“Primera acción GPU de la Bolsa de Hong Kong” presenta su informe anual, ¿cuanto más pierde, más gana?

This is the first annual report published after Biren Technology’s listing on Hong Kong’s stock exchange.

On January 2, 2026, as the echoes of the New Year bell rang out, Biren Technology officially struck the bell to begin trading on the Hong Kong Stock Exchange, becoming the first GPU-related stock in the Hong Kong market, and firing the first shot of the New Year’s hard-tech IPO.

Recently, the company officially released its full-year 2025 performance announcement—at a time when key milestones are being realized across the global artificial intelligence industry on a large scale, and demands for domestically produced computing power replacement continue to be released. Biren Technology achieved rapid growth in key figures such as revenue through commercial breakthroughs of its core products: in 2025, the company’s revenue first exceeded the 1 billion mark, reaching RMB 1.03B, up 207.2% year over year; gross profit rose in parallel to RMB 557 million, up 210.8% year over year; gross margin remained stable at 53.8%; revenue from core intelligent computing solutions contributed more than 90%, and the solid foundation of its business brought substantial performance growth.

However, behind the impressive growth curve, the company’s loss scale also expanded at the same time. The loss recorded on its accounts for the full year reached RMB 16.49B, up 972.3% year over year. This figure is mainly affected by non-operating one-off factors such as an increase in the fair value of redemption liabilities related to investors’ redemption rights and listing expenses. The company’s redemption rights were permanently terminated after listing, and the related liabilities were converted into equity at the same time, so no such profit or loss will arise in the future. Even so, after excluding the above one-off factors, the adjusted loss still stood at RMB 874 million, widening 13.9% year over year.

The Biren Technology team told Beijing Business Today reporter that the year’s high growth in performance came from the large-scale roll-out and delivery of its intelligent computing solutions business. The company is still in a phase of rapid growth. Strategic resources are prioritized for long-term capability building. At the current stage, its focus is on R&D investment and market expansion, and with sufficient funding reserves, the company is stepping up R&D for the next-generation chips and system-level technologies to seize early opportunities in the computing power market for the AI inference era.

1. The windfall period of rising demand

The full-scale surge of large-model technology is reshaping the global intelligent computing chip market landscape, with training and inference becoming the two key core application scenarios in the industry.

Early industry demand was concentrated on the large-model pre-training stage. Developing models with trillion-scale and even ten-trillion-scale parameters requires massive server clusters to provide exceptionally strong computing power support; high bandwidth and fast interconnection are core requirements. As large-model functions and application scenarios continue to mature, the frequency of terminal usage and coverage rapidly expand, and computing-power demand for inference scenarios experiences explosive growth: while computing demand per single task is relatively moderate, the total computing power consumption driven by high concurrency keeps rising, and low latency also becomes a core requirement for inference scenarios.

According to materials from CIC (China Insights Consulting) cited in Biren Technology’s IPO prospectus, over the next one to two years, the investment growth rate in inference chips is expected to exceed that of training, becoming the core driving force for industry growth.

In terms of market size, the global intelligent computing chip industry maintains leapfrogging growth. From 2020 to 2024, market size rapidly increased from 6.6 billion USD to 119 billion USD, with a CAGR of 106%. It is expected to surpass 585 billion USD in 2029, and the CAGR from 2024 to 2029 is expected to remain at 37.5%. As the world’s core AI market, China’s growth rate continues to run ahead of the global average. From 2020 to 2024, market size grew from 1.7 billion USD to 30.1 billion USD, with a CAGR of 105%. It is expected to reach 201.2 billion USD by 2029, and the CAGR from 2024 to 2029 is 46.3%.

Among them, GPGPU (general-purpose graphics processing unit) has become the mainstream choice for intelligent computing chips thanks to its general-purpose nature, flexibility, and mature software ecosystem. In 2024, GPGPU’s share in the global intelligent computing chip market reached 92%, and in the China market it was 78.1%. China’s GPGPU market is expected to have a CAGR of 49% from 2024 to 2029, indicating significant growth potential.

According to the prospectus and public information, in the context of industry competition shifting from solely comparing chip performance to full-stack competition across dimensions such as hardware iteration, software ecosystems, and cluster system coordination, large-scale computing power clusters are becoming an industry trend. High R&D spending and ecosystem building are also becoming a rigid need for enterprises to be competitive.

2. Product implementation—waiting for the blooms to open

Biren Technology’s breakthrough in 2025 is directly reflected in the doubling growth of revenue figures, and behind the numbers is an effective improvement in business implementation capabilities.

According to the company’s introduction to Beijing Business Today reporter, the core support for the high growth in revenue is the large-scale rollout and delivery of intelligent computing solutions. Its two flagship GPUs, BR106 and BR166, achieved full-form-factor mass production. The company also successfully upgraded from a single-chip supplier to an overall solution provider capable of offering k-card-level intelligent computing clusters. Benchmark projects such as 2048-card optical interconnect super-node cluster have been successfully delivered, and customer coverage includes national-level computing power platforms, telecommunications operators, and leading AI enterprises.

In a horizontal comparison, China’s domestic GPU leader Moore Threads delivered impressive results in the same period. According to its 2025 annual performance quick report announcement, the company’s annual revenue increased 243.37% year over year, while its loss narrowed by 36.70% year over year. The two leading enterprises have different rhythms in narrowing losses to explore profitability, but the mutual confirmation from their high growth in revenue testifies to the strong momentum in the industry.

As a technology company in a phase of rapid expansion, Biren Technology’s R&D and operating investment must rise in tandem. In 2025, the company’s R&D expenses reached RMB 1.48B, up 78.5%. It is understood that the related funds are mainly invested in next-generation BR20X series chips, full-stack software optimization, and optical interconnect system technology R&D, with targeted adaptation to computing power needs in the AI inference era. Sales and marketing expenses were RMB 57.5 million, up 11.6%, used to expand the sales team and develop downstream markets. General and administrative expenses rose to RMB 332 million, up 35.8%. The core reason is that team-size expansion increased manpower and operating costs, which falls under ordinary operating expenditures for a company in its growth stage.

On the ecosystem adaptation front, Biren Technology completed rapid connectivity with mainstream large models such as DeepSeek, Zhipu GLM, Alibaba Qwen, and Tencent Hunyuan, and connected the cooperation chain among chips, large models, and cloud services, greatly lowering the deployment threshold for domestically produced computing power. At a time when industry competition is shifting toward system-level efficiency, the launch of its optical interconnect super-node solution precisely targets a key turning point in which the industry shifts from competing on single-node performance to competing across the overall system.

Wang Peng, associate researcher at the Beijing Academy of Social Sciences, said in an interview with Beijing Business Today reporter that domestic GPU companies are at a crucial crossroads of development. Companies need to rely on large-scale business to hedge the rigid R&D spending, build a mature software ecosystem to reduce customers’ switching costs, and gradually replace single hardware sales with integrated computing power services.

Wang Peng believes that real commercial demand is already the core support for market growth, and that coordination among industry-chain clusters and scale effects will directly determine whether enterprises can absorb high R&D investment and achieve sustainable profitability.

Beijing Business Today reporter Tao Feng, Wang Tianyi