KuaiLin Optoelectronics, founded by ShanghaiTech Chen Baile lab alumni, closed a tens of millions yuan angel round to ship by 2027 a uni travelling carrier (UTC PD) photodiode — the light to electrical sensor at the receive end of next generation 1.
Every fiber inside an AI data center ends at the same unglamorous component, a photodetector chip that converts incoming light back into an electrical signal. As the industry moves from 800G toward 1.6T per lane, that chip has to handle more bandwidth in the same power budget. A four-PhD team from ShanghaiTech, founded as Shanghai KuaiLin Optoelectronics in 2024, is betting a modified uni-travelling-carrier (MUTC) structure can break the bandwidth-efficiency tradeoff that has kept domestic photodetectors out of the highest-speed optical modules.
The startup has closed an angel round in the "tens of millions of yuan," led by Xiaomiao Langcheng (小苗朗程) with Yicun Capital (一村资本) following, and Zhonglan Tech Innovation (中岚科创) engaged as long-term financial advisor. The capital is earmarked for chip R&D iteration, sample validation, mass-production ramp, and senior engineering hires. Founder and CEO Li Linze (李林泽) leads a four-person co-founder team drawn from the same ShanghaiTech research group of Prof. Chen Baile (陈佰乐). The co-founders are Wang Jingyi (王景熠), Wang Luyu (王鲁玉), and Long Tianyu (龙天宇), all ShanghaiTech PhDs and all under 30, from a group with prior publications in Nature and Nature Photonics.
A Nature Photonics paper (s41566-025-01784-0) reports modified uni-travelling-carrier photodiodes with 206 GHz bandwidth and 0.81 A/W external responsivity, and an arXiv preprint (2501.02812) reports MUTC photodiodes with a bandwidth-efficiency product above 130 GHz × 100%. Both come from the Chen group lineage at ShanghaiTech. The startup positions the modified UTC-PD it commercializes as the same family of devices, targeting bandwidth above 200 GHz at responsivity around 0.8 A/W. Those numbers describe research demonstrations, not shipping product specs.
The market KuaiLin is positioning into is the back half of an optical-module upgrade cycle. Today the sector is moving from 800G to 1.6T per-lane modules, with 3.2T on the industry roadmap. The high-end photodetector side has historically been dominated by overseas suppliers (Coherent/II-VI, Lumentum, Broadcom, Mitsubishi, and Semtech are commonly cited), with thin domestic-substitution presence at the top tier. The team frames its bet as closing that specific gap. Per the company's own roadmap, core product mass delivery is targeted for 2027; today it is in customer sampling and production-process development. A separate track in 6G terahertz chips is described by the company as already generating "stage-level revenue," a claim that has no independent corroboration in the available source set and should be read as company-stated.
The Chen group's published work shows the device physics is already credible. What separates a usable photodetector from a research result, per the founders, is the transfer of process know-how onto a foundry's mass-production line: epitaxy, fab, and test yield. Customers in the next module cycle are asking for reliability and capacity, not peak single-chip specs. The team explicitly frames this part of the build as a "race against time" against the 1.6T-to-3.2T module roadmap. A ShanghaiTech English-language news feature from February 2026 documents related research achievements from the same group, providing independent university-side corroboration of group-level capability, though not of the company itself.
The startup also pitches itself against a domestic landscape the founders describe as dominated by piecemeal or follow-on innovation, projects that lack the full stack from epitaxial growth through wafer-level packaging. KuaiLin's argument is that its team can run all three layers because the founders came from the same lab, working on devices the lab had already peer-reviewed. If the foundry transfer holds, 2027 would put volume production roughly in line with the next generation of optical-module deployment. If it does not, the lab demo will sit on a long shelf of "world record" photodetector results that never cleared mass-production qualification.