刚刚！百度官宣：昆仑芯赴港IPO

Group 1 - Baidu announced that Kunlun Chip has submitted a listing application to the Hong Kong Stock Exchange for approval to list and trade its shares on the main board [3] - The proposed spin-off will involve a global offering of Kunlun Chip shares, including a public offering in Hong Kong and a placement to institutional and professional investors [3] - As of the announcement date, Kunlun Chip is a non-wholly owned subsidiary of Baidu, and it is expected to remain a subsidiary after the spin-off [3] Group 2 - Kunlun Chip, formerly Baidu's Intelligent Chip and Architecture Department, completed its first round of independent financing in April 2021, with an initial valuation of approximately 13 billion [5] - As of July 2025, Kunlun Chip has undergone a D round of financing, with investors including various capital firms and funds [5] - Baidu holds a 59.45% stake in Kunlun Chip, while other partners hold smaller percentages [5] Group 3 - The new generation of Kunlun Chips, M100 and M300, was announced at the Baidu World Conference, with M100 designed for large-scale inference and expected to launch in 2026, while M300 targets ultra-large multimodal model training and inference tasks, expected in 2027 [5] - Baidu has already deployed a cluster of 30,000 Kunlun Chips, capable of supporting multiple trillion-parameter model trainings, with plans to expand to a million-chip scale [5] Group 4 - Morgan Stanley predicts that Kunlun Chip's revenue could reach 8.3 billion by 2026, representing a sixfold increase compared to the expected revenue in 2025 [6] - The AI chip industry is currently experiencing a wave of IPOs, indicating a growing interest and investment in this sector [7] Group 5 - Suiyuan Technology, one of the domestic GPU companies, has appointed CICC as its advisory institution for its listing process, indicating a strategic move based on market conditions [10] - Suiyuan Technology focuses on AI cloud computing products and has a registered capital of 387 million [10] - Hanbo Semiconductor, established in December 2018, specializes in high-end GPU chips and has a strong R&D team with over 500 members, many of whom have extensive experience from leading companies like AMD and NVIDIA [13]