STARLight Project chosen as the European consortium to take the lead in next-generation Silicon Photonics on 300mm wafers
The STARLight project brings together a consortium of leading industrial and academic partners to position Europe as a technology leader in 300mm silicon photonics (SiPho) technology by establishing a high-volume manufacturing line, developing leading-edge optical modules, and fostering a complete value chain. From now until 2028, STARLight aims to develop application-driven solutions focusing on key industry sectors such as datacentres, AI clusters, telecommunications, and automotive markets.
Led by ST the STARLight consortium has been selected by the European Commission under the EU CHIPS Joint Undertaking initiative.
“Silicon Photonics technology is critical to put Europe at the crossroads to the AI factory of the future and the STARLight project represents a significant step for the entire value chain in Europe, driving innovation and collaboration among leading technology companies. By focusing on application-based results, the project aims to deliver cutting-edge solutions for datacentres, AI clusters, telecommunications, and automotive markets. With well-recognised pan-European partners, the STARLight consortium is set to lead the next generation of silicon photonics technologies and applications,” said Remi El-Ouazzane, President, Microcontrollers, Digital ICs and RF products Group at STMicroelectronics.
Silicon photonics is a preferred technology to support datacentres and AI clusters optical interconnects for scale-out and scale-up growth, as well as for other technologies such as LIDAR, space applications and AI photonic processors that require better energy-efficiency and power efficient data transfer. It combines the high-yield manufacturing capabilities of CMOS silicon, commonly used in electronic circuits, with the benefits of photonics, which transmits data using light.
The STARLight project has an initial focus to build datacom demonstrators for datacentres, based on PIC100 technology, capable of handling up to 200Gb/s with key actors including ST, SICOYA and THALES. It will also develop prototypes for free-space optical transmission systems, designed for both space and terrestrial communication.
Additionally, the project will leverage the multidisciplinary experience of major contributors to shape the research effort towards a 400Gbps per lane optical demonstrator using new materials, targeting the next generation of pluggable optics.
The STARLight project aims to develop a cutting-edge photonic processor optimised for tensor operations, such as matrix vector multiplication and multiply-accumulate, with superior characteristics in terms of size, data processing speed, and energy consumption compared to existing technologies. Since neural networks – the core algorithms behind AI – rely heavily on tensor operations, enhancing their efficiency is critical for AI processing performance.
The STARLight project plans to develop and showcase innovative silicon photonic devices specifically designed for the telecommunications industry. Ericsson will focus on two concepts to improve mobile network efficiency. The first involves the development of an integrated switch to enable optical offload within Radio Access Networks, allowing for more efficient handling of data traffic. The second concept explores Radio over Fibre technology to relocate power-intensive processing ASICs away from antenna units, thus providing enhanced capacity and savings in embodied CO2. Additionally, MBRYONICS will develop a free space to fibre interface at the reception of Free Space Optical (FSO) communication, which is a key element in the design of an optical communication system.
The STARLight project will also demonstrate how it performs in sensing applications, and the close relationships of STEERLIGHT, a LiDAR sensors maker, with leading car manufacturers will help make this an industrial reality.
Within the project, THALES will develop sensors that accurately generate, distribute, detect, and process signals with intricate waveforms to demonstrate key functionalities. More broadly, the outcomes of this project are also intended to benefit the wider ecosystem of indoor and outdoor autonomous robot manufacturers.
