From 8,000 meters underground to high-temperature robot joints, Tsinghua-affiliated special power chip company completes hundreds of millions of yuan in Series C financing | 36Kr Exclusive

Written by Zhang Ziyi

Edited by Yuan Silai

36Kr learned that Hangzhou Puxi Optoelectronics Semiconductor Technology Co., Ltd. (hereinafter referred to as "Puxi Optoelectronics"), an enterprise focusing on special power chips and systems for extreme environments, has recently completed a Series C financing of several hundred million RMB. New institutional investors include 10 state-owned and industrial capital entities such as Zhejiang Jiahe Changyue, Chengdu Venture Capital Co., Ltd. (Chengdu Sci-Tech Innovation Investment), Shanghai Semiconductor Equipment and Materials Industry Investment Fund, Shandong Finance Group-affiliated Funds, and Guizhou Guochuang.

This round of financing will be mainly used for the iteration of fourth-generation semiconductor gallium oxide chips, capacity expansion of 230℃-grade high-temperature chip modules, R&D of high-temperature robot joints, and pre-research of space computing power and controllable nuclear fusion power supplies.

36Kr understands that Puxi Optoelectronics was founded in 2020. In 2025, the company's revenue reached 230 million RMB, making it one of the few profitable enterprises in the industry track.

Around 2020, power semiconductors began to rise, and a large number of enterprises entered the power semiconductor markets covering new energy vehicles, photovoltaics and energy storage. Puxi Optoelectronics did not participate in the price competition in the general market.

Puxi Optoelectronics has selected a more specific market segment — special power chips and systems applied in extreme environments.

Different from CPUs and GPUs that are responsible for computing, power chips withstand relatively high voltage and current. The higher the power, the higher the requirements for efficiency, heat dissipation and reliability.

In addition to withstanding high power, special power chips also need to cope with extreme conditions such as high temperature, radiation and narrow space.

Xu Yili, Founder and CEO of Puxi Optoelectronics, said that the company's current products can withstand a maximum ambient temperature of over 250℃, and feature smaller size and lighter weight under the same power, lower self-heating, and high reliability in harsh environments.

The first market Puxi Optoelectronics entered was deep underground energy exploration.

When oil exploration equipment goes 8000 meters or even deeper underground, the interior of the Earth gets increasingly hot, with the ambient temperature exceeding 200℃. Meanwhile, the internal space of downhole drilling tools is limited, so the power chip system must not only resist high temperatures, but also minimize its volume and self-heating.

"The ambient temperature is already 200 to 300 degrees Celsius, if the equipment itself generates additional heat, the internal temperature may reach 300 to 400 degrees Celsius," said Xu Yili.

36Kr understands that to make a power system operate above 200℃, it is not enough to only switch to high-temperature-resistant chips. Resistors, capacitors, transformers, substrates and packaging materials may fail first, and once any part fails, the entire system will crash.

Therefore, Puxi Optoelectronics has developed the HS-MCM platform technology, which adopts fully enclosed hybrid packaging for special power chips and peripheral circuits such as driving and feedback. Many of the components involved do not even have mature supply chains at present.

Since 2023, Puxi Optoelectronics has set its sights on the commercial aerospace sector beyond 10000 meters in space, and started to enter markets such as special robot joints.

Take robots as an example: to run faster and jump higher, the joints need greater instantaneous power, but their volume and weight cannot increase accordingly. After the power is improved, heat generation becomes another prominent problem.

In industrial scenarios such as refining and chemical, steel, and electrolytic aluminum, the ambient temperature itself may have reached 60 to 70 degrees Celsius. After the joints work, the internal heat generation may further push the temperature above 100℃. The reliability of ordinary electronic systems will be significantly affected if they work in such an environment for a long time.

Therefore, Puxi Optoelectronics started with high-temperature industrial robot joints, leveraging its accumulation in high temperature resistance, low heat generation and miniaturization to develop joint system modules; as robots put forward higher requirements for high power density, the company has also begun to expand to more conventional robot scenarios.

In the next step, Puxi Optoelectronics plans to further extend this set of special technologies. In the second half of this year, the company will start to expand the power conversion system business for AI computing centers, and is also developing emerging application directions such as controllable nuclear fusion power supplies and high-power drones.

After the Series C financing, Puxi Optoelectronics will continue to expand its business, carry out research on fourth-generation semiconductor chips, and realize mass shipment of "SiC + gallium oxide" hybrid power module products, becoming the first company in the industry that covers both third-generation and fourth-generation semiconductors. At the same time, it will carry out cutting-edge research including computing-power coordination for AI computing centers, space computing power energy OS and solid-state fusion power supplies.

In terms of team, Puxi Optoelectronics was co-founded by four undergraduate classmates of the 1998 grade in the Department of Electronic Engineering of Tsinghua University. Xu Yili, Founder and CEO of the company, holds bachelor's and master's degrees from the Department of Electronic Engineering of Tsinghua University, and is a leading scientific and technological entrepreneurship talent of the Central Organization Department's "Ten Thousand Talents Plan", responsible for the company's overall strategy and technical direction. Yang Qi, CTO, worked at Schlumberger for 15 years, long engaged in the design of extreme high-temperature downhole systems, with rich experience in R&D of electronic systems for extreme environments. Li Xin, CMO, once worked in China's electronic industry system, long responsible for sales and marketing of high-end electronic equipment, and has accumulated customer resources in the energy, aerospace and other fields.

CEO Dialogue

36Kr: The company focused on small-batch, high-unit-price special products in the early stage, and now you are expanding the production lines. What is the biggest change at this stage?

Xu Yili: The business will be very different when it reaches a certain scale. When customer orders grow larger, they will come to audit the factory, inspect your production lines, equipment, production systems and quality control. At this time, the problem is no longer "whether the product can be made", but the yield rate and consistency of the products.

R&D can allow engineers to find solutions flexibly, but production cannot be too flexible. The workflow of each process, the placement of items, and the management of materials must all be specified in extremely detailed rules.

After large-scale production, the biggest challenge is actually production consistency, and we will allocate a large part of the funds to the standardization and expansion of production lines.

36Kr: From oil exploration and robotics to the current development of computing center business, these industries seem to be very different. How do you judge whether a technology can be replicated to new markets?

Xu Yili: First of all, the underlying technology is reusable, because what we develop are all power chips and power systems. Different industries have different requirements for power, voltage and current, but the core technical principles will not change much.

When we evaluate a new industry, we first check whether it has the problems that we are good at solving. For example, energy exploration faces problems of high temperature and small space; robots require high power without making the joints too large or too heavy, and need to control heat generation; the same is true for computing centers: as the power gets higher and higher, power supply and heat dissipation will become increasingly important.

We will not enter a market just because it is large in size. Instead, we first confirm whether our "specialty advantages" can be brought into play, and then judge whether the market is large enough.

I think the demand of "higher power without larger volume and higher self-heating" will remain unchanged for a long time. As long as this contradiction exists, the technologies we have accumulated will have opportunities to extend to new scenarios.