Denglin Technology Raises HK$1.41 Billion in Hong Kong GPU IPO, Revenue Surges

Engineers working in a Chinese semiconductor lab inspecting GPU chips and testing equipment.

Original source: technode.com — In Nvidia’s Shadow, a Chinese Chipmaker Takes a Different Route

Denglin Technology, a Chinese maker of general-purpose GPUs for AI computing, went public in Hong Kong on November 10, raising HK$1.41 billion. According to technode.com, the company’s revenue reached RMB 200.4 million in the first four months of 2026—already exceeding its full-year 2025 total. But despite rapid growth, Denglin holds just 1% of China’s enterprise AI GPGPU market, far behind Nvidia and AMD.

What happened

Founded in Shanghai in 2017 by four former Vivante colleagues, Denglin Technology initially focused on smart-city applications. Its first chip samples arrived in early 2021. In 2024, the company moved its headquarters to Suzhou to be closer to the semiconductor supply chain.

Denglin’s GPUs were designed with an emphasis on cost and energy efficiency. The company chose smartphone-style low-power double data rate (LPDDR) memory instead of the high-bandwidth memory (HBM) common in high-end accelerators. LPDDR is cheaper and more energy-efficient, but its lower bandwidth can limit performance for large AI models. To address this, Denglin added more on-chip static random-access memory (SRAM) and developed software to reuse data, reducing the need to fetch it from external memory.

By October 2026, Denglin’s KS GPUs had completed compatibility and efficiency testing for the DeepSeek-V4.1-Flash-1017 AI model, demonstrating their ability to support major AI workloads. The company now plans to move into higher-performance AI computing, with its first Minsky IV GPU version expected in December 2026 and a second version with advanced memory (such as HBM3e) targeted for late 2027.

Denglin’s business focus has shifted. In 2023, about 80% of its GPGPU product revenue came from smart-city projects. By early 2026, enterprise private cloud made up 83% of revenue, while smart cities contributed just 12.2%. The company sold 10,246 KS inference cards in the first four months of 2026, generating RMB166 million.

Despite sales growth—RMB27.9 million in 2023, RMB82.6 million in 2024, RMB171.3 million in 2025, and RMB200.4 million in just the first four months of 2026—Denglin remains a small player. Frost & Sullivan ranked it third in China’s enterprise AI GPGPU market by 2025 revenue, but its share was just 1%, compared to Nvidia’s 83.1% and AMD’s 12.3%.

Profitability remains a challenge. Denglin recorded net losses of RMB332 million in 2023, RMB296.5 million in 2024, and RMB281 million in 2025, mainly due to high R&D spending. In the first four months of 2026, net loss narrowed to RMB16.5 million, and adjusted net profit reached RMB20.6 million after accounting for share-based payments and listing expenses. However, the company’s customer base is concentrated: its five largest customers accounted for 90.4% of revenue in early 2026, with the biggest contributing 38%.

Denglin’s Hong Kong listing was under Chapter 18C, which allows specialist-technology companies to go public before reaching the HK$250 million annual revenue threshold. The company plans to use 35% of proceeds to upgrade GPU lines, 25% for new product categories, and 20% on software tools.

Why it matters for researchers

Denglin’s story highlights the shifting landscape for AI hardware startups in China. The company’s focus on cost and power efficiency over peak performance signals a market for specialized, energy-conscious AI solutions, especially in enterprise and edge applications. For researchers, this illustrates the practical trade-offs between memory architecture, energy use, and model performance in real-world deployments. Denglin’s rapid revenue growth—despite a small market share and ongoing losses—also shows that there is investor interest and industry demand for alternatives to dominant players like Nvidia and AMD. However, the high R&D costs and concentrated customer base underline the challenges of scaling and sustaining a hardware startup in a competitive sector. For PhD students and early-career researchers, Denglin’s trajectory is a case study in how technical decisions, market focus, and funding environments shape career opportunities and research directions in semiconductors and AI hardware.

Where to dig deeper

This article summarises public reporting. Facts as reported by technode.com; read the original for full details.

Share this story · 分享

X LinkedIn Facebook 微博 Weibo

Leave a Reply

Your email address will not be published. Required fields are marked *