"Hong Kong Stock Exchange's first GPU stock" annual report submitted, losing more but earning more?

Ask AI · How does Bairen Technology, which is rapidly growing, balance R&D investment and profitability?

This is Bairen Technology’s first annual report after listing on the Hong Kong stock market.

On January 2, 2026, amid the lingering chimes of the New Year, Bairen Technology officially rang the bell and listed on the Hong Kong Stock Exchange, becoming the first GPU target in the Hong Kong market, also marking the first shot of the new year’s hard tech IPO.

Recently, the company officially released its 2025 full-year performance announcement—coinciding with key milestones such as the large-scale implementation of the global artificial intelligence industry and the continued release of domestic computing power replacement demands. Bairen Technology achieved rapid growth in key indicators driven by breakthroughs in the commercialization of core products: in 2025, the company’s revenue first exceeded 1 billion yuan, reaching 1.03B yuan, a year-on-year increase of 207.2%; gross profit also rose to 557 million yuan, up 210.8% year-on-year; gross profit margin remained stable at 53.8%. Revenue from core intelligent computing solutions contributed over 90%, and solid business foundations brought considerable performance growth.

However, behind this impressive growth curve, the company’s losses also expanded proportionally, with an annual net loss of 16.49B yuan, a 972.3% increase year-on-year. This figure was mainly affected by non-operational one-time factors such as the fair value increase of redemption liabilities related to investor redemption rights and listing expenses. The company’s redemption rights have been permanently terminated after listing, and related liabilities have been converted into equity, so such gains and losses will no longer occur in the future; but even excluding these one-time factors, the adjusted loss still reached 874 million yuan, a 13.9% increase year-on-year.

Bairen Technology’s team told Beijing Business Today that the high growth in performance this year stems from the large-scale implementation and delivery of intelligent computing solutions. The company is still in a high-speed growth phase, prioritizing strategic resource allocation toward long-term capability building. The current focus is on R&D investment and market expansion, leveraging ample financial reserves to increase investment in next-generation chip and system-level technology R&D, aiming to seize the market opportunity in AI inference computing power.

1. The Boom Period of Demand Growth

The comprehensive explosion of large model technology is reshaping the global intelligent computing chip market landscape, with training and inference becoming the industry’s two core application scenarios.

Early industry demand focused on the pre-training stage of large models, where the development of models with hundreds of billions or trillions of parameters required large-scale server clusters to provide super-strong computing support. High bandwidth and high-speed interconnection became core requirements. As the functionality and application scenarios of large models continue to mature, the frequency of terminal use and coverage expand rapidly, leading to an explosive growth in computing power demand for inference scenarios: while the computing power needed for a single task is relatively moderate, the overall consumption continues to rise due to high concurrency, and low latency has become a key requirement for inference.

According to the prospectus of Bairen Technology citing CIC data, in the next one to two years, investment growth in inference chips is expected to surpass training chips, becoming the main driver of industry growth.

From a market size perspective, the global intelligent computing chip industry has maintained rapid growth: from $6.6 billion in 2020 to $119 billion in 2024, with a compound annual growth rate (CAGR) of 106%. It is expected to surpass $585 billion by 2029, with a CAGR of 37.5% from 2024 to 2029. China, as a key global AI market, continues to outperform globally: from $1.7 billion in 2020 to $30.1 billion in 2024, with a CAGR of 105%, projected to reach $201.2 billion by 2029, with a CAGR of 46.3% from 2024 to 2029.

Among them, GPGPU (general-purpose graphics processing units), due to their versatility, flexibility, and mature software ecosystem, have become the mainstream choice for intelligent computing chips. In 2024, GPGPU accounted for 92% of the global intelligent computing chip market and 78.1% in China; China’s GPGPU market is expected to grow at a CAGR of 49% from 2024 to 2029, with significant growth potential.

According to the prospectus and public data, as industry competition shifts from single-chip performance to full-stack competition involving hardware iteration, software ecology, and cluster system collaboration, large-scale computing clusters have become an industry trend. High R&D investment and ecosystem construction are also essential for companies to establish a foothold.

2. Product Implementation, Waiting for Bloom

In 2025, Bairen Technology’s breakthrough is visually reflected in the doubling of revenue, but behind the numbers is an effective improvement in business implementation capability.

According to the company’s information provided to Beijing Business Today, the core support for the high revenue growth is the large-scale deployment and delivery of intelligent computing solutions—its flagship GPUs, BR106 and BR166, have achieved full form mass production. The company has successfully upgraded from a single-chip supplier to a provider of comprehensive solutions capable of supporting AI compute clusters with thousands of cards. Landmark projects such as 2048-card optical interconnect super-node clusters have been successfully implemented, with clients including national-level computing platforms, telecom operators, and leading AI companies.

Comparatively, Mooresoft, a leading domestic GPU company listed on the STAR Market, also delivered impressive results: according to its 2025 performance quick report, the company’s annual revenue increased by 243.37% year-on-year, while its loss narrowed by 36.70%. Although the two top companies are at different stages of narrowing losses and exploring profitability, the high growth in revenue confirms the industry’s strong momentum.

As a technology enterprise in a rapid expansion phase, Bairen Technology’s R&D and operational investments are naturally increasing in tandem. In 2025, R&D expenditure reached 1.48B yuan, up 78.5%. The funds are mainly invested in next-generation BR20X series chips, full-stack software optimization, and optical interconnect system technology R&D, targeting AI inference power needs; sales and marketing expenses amounted to 57.5 million yuan, up 11.6%, used for expanding sales teams and downstream market development; general and administrative expenses rose to 332 million yuan, up 35.8%, mainly due to team expansion and increased operational costs, typical of a growth-stage company.

On the ecosystem adaptation front, Bairen Technology has completed rapid integration with mainstream large models such as DeepSeek, Zhipu GLM, Alibaba Qianwen, and Tencent Moxuan, connecting chips, large models, and cloud collaboration chains, greatly lowering the barriers for domestic computing power deployment. In the current industry shift toward system-level efficiency, the launch of its optical interconnect super-node solution precisely targets the key transition from single-point performance comparison to overall system competition.

Wang Peng, associate researcher at Beijing Academy of Social Sciences, told Beijing Business Today that domestic GPU companies are at a critical development juncture. They need to rely on scaled business to hedge R&D rigid costs, build mature software ecosystems to reduce customer migration costs, and gradually replace single hardware sales with integrated compute services.

Wang Peng believes that genuine market demand is the core support for market growth. The synergy of industry chain clusters and scale effects will directly determine whether companies can absorb high R&D costs and achieve sustainable profitability.

Beijing Business Today reporters Tao Feng and Wang Tianyi