Japanese Media: China's "NVIDIA" Achieves First Quarterly Profit After Listing

Moorer Technology, a Chinese graphics processing semiconductor company, has recently released its full-year 2025 financial results and the financial report for the period from January to March 2026.

The company’s products cover AI computing acceleration, graphics rendering, and video processing. It is regarded as the primary alternative to NVIDIA in China, having listed on the STAR Market of the Shanghai Stock Exchange on December 5, 2025.

According to data from January to March 2026, revenue increased by 155.35% year-on-year to RMB 738 million (approximately JPY 17 billion). Net profit attributable to shareholders of the parent company reached RMB 29.36 million (about JPY 680 million), marking the first quarterly profit since its listing, up by RMB 142 million (around JPY 3.3 billion) year-on-year.

Notably, although after excluding non-recurring gains and losses such as government subsidies, the adjusted net profit remained at a loss of RMB 54.28 million (approximately JPY 1.2 billion), this deficit narrowed by 60.10%. Full-year revenue rose by 243.37% to RMB 1.505 billion (about JPY 350 billion), with gross profit increasing by 218.43% to RMB 987 million (approximately JPY 230 billion). Net profit attributable to shareholders of the parent company decreased by 38.16%, while the loss after excluding extraordinary income narrowed by 33.38%.

In terms of R&D investment, the company spent RMB 1.305 billion (about JPY 30 billion) on R&D throughout 2025, accounting for 86.68% of total revenue. With the surge in demand for AI computing power, Moorer Technology’s commercialization process continues to accelerate.

In March this year, the company announced it had secured a contract worth RMB 660 million (approximately JPY 15 billion) for an intelligent computing cluster called “KUAE.” This project accounted for nearly 90% of sales during the January–March 2026 period. The cluster is an intelligent computing system capable of training using 10,000 GPUs, enabling end-to-end training of large language models with up to one trillion parameters.

Moorer Technology is among the few suppliers in the market that have successfully commercialized large-scale GPU clusters with 1,000 and even 10,000 GPUs. Based on its next-generation GPU architecture “Huagang,” the company is expected to drive the development of ultra-large-scale intelligent computing clusters featuring 100,000 GPUs, as well as core technologies such as super nodes.

Original article: toutiao.com/article/1864593668979788/

Disclaimer: The views expressed in this article are those of the author(s).