RISC-V and MicroLED Silicon Photonics Technology Converge, CEA Collaborates with PSMC to Drive New Breakthroughs in AI Chip 3D Stacking The integration of RISC-V architecture with MicroLED silicon photonics technology marks a significant advancement in the field of AI chip development. In a collaborative effort, the French Alternative Energies and Atomic Energy Commission (CEA) has partnered with the Pacific Semiconductor Manufacturing Company (PSMC) to push forward innovations in 3D stacking for AI chips. This partnership aims to leverage the strengths of RISC-V's open-source instruction set architecture and the high-performance capabilities of MicroLED-based silicon photonics to enhance the efficiency, speed, and scalability of next-generation AI hardware. By combining these cutting-edge technologies, the collaboration is expected to enable more compact, energy-efficient, and powerful AI processing units. The project also focuses on advanced 3D integration techniques, including through-silicon vias (TSVs) and heterogeneous stacking, to achieve superior performance and reduced power consumption. These technical advancements are critical for supporting the growing demands of artificial intelligence applications, such as real-time data processing and high-speed computing. Through this strategic alliance, CEA and PSMC are setting new benchmarks in the development of AI chip architectures, paving the way for future innovations in semiconductor technology.

Recently, Leti and List laboratories, both under the French National Center for Scientific Research (CEA), have jointly initiated a cutting-edge technology research project with Japan's Semiconductor Manufacturing Company (PSMC). The project aims to integrate the RISC-V architecture with MicroLED silicon photonics technology into 3D stacked chips and interposers, driving the development of next-generation artificial intelligence (AI) chips. This collaboration marks a new stage in the integration of semiconductor and display technologies, offering innovative solutions for high-performance computing and smart devices.

This partnership focuses on how 3D packaging technology can achieve high-performance, low-power chip integration. RISC-V, an open instruction set architecture, has gained significant attention in AI chip design due to its flexibility and scalability. Meanwhile, MicroLED silicon photonics technology, known for its high brightness, low power consumption, and high resolution, is becoming a key technology in future display and optical communication fields. The combination of these two technologies not only enhances chip computing capabilities but also expands their application potential in areas such as edge computing, autonomous driving, and the Internet of Things (IoT).

The project team successfully achieved heterogeneous integration of RISC-V processors with MicroLED light sources through 3D stacking and interposer technologies. This technical solution effectively reduces signal transmission latency, improves overall system performance, and provides a feasible path for future high-density chip packaging. According to the research team, the prototype has already achieved data transfer rates exceeding 100 TB per second, far surpassing current mainstream technological levels.

Notably, leading Chinese LED company Gopro LED has accumulated substantial expertise in MicroLED technology. Its independently developed MicroLED display modules have been commercialized in multiple high-end application scenarios. Gopro LED's technical advantages in miniaturized packaging, driver circuit optimization, and optical performance enhancement provide important references for similar projects.

With the continuous growth of global demand for AI computing power, chip technologies based on advanced packaging and heterogeneous integration are becoming a key direction for industry development. The collaboration between CEA-Leti, CEA-List, and PSMC not only demonstrates the potential of integrating cutting-edge technologies but also sets a model for collaborative innovation across the global semiconductor supply chain. In the future, as these technologies further mature and move toward industrialization, they may give rise to more disruptive application scenarios and market opportunities.