 GHz) and 5G (up to 28 GHz) offer remarkably high speeds but rely on a dense infrastructure of antennas to compensate for their limited range.
To keep up with the future explosion of digital activity (photos, videos, online interactions and much more) a new network will be needed. This is where 6G comes in, promising speeds 10 to 100 times greater than its predecessor, along with enhanced data capacity and significantly reduced latency.
With 6G, wireless communication enters the realm of ultra-short waves, exceeding 100 GHz. These frequencies will enable near-instantaneous data transmission but will also be more susceptible to physical obstacles, requiring an even denser infrastructure of small antennas. This “last mile” of communication — the final stretch between a device and the network — is precisely the focus of the European HERMES project, initiated by the IMS laboratory.
A major challenge
The purpose of the HERMES project is to develop increasingly compact radio-frequency circuits while enhancing their performance by utilizing more spectrum at higher frequencies for 6G. These next-generation chips must also be more reliable, cost-effective, and energy-efficient to address today’s critical environmental challenge.
Currently, digital technologies account for 10% of global greenhouse gas emissions. Since controlling user behavior is not possible, efforts must focus on reducing energy consumption at the source: within the electronic circuits themselves.
As early as 2020, François Rivet, the project’s coordinator, had the intuition that artificial intelligence could play a key role in achieving these ambitious goals.
Pioneering the future of 6G
On September 1st, 2021, the HERMES project officially launched with a bold objective: integrating radio-frequency circuits (RFIC) into a single CMOS chip.
While this technology has been widely used since the 1990s, it needed adaptation to overcome its limitations at extremely high frequencies, thus making it viable for 6G applications.
By leveraging Walsh’s mathematics, research led to the development of a chip that shifts from the time domain to the frequency domain.
In practice, digital data — stored as binary code on devices — is converted into an analog signal compatible with radio waves, allowing it to travel through the air. A receiving chip then captures the signal and reconverts it back into digital form for use.
This process integrates radio-frequency modulators and demodulators into electronic chips, forming the modem that enables wireless communication under the 6G standard. To enhance performance, the RF Front-End (the set of circuits responsible for handling radio signals) was optimized using artificial intelligence. AI helps correct distortions naturally introduced by electronic components, significantly reducing the energy consumption of signal processing compared to conventional linearization techniques.
But the role of AI in HERMES goes even further. The project’s core innovation lies in its flexibility: AI dynamically adjusts frequency availability to maximize spectral efficiency. For instance, if a 6G antenna detects congestion in an urban area due to high video traffic, AI can automatically redistribute data streams to less crowded frequency bands.
In short, HERMES combines AI with Walsh’s mathematics for 6G (up to 150 GHz), using CMOS technology that seamlessly integrates digital and analog processing.
A breakthrough technological innovation
The circuits were developed by a team from the IMS Laboratory, including François Rivet, Hervé Lapuyade, Yann Deval, Nathalie Deltimple, Éric Kerhervé, Antoine Lhomel, Maxandre Fellmann and Pierre Ferrer.
Once designed, the chips were sent to STMicroelectronics, the only purely European manufacturer of electronic circuits. This collaboration was made easier by a long-standing joint laboratory between IMS and STMicroelectronics, fostering productive research and development efforts.
With the HERMES project, Europe has maintained full sovereignty over both the design and production of this cutting-edge technology.
The first chip tests, conducted in late 2024, confirmed that the innovation not only met but exceeded expectations. Everything worked flawlessly. Communication speeds increased tenfold, far surpassing the initial goal of a threefold improvement.
Additionally, the AI-driven system demonstrated remarkable energy efficiency, consuming ten times less power than conventional technologies.
The HERMES team has successfully developed a breakthrough innovation that is cost-effective, suitable for the mass market, and delivers a tangible environmental benefit.
A true success story!
—
The HERMES project is led by the University of Bordeaux and funded by the European Innovation Council (EIC) under the FET Open program. It brings together a consortium of leading research institutions and industry partners, including the IMS laboratory (France), CEA (France), Baltijos Pažangių Technologijų Institutas (Lithuania), Silicon Austria Labs (Austria), Katholieke Universiteit Leuven (Belgium), and Intracom Defense (Greece).
François Rivet, holding the HERMES project’s wafer packed with 6G-ready chips. Photo credit: © Gautier DUFAU.
