"Optical Standard" Completes a New Round of Strategic Financing to Accelerate the Commercialization of Optical Computing Chips | Exclusive on 36Kr

Written by | Wang Fangyu

Edited by | Su Jianxun

36Kr recently learned that Photon AI Computing Chip Company LightStandard Technology has announced the completion of a new round of strategic round financing. The investor is a domestic first-tier Internet giant, and the two sides will cooperate on AI computing power hardware.

The old shareholders Zhongying Venture Capital and Xiaomiao Langcheng also continue to increase their holdings in this round, and Mushi Capital serves as the company's exclusive financing financial advisor. The funds raised in this round will be used for continuous investment in the photon AI computing hardware system, server ecology, and software ecology to accelerate the commercialization of optical computing in artificial intelligence scenarios.

Unlike traditional chips, optical chips achieve information processing through optical technology rather than electrical signals. Relying on light as the transmission carrier, optical chips have higher transmission speed, lower power consumption, and greater bandwidth.

LightStandard Technology has chosen a special route in the technical route of optical chips - LightStandard adopts the heterogeneous integration of silicon photonics + phase change materials and the unique Crossbar photonic matrix computing structure, becoming the first commercial company to achieve the integration of storage and computing in optical computing chips.

Publicly available information shows that this technical route has some natural advantages: it is easier to make a large matrix scale due to the small unit size and high integration, and the low power consumption, low latency, and high stability brought by in-memory computing.

The different technical routes have enabled LightStandard Technology to achieve the tape-out of optical computing chips early, and the computing power density and computing power accuracy of the chips have reached commercial standards.

Xiong Yinjiang, the co-founder of LightStandard, told 36Kr that the company has completed five tape-outs so far, and 2025 will be the first year of commercialization of the company's products.

"Introducing a strategic investor at this time is a decision focused on the development in the next 5 - 10 years, because the technological advancement determines the starting point, while the product practicality determines how far it can go."

The actual photo of LightStandard's optical computing wafer               Source: Authorized by the enterprise

The commercialization of optical computing products has attracted much attention from the government, capital market, and enterprises in the computing power ecological chain in recent years. From the perspective of market application, the integrated storage and computing optical computing chip is naturally suitable for cloud-side AI reasoning and training scenarios, with the characteristics of large computing power and low power consumption.

Xiong Yinjiang told 36Kr that in the field of general AI reasoning computing power, the company has reached cooperation intentions with AI computing power leasing platforms and intelligent computing centers led by local governments, and is conducting in-depth technical exchanges with Internet industry parties including the strategic investor in this round. In the vertical application field, the company has carried out joint technical research with vertical large model manufacturers, autonomous driving manufacturers, satellite manufacturers, scientific research institutes, etc., to explore the potential of photon storage and computing in various application scenarios.

In terms of the core team, LightStandard has a strong scientific research background with post - 95 founders and post - 70 and post - 80 engineering executives with 10 - 20 years of experience.

Among them, Cheng Tangsheng, the co-founder of the company, is a student of Harish Bhaskaran, the world's first "Phase Change Material Optical Computing" expert, a professor in the Department of Materials at the University of Oxford, and a fellow of the Royal Academy of Engineering in the UK. He is responsible for R & D and engineering implementation. The engineering executives have been deeply engaged in the industry for many years and have rich experience in R & D and industrialization.