Shanghai's 10-billion-yuan unicorn is about to go public.

Another automotive-grade chip unicorn is knocking on the door of the A-share market.

Recently, Calterah Semiconductor Technology (Shanghai) Co., Ltd. (hereinafter referred to as "Calterah") officially submitted its prospectus to the Science and Technology Innovation Board of the Shanghai Stock Exchange. This automotive-grade millimeter-wave radar chip company, born out of the Berkeley Lab, is about to land on the capital market, with a planned fundraising of 3.489 billion yuan.

As a benchmark enterprise for domestic substitution, Calterah broke the decades-long monopoly of international giants with the CMOS technology route. In 2025, its domestic market share of in-vehicle millimeter-wave radar chips reached 31.1%, ranking second in China and fourth in the world. Its cumulative shipment exceeded 30 million units, with BYD, Geely, NIO and Volvo all on its customer list.

Founded 12 years ago, the company has completed 11 rounds of financing intensively, with China Integrated Circuit Industry Investment Fund Phase II, Walden International, Noah Holdings, GAC Capital, Pudong Venture Capital, Silergy, Zhangjiang Sci-Tech Investment, and Greater Bay Area Science and Innovation Fund all participating, pushing its valuation to over 10 billion yuan.

Berkeley Mentor and Apprentice Made the Right Bet

The story of Calterah began in 2014.

After receiving his doctorate in electrical engineering from the University of California, Berkeley, 30-year-old CHEN Jiashu returned to China and was determined to start a business in Shanghai, focusing on millimeter-wave radar. During his time in the United States, CHEN Jiashu led the team to develop the world's first 60GHz WiGig CMOS SoC chip, and won the Fulbright Science and Technology Award issued by the U.S. Department of State. He was the only Chinese student among the 27 winners that year.

At that time, China's automotive industry was on the eve of an intelligentization boom, but the core component of millimeter-wave radar chips was completely controlled by international giants such as Texas Instruments, NXP, and Infineon, with a domestic localization rate close to zero.

He also invited his mentor, Professor Ali Niknejad, Director of the Berkeley Wireless Research Center (BWRC), to start the business together. The company was named "Calterah", where "CAL" pays tribute to Berkeley, and "TERAH" is taken from Terahertz.

This pair of mentor and apprentice chose a different path from the mainstream. The common practice in the industry is to use gallium arsenide or silicon germanium processes to make a multi-chip assembly solution, which has stable performance but high cost and low integration. From the very beginning, Calterah bet on the CMOS process for its low cost and high integration, but its performance in high frequency bands has long been questioned as not meeting automotive-grade standards.

Many people thought that the Berkeley PhDs were being overly idealistic. But facts have proved that they made the right bet.

In 2015, only one year after its establishment, Calterah taped out the world's first fully integrated 77GHz radar transceiver single chip. In 2017, Yosemite, the world's first automotive-grade CMOS 77GHz millimeter-wave radar RF front-end chip, went into mass production, marking the first breakthrough for domestic millimeter-wave radar chips. In 2019, the antenna was embedded into the package, and the 77/60GHz millimeter-wave radar AiP SoC chip was launched, further lowering the design threshold of radar modules. In 2024, the Kunlun platform, which is mainly based on the domestic supply chain, was launched.

The real power of the CMOS route lies in cost. Traditional gallium arsenide solution-based radar modules often cost hundreds of US dollars, which are only affordable for million-dollar luxury cars. The CMOS single-chip solution cuts the cost to a fraction, making millimeter-wave radar no longer exclusive to luxury cars, but a standard feature for 100,000-yuan family cars. In other words, Calterah not only realized domestic substitution, but also expanded the entire market cake.

By 2025, Calterah's domestic market share reached 31.1%, ranking fourth in the world, with cumulative shipments exceeding 30 million units, and more than 30 cooperative car companies covering more than 300 vehicle models. The list of domestic customers is very long: BYD, Geely, Changan, Chery, NIO, Leapmotor. Interestingly, it has also expanded into Europe, becoming the first domestic millimeter-wave radar chip supplier for the global platforms of two leading Tier 1 suppliers, and its products have been used in Volvo and Rivian models. It was almost unthinkable before that Chinese automotive-grade chips could enter the pre-installed supply chain of overseas OEMs.

Valuation Reached 10 Billion Yuan After 11 Rounds of Financing

Since its establishment, Calterah has completed 11 rounds of financing in total, with a luxurious investor lineup. Almost all well-known institutions in the chip track are on the shareholder list: China Integrated Circuit Industry Investment Fund Phase II, CICC Capital, Walden International, Noah Holdings, GAC Capital, Pudong Venture Capital, Silergy, Zhangjiang Sci-Tech Investment, Greater Bay Area Science and Innovation Fund.

The first to enter was Silergy, which invested in the angel round in November 2015. At that time, Calterah had just completed the tape-out of its first 77GHz chip, and the company's registered capital was only 50,000 yuan. As of the date of signing the prospectus, Silergy is still an important shareholder. From a registered capital of 50,000 yuan to a valuation of 10 billion yuan, it has taken a long 12-year journey in the capital market.

After that, the financing pace accelerated year by year:

2017 Series A: Frees Fund, Zhongguancun Xingye Investment entered the market;

2019 Series B: CICC Capital, Shaanxi Hongchuang, China Renaissance Capital, China Mobile Innovation Industry Fund, Golris Capital and others followed up;

July 2020 Strategic Round: Shangqi Capital, GAC Capital, Nest Capital joined in;

September 2021 Series C: Led by SDIC Investment, with participation from Langmafeng Venture Capital, BOCOM International, China Renaissance New Economy Fund, the total amount reached several hundred million RMB;

January 2022 Series C+: Fosun Chuangfu, China Merchants Capital, Noah Holdings, Yinggang Capital, Juntong Capital continued to increase their stakes.

July 2024 Series D: Led by China Integrated Circuit Industry Investment Fund Phase II, with participation from Guoxin Venture Capital, Fu Venture Capital, SDIC Investment, Walden International and others, the total amount reached several hundred million yuan.

The financing pace was particularly intensive during the reporting period.

In the capital increase in January 2025, China Integrated Circuit Industry Investment Fund Phase II contributed 50 million yuan, Shanghai Guoxin contributed 49 million yuan, Advanced Manufacturing Phase II contributed 30 million yuan, Chuangxin Chuangke contributed 20 million yuan, and Hong Kong Wisteria contributed 1 million US dollars, injecting a total of about 179 million yuan. In the E1 round in September 2025, Leading Area Fund, Zhangke Yaokun, Ruishi Phase IX, Greater Bay Area Fund and others contributed a total of 208 million yuan. In the E2 round in 2026, Huaxin Dingxin invested 300 million yuan alone, CCTV Converged Media, Ruishi Phase VII, and Guotou Xinzhi each contributed 100 million yuan, raising about 955 million yuan in a single round. The three rounds totaled about 1.34 billion yuan.

Among them, Huaxin Dingxin obtained a 3.63% stake with 300 million yuan in the E2 round, and the estimated valuation is about 8.26 billion yuan. For this IPO, the planned fundraising is 3.489 billion yuan with no less than 25% of the equity to be issued. Based on this calculation, the issuance valuation is about 14 billion yuan. In other words, less than a year after the E2 round institutions entered the market, the floating profit on the book has reached 70%.

Parallel to the intensive financing transfusion, there were also a large number of old share transfers.

From September 2025 to March 2026, 23 new shareholders including Leading Area Fund, Huaxin Dingxin, and CCTV Converged Media concentratedly took shares or transferred old shares. The price for capital increase and share subscription was 292.25 yuan per share, and the price for old share transfer ranged from 176.99 yuan per share to 287.93 yuan per share.

CHEN Jiashu directly holds 12.1376% of the shares, and indirectly controls 20.8730% via Nanchang Xichuang, jointly controlling 33.0106% of the voting rights, making him the actual controller of the company. There is no single shareholder holding more than 30% of the shares. A highly dispersed equity is not uncommon for chip companies on the Science and Technology Innovation Board. In addition, China Integrated Circuit Industry Investment Fund Phase II holds 0.6266% of the shares, a relatively low proportion. However, the endorsement of the national team is far more symbolic than financially meaningful.

Not Profitable Yet, All Funds Burned in R&D

From the data, Calterah's growth momentum is indeed strong.

From 2023 to 2025, the operating revenue rose from 206 million yuan to 303 million yuan, and then to 632 million yuan, with a three-year compound annual growth rate of 75.28%. It more than doubled in 2025 alone, with a year-on-year increase of 108.44%. In the first quarter of 2026, the single-quarter revenue reached 154 million yuan, and automotive-grade millimeter-wave chips contributed more than 99% of the revenue.

The flip side of high growth is continuous cash outflow.

From 2023 to 2025, the attributable net losses were 323 million yuan, 334 million yuan and 193 million yuan respectively; in the first quarter of 2026, it lost another 60.28 million yuan. The cumulative loss in three and a half years exceeded 900 million yuan. As of the end of March 2026, the cumulative unrecovered loss was 172 million yuan.

All of this money was spent on R&D.

In the automotive-grade chip industry, the R&D cycle is long, the verification standards are strict, and the supply chain onboarding of OEMs usually takes at least three to four years, so burning cash is the norm. During the reporting period, Calterah's cumulative R&D investment exceeded 1.039 billion yuan. The R&D expense in 2023 was 304 million yuan, 1.5 times the revenue of that year. By 2025, this ratio dropped to 58.55%, which is still at a high level. As of March 2026, the company had 275 R&D personnel, accounting for 65.63% of the total employees, which means two-thirds of the company's personnel are engaged in R&D. With long-term R&D investment exceeding revenue, it is almost inevitable to lose money temporarily.

But the gross profit margin is not low. The gross profit margin of main business in each period of the reporting period was 47.80%, 43.81%, 47.25% and 48.97% respectively, continuously higher than the average level of domestic peers. In other words, the products themselves are profitable, but all the profits are burned by R&D.

Objectively speaking, the company's current loss is narrowing. The loss in 2025 was 193 million yuan, 42% narrower than the 334 million yuan loss in 2024. Operating cash flow also improved from -249 million yuan in 2024 to -18 million yuan in 2025, almost breaking even. The inflection point of profitability seems to be approaching.

But whether the inflection point can really come depends on two variables.

The first variable is customers. The revenue proportion of the top five customers in each period of the reporting period was as high as 99.77%, 99.07%, 99.90% and 99.97% respectively. In 2025, the revenue from BYD, the largest end customer, accounted for more than 50%. The prospectus admits that since 2026, affected by the downstream OEMs' supply chain diversification strategies such as introducing second suppliers, the revenue from this customer has shown signs of continuous pressure. This may mean that BYD has begun to compare products from different suppliers. The onboarding cycle for new automotive-grade chip customers is as long as three to four years. Once the core customer reduces procurement, the revenue may shrink significantly.

The second variable is the supply chain. The procurement proportion of the top five suppliers rose all the way from 54.50% in 2023 to 80.16% in the first quarter of 2026, among which the procurement proportion of one major supplier rose from 37.20% to 58.68%. The proportion of overseas procurement is above 50% in each period, and EDA tools and some interface IP also rely on overseas suppliers.

Both customer concentration and supply chain concentration bring risks on two fronts. This is also the reason why Calterah must go public: it needs sufficient funds to expand production, expand customer base, and reduce dependence.

This article is from the WeChat official account "Dong 40 Tiao Capital" (ID: DsstCapital), written by Li Man, and published with authorization from 36Kr.