Silicon Photonics Startup Lingdong Xinguang Secures Funding

A Chinese silicon photonics startup, Lingdong Xinguang, has successfully closed a multi-million dollar pre-Series A funding round. The investment, led by Panlin Capital with participation from Tongfang Investment and a joint fund of Deep Angel and Huize Tiancheng, will accelerate the development of advanced optical interconnect technologies crucial for overcoming the computational bottlenecks in artificial intelligence and other high-performance computing sectors.

The capital infusion is earmarked for critical research and product commercialization in chip-to-chip optical communication. Specifically, Lingdong Xinguang will focus on advancing its proprietary SmartComb multi-wavelength dense light source and the SmartPHY optical I/O micro-chiplet. These innovations are designed to significantly enhance data transfer speeds between processors, a growing challenge as AI models demand exponentially more processing power.

Founded in 2022, Lingdong Xinguang leverages over a decade of research from its founder, Professor Chen, a distinguished figure in silicon photonics from Tsinghua University. The company's core competency lies in its silicon photonics integration platform, which utilizes standard integrated circuit manufacturing processes to create miniaturized and scalable optical components. This approach promises to reduce costs and power consumption compared to traditional solutions, positioning silicon photonics as a key enabler for next-generation computing infrastructure.

The market for optical interconnects is experiencing rapid growth, driven by the insatiable demand for bandwidth in AI training, autonomous driving, and advanced telecommunications. Traditional copper interconnects are reaching their physical limitations, making optical solutions, particularly silicon photonics, a vital area of innovation. While established players in Europe and the US have historically led in advanced processes, Chinese companies like Lingdong Xinguang are rapidly closing the gap, supported by significant domestic investment and a burgeoning AI ecosystem.

Lingdong Xinguang's product pipeline includes ultra-narrow linewidth lasers, optical I/O solutions, and FMCW lidar light sources. The company has already achieved significant traction with its domestically produced ultra-narrow linewidth lasers, which offer superior performance at lower costs and power consumption. The strategic emphasis on optical I/O is particularly noteworthy, as it directly addresses the critical need for high-bandwidth communication within large-scale AI compute clusters, enabling thousands of chips to function cohesively.

During the funding discussions, investors highlighted the transformative potential of silicon photonics. Panlin Capital's Managing Partner, Li Yuhui, commented, "The increasing demand for computational and communication speeds in sectors like AI and autonomous driving is pushing traditional integrated circuits to their physical limits. Silicon photonics is emerging as a critical technology. Lingdong Xinguang's deep technical foundation, validated by the successful commercialization of its ultra-narrow linewidth lasers, positions it strongly. We are particularly optimistic about their advancements in chip-to-chip optical interconnects, where their multi-wavelength dense light source offers a distinct competitive advantage over existing solutions."

The company's leadership, including co-founder Wang Hui, a Tsinghua alumnus with extensive industry experience, emphasizes the platform nature of their technology. "Our silicon photonics platform allows for the integration of optical and electrical functions on a single silicon substrate, enabling mass production through CMOS processes," Wang explained. "This high integration, cost-effectiveness, and low power consumption are key advantages. We believe our advancements in dense wavelength division multiplexing (DWDM) technology for optical I/O will be a significant growth driver, enabling higher bandwidth density and lower system costs compared to current coarse wavelength division multiplexing (CWDM) approaches."