China’s AI hardware buildout is expanding beyond processors and memory into the components that move information and present it to users. Raysolve Semiconductor has completed a Pre-A4 financing round of nearly 100 million yuan, according to a September 10 report by 36Kr. The startup says it will use the funds for technology iteration, product verification, industrial cooperation, and commercialization in near-eye display and optical interconnection. That makes the deal relevant to AI infrastructure, although Raysolve should not be described as an AI-accelerator company.
CMB International Capital led the reported round, with Fengtu Capital and an unnamed industrial strategic investor participating. Muyan Capital served as exclusive financial adviser, 36Kr said. The announcement does not disclose the startup’s valuation, production volume, revenue, customer contracts, or the amount of funding it raised in prior rounds. The available information is strongest on the intended technical directions, not on financial outcomes or market share.
Raysolve works on two areas that can look unrelated at first glance. One is Micro-LED micro-display technology for near-eye devices such as augmented- and mixed-reality hardware. The other is chip-level optical interconnection, where the company says it is working with optical-module manufacturers on high-speed, high-density connections. Both are relevant to AI systems because models require data to move through hardware, while increasingly capable visual devices may need efficient local or cloud-connected inference. Neither connection means that Raysolve has announced an AI processor or a deployed AI headset.
Micro-LED Is Raysolve’s First Hardware Foundation
PEdaily describes Raysolve as a developer of monolithic full-color Micro-LED micro-display chips. The company’s stated technical approach uses Quantum Dot Photoresist, or QDPR, technology for near-eye display applications. Micro-displays are small panels designed for compact devices, where brightness, color, pixel density, power consumption, and manufacturing yield can all determine whether a product is viable.
The cited reports do not provide a specification sheet for a commercial Raysolve display. There is no disclosed resolution, brightness figure, yield rate, unit price, or named AR/MR device customer. It would therefore be inaccurate to compare the company directly with a global display supplier or claim that its hardware has reached consumer volume. The more limited conclusion is that the startup is raising capital to advance a particular Micro-LED and quantum-dot manufacturing approach.
Near-eye displays are relevant to AI because many prospective AR and MR products depend on computer vision, voice input, real-time scene understanding, or cloud-connected assistance. Yet the display is only one component in that stack. A capable AI model does not automatically require Raysolve’s technology, and a Raysolve display does not automatically contain advanced AI. The financing announcement is better understood as evidence that capital is flowing toward specialized hardware components that could support future AI interfaces.
China’s recent edge-AI announcements demonstrate why that distinction matters. EastFrontier covered OPPO’s new benchmark for on-device AI memory, a story about memory characteristics inside devices. Raysolve’s Micro-LED work is about the output layer, while optical interconnection concerns data movement. A complete AI device will need both, but the companies occupy different positions in the hardware chain.
Optical Interconnection Links Displays to the AI Infrastructure Debate
The second reported direction is high-speed, high-density chip-level optical interconnection. 36Kr said the company intends to pursue industrialization in this field, and PEdaily said it is collaborating with optical-module manufacturers. The reports do not name those partners or disclose a production timetable. Raysolve’s stated role is development and collaboration, not a confirmed commercial supply agreement.
Optical links are drawing attention because AI systems must transfer larger volumes of data among processors, memory, and servers. As accelerators become more powerful, moving information can become a bottleneck. Companies across the semiconductor ecosystem are exploring ways to increase bandwidth and reduce the energy cost of interconnects. Raysolve’s financing places it in that broader design space, even if its public materials emphasize display technology as well.
The infrastructure context is visible in China’s wider chip-funding cycle. EastFrontier has separately tracked Nano-Core Chip’s recent AI-chip financing. Nano-Core and Raysolve are not direct peers. One announcement concerns AI-chip funding, while Raysolve’s stated work spans displays and optical interconnection. The comparison is useful because AI hardware is not one market. It is a chain of specialized components whose constraints can limit the final system.
This is also why optics can become strategically important without generating the headline attention of an AI accelerator. If an AI server cannot move data rapidly enough, more compute does not automatically create higher throughput. If an AR device cannot deliver a clear, efficient display, better software may not yield a usable product. The technical value is often determined by how components interact rather than by a single chip’s peak number.
The Financing Funds Development, Not Proven Market Scale
The Pre-A4 label shows that Raysolve remains at an early private-company stage. The company said the new funding will support product iteration, verification, and industrial cooperation. Those are precisely the activities needed before a deep-tech hardware startup can turn a laboratory process into a reliable supply business. They are also costly, because components must be manufactured, tested, qualified, and integrated with partners’ systems.
The investor mix gives limited clues. CMB International Capital led the round, Fengtu Capital participated, and an unnamed industrial strategic investor joined. Without the strategic investor’s identity or a disclosed partnership agreement, there is no basis for claiming a specific customer channel or application. The announcement similarly provides no statement about export markets, revenue targets, or an initial public offering.
For readers following China AI infrastructure, the key point is not that Raysolve has solved interconnection or display manufacturing. It has not publicly shown that. The point is that specialized optical and display technologies are receiving financing at the same time as China’s larger AI-compute and edge-device markets are expanding. The country’s AI ambitions depend on processors and HBM, but they also depend on the less visible layers that connect machines and let people interact with them.
Raysolve’s next milestones will need to be more concrete: a verified product specification, a qualified customer program, a disclosed manufacturing result, or a named optical-interconnect partner. Until then, the nearly 100 million yuan round should be viewed as capital for engineering and validation. It reflects investor interest in a hardware segment that may become more important as AI requires faster links, denser devices, and new forms of visual computing, but it is not yet proof of a commercial AI-infrastructure breakthrough.