Chinese Chipmaker Moore Threads Gets Fast Track Approval to $1.1 Billion IPO
By Liu Peilin
Chinese AI chipmaker Moore Threads Intelligent Technology (Beijing) Co. Ltd. has won speedy approval for a STAR Market initial public offering (IPO), clearing the way for a major fundraising despite deep losses and U.S. sanctions.
The Shanghai Stock Exchange’s listing committee cleared the IPO on Friday, less than three months after receiving the firm’s application. The company, founded by a former Nvidia Corp. executive, aims to raise 8 billion yuan ($1.1 billion) to fund research and development and shore up working capital, according to its prospectus.
The fast-track greenlight underscores Beijing’s urgency to build a self-sufficient semiconductor industry, particularly in AI, as U.S. restrictions tighten. Moore Threads illustrates a broader trend among sanctioned Chinese chipmakers: leaning on large-scale computing clusters to offset weaker single-chip performance while burning cash on an uphill path to profitability.
Founded in October 2020, Moore Threads grew revenue from 45.8 million yuan in 2022 to 432 million yuan in 2024. However, its net losses have been severe, totaling 4.61 billion yuan over the same period after deducting nonrecurring items. The company’s R&D spending from 2022 to 2024 reached 3.8 billion yuan — more than six times its cumulative revenue.
The company’s growth ambitions have been hampered by U.S. sanctions. Washington added Moore Threads to its Entity List in October 2023, cutting off access to Taiwan Semiconductor Manufacturing Co. Ltd. for producing high-performance chips. In its prospectus, the company said this has forced it to find new suppliers, potentially adding costs and slowing R&D.
In response, Moore Threads has shifted focus to GPU clusters. Revenue from its AI cluster products reached 184 million yuan in 2024, accounting for more than 42% of total sales. The firm launched a 1,000-card cluster in December 2023 and rolled out a 10,000-card cluster by the end of 2024.
Moore Threads’ latest S5000 chip delivers peak performance of 32 TFlops at FP32 precision, placing it between Nvidia’s A100 and H100 GPUs. The company did not disclose benchmarks for newer precision formats critical for large AI model training.
Moore Threads was founded by Zhang Jianzhong, a former general manager of Nvidia China. Its backers include HongShan Capital Group and a fund backed by China Mobile Ltd.
caixinglobal.com is the English-language online news portal of Chinese financial and business news media group Caixin. Global Neighbours is authorized to reprint this article.
Image: William W. Potter – stock.adobe.com