Domestic GPU vendor MetaX reported its first-ever profitability alongside the mass production of its 7nm graphics processors, reaching a market valuation of 270 billion RMB.
On August 30, Metax, one of China’s “Four Little Dragons” of domestic GPUs, released its half-year report, revealing revenue of 1.324 billion RMB, a 44.67% year-on-year increase. For the same period, the company recorded a net profit attributable to shareholders of 612 million RMB, marking its transition from losses to profitability.
Metax is the first among the four leading domestic GPU firms to achieve positive earnings, signaling that the domestic GPU development pathway is viable.
Last December, the company listed on the STAR Market. Its share price briefly surpassed 1,000 RMB, pushing its market capitalization above 400 billion RMB. It currently maintains a market cap of approximately 270 billion RMB.
The other three members of the “Four Little Dragons” are Moore Threads, Biren Technology, and Enflame Technology. Three of these companies have already listed on A-share or Hong Kong exchanges, with Enflame expected to join the STAR Market later this year. While none of the other three have yet turned a profit, their financials are improving sharply. Moore Threads, which holds the largest market cap among them, reported H1 revenue of 1.736 billion RMB, up 147.42% year-on-year. Its net loss narrowed significantly to -11.5631 million RMB, compared to a 271 million RMB deficit in the first half of last year, making near-term profitability highly probable.
This industry-wide progress is underpinned by continuous technological advancements and surging sales volumes. Metax attributed its revenue growth primarily to widespread customer adoption and procurement of its products and services throughout the reporting period, which drove a substantial increase in GPU shipments and overall revenue scale.
During the first half of the year, Metax invested 525 million RMB in research and development, representing 39.65% of its total revenue. Guided by its “mass-produce one generation, develop one generation, plan one generation” strategy, the company announced that its Xi Yun C600 GPU successfully entered mass production in May 2026 and is poised to become its next-generation flagship product.
The C600 features 144GB of HBM3e memory with a bandwidth exceeding 3.35 TB/s. It delivers a peak FP8 compute capacity of approximately 1,000 TFLOPS and an INT8 compute capacity of up to 2,000 TOPS. The chip is fabricated using a fully domestic 7nm manufacturing process.
Meanwhile, the next-generation flagship C700 is currently in development. It aims to deliver substantial improvements in compute, memory, interconnect, and energy efficiency, positioning it close to the performance level of NVIDIA’s H100 series. Initiated in April 2025, the C700 project is now in the hardware and software procurement and design phase. Core chip design and functional verification are largely complete, with ongoing deep-dive performance tuning and iterative physical implementation optimization based on advanced domestic fabrication processes.
Copyright (C) Mydrivers.com, All Rights Reserved.