◎记者 严晓菲 这种看似矛盾的现象，源于各方对国产GPU市场广阔前景的共识。据国际数据公司（IDC）预测，到 2029年，中国加速服务器市场规模将超过千亿美元。交银国际研报认为，国产算力芯片行业未来将呈现 百花齐放态势，技术上取得长足进步，包括晶圆代工、存储器、半导体设备等在内的国产算力产业链有 望持续受益。 12月以来，国产GPU领域的两位主要玩家——摩尔线程和沐曦股份，先后叩开了资本市场的大门。这两 家成立均仅5年的公司，正以超预期的市场表现，点燃投资者的巨大热情。 沐曦股份上市首日便坐上A股史上"最赚钱新股"的宝座，股价一度飙涨到逼近900元/股，总市值也已突 破3300亿元。在此之前，摩尔线程在上市后短短几个交易日内，股价最高涨到941.08元/股。 在这两家公司的背后，是两份密密麻麻的出资方名单：沐曦股份股东超过120家，摩尔线程股东超过80 家。在这场从"非共识"到"共识"的转变中，这些"耐心资本"为何能如此守得住寂寞？ 高光时刻： 亏损与高估值反差折射广阔前景 在国产GPU的高光时刻，难以无视的是这两家公司依旧严峻的财务现实。根据招股书，它们都仍处于持 续亏损的状态。 沐曦股份2022至2024 ...

Core Viewpoint - The article discusses the strategic cash management approach of Moore Threads, emphasizing its commitment to responsible fund usage while pursuing ambitious GPU technology development in China, aiming for a breakthrough in the AI computing sector [6][11][24]. Group 1: Cash Management Strategy - Moore Threads plans to utilize up to 7.5 billion yuan of idle raised funds for cash management, ensuring it does not affect project implementation or fund safety [7][11]. - The company has a clear investment plan for the raised funds, with projects including AI chip development and a total investment of 800 million yuan over three years [8][9]. - The cash management strategy involves conservative investments in safe, liquid products like deposits, reflecting a responsible approach to fund management [12][13]. Group 2: GPU Technology Development - Moore Threads aims to achieve domestic GPU replacement, positioning itself as a key player in China's AI computing landscape [14][24]. - The company has chosen the challenging path of developing a general-purpose GPU, a route that has seen limited success for other Chinese firms [19][20]. - Its proprietary MUSA ecosystem integrates various capabilities, making it a strong contender against international competitors like NVIDIA [21][22]. Group 3: Market Position and Growth - The launch of the MTT S80 graphics card marked a significant achievement, becoming the only domestic card capable of running the popular game "Black Myth: Wukong" smoothly [27][30]. - Moore Threads has established a comprehensive product system across both consumer and enterprise markets, with a notable shift towards AI computing products expected to account for 77.63% of revenue by 2024 [34][31]. - The Chinese AI chip market is projected to grow significantly, with an expected compound annual growth rate of 53.7% from 2024 to 2029, indicating a favorable environment for Moore Threads [48][49]. Group 4: Competitive Landscape - The article highlights the challenges posed by NVIDIA's CUDA ecosystem, which has become a critical barrier for competitors [35][36]. - Moore Threads' MUSA architecture is designed to be independent of CUDA, allowing for easier migration for developers, thus enhancing its competitive edge [39][40]. - The anticipated growth of the GPU market, alongside Moore Threads' rapid revenue increase from 0.46 billion yuan in 2022 to 7.02 billion yuan in the first half of 2025, underscores the company's potential for future success [49][50].

Core Viewpoint - Moore Threads officially listed on the Shanghai Stock Exchange's Sci-Tech Innovation Board on December 5, becoming the first fully functional GPU company to enter the A-share market in China, marking a significant milestone in the company's development and the advancement of China's computing power [2] Group 1: Company Milestones - The company achieved a listing price of 650 CNY per share, opening with a remarkable increase of 468.78%, and reaching a peak of 688 CNY during trading, allowing investors to earn over 267,000 CNY per share [4] - The founder and chairman, Zhang Jianzhong, emphasized the company's commitment to independent innovation and the development of a unified system architecture, MUSA, which supports AI computing acceleration, graphics rendering, physical simulation, and scientific computing on a single chip [6][11] - The company has successfully launched four generations of GPU architectures and five chips, establishing a comprehensive product layout from chips to clusters, serving various sectors including government, enterprises, data centers, and consumer terminals [9] Group 2: Financial Performance - From 2022 to June 2025, the company invested over 4.3 billion CNY in R&D, with over 77% of its workforce dedicated to research, resulting in a compound annual growth rate of 208.44% in revenue from 2022 to 2024 [12] - In the first half of 2025, the company reported revenue of 702 million CNY, surpassing the total revenue of the previous three years, with AI computing business contributing 94.73% of this revenue and a gross margin of 69.14% [12] Group 3: Industry Context - The listing of Moore Threads reflects the increasing support and determination from policies and market players for the development of high-tech and hard-tech enterprises under national strategies [14] - The company aims to contribute to the national strategy of enhancing computing power as a strategic resource, emphasizing the importance of GPU technology in the context of global technological competition [11][13] - The company plans to use the funds raised from the listing to focus on the development of next-generation AI training and inference chips, graphics chips, and AI SoC chips, aiming to strengthen its technological leadership in AI computing and graphics [20]

Core Viewpoint - Moer Technology has officially listed on the STAR Market, becoming China's first fully functional GPU company to enter the capital market, marking a significant milestone in its development and the domestic high-end GPU sector [2][4]. Group 1: Company Overview - Moer Technology's stock debuted at a price of 114.28 yuan, the highest for the year, and surged by 468.78% on the first day to reach 650 yuan per share, resulting in a market capitalization exceeding 305.5 billion yuan [2]. - The company aims to raise 8 billion yuan for the development of next-generation AI training and inference chips, graphics chips, and AI SoC chips, with a total fundraising of approximately 7.997 billion yuan after deducting issuance costs [4]. Group 2: Technological Innovation - Moer Technology has established a unified MUSA architecture that supports AI computing acceleration, graphics rendering, physical simulation, and scientific computing on a single chip, setting a key technological benchmark for domestic high-end GPU chip autonomy [6]. - The company has successfully mass-produced five chips and iterated four generations of GPU architectures, forming a comprehensive product matrix that covers AI, scientific computing, and graphics rendering [6]. Group 3: Market Position and Growth - The company has achieved a compound annual growth rate of 208.44% in revenue from 2022 to 2024, with revenue in the first half of 2025 reaching 702 million yuan, surpassing the total of the previous three years [8]. - Moer Technology's AI computing business contributed 94.73% of its revenue, reflecting its focus and leading advantage in core sectors [8]. Group 4: Strategic Alignment - The company's listing aligns with national strategies for technological self-reliance and the development of high-end chips, which are crucial for advancing AI and digital economy initiatives [9]. - The funds raised will enhance Moer Technology's competitive edge in AI computing and graphics processing, supporting the intelligent transformation across various industries [9].

Core Viewpoint - Moore Threads has officially listed on the Shanghai Stock Exchange's Sci-Tech Innovation Board, marking a significant milestone as the first fully functional GPU company in China to enter the capital market, reflecting recognition of its commitment to independent innovation and product iteration in the GPU computing power sector [1][3]. Company Development - Since its establishment in 2020, Moore Threads has focused on independent innovation and has become a leading domestic full-function GPU enterprise, aligning with national strategies for technological self-reliance and building a secure computing power system [3][4]. - The company has successfully mass-produced five chips and iterated four generations of GPU architectures, establishing a diverse product layout that covers AI, scientific computing, and graphics rendering [4][5]. Technological Advancements - Moore Threads has achieved a technological breakthrough with its MUSA unified system architecture, enabling a single chip to support AI computing acceleration, graphics rendering, physical simulation, and scientific computing [4]. - The company has developed the KUAE intelligent computing cluster, capable of supporting large-scale expansion and providing a reliable computing infrastructure for training trillion-parameter models, positioning itself as a key player in the AGI era [6]. Market Performance - From 2022 to 2024, the company's revenue compound annual growth rate reached 208.44%, with 2025's first half revenue reaching 702 million yuan, surpassing the total of the previous three years [7]. - The AI computing business contributed 94.73% of the revenue, highlighting the company's focus and competitive advantage in its core sector [7]. Strategic Positioning - The national strategy emphasizes high-end chips as a key area for development, establishing GPUs as central to driving industrial upgrades and future competitiveness [8]. - The funds raised from the IPO will be directed towards the development of next-generation AI training and inference chips, graphics chips, and AI SoC chips, aiming to enhance the company's technological leadership in AI computing and graphics [8]. Vision and Goals - Moore Threads aims to become a globally competitive GPU leader, providing advanced computing platforms for the integration of AI and digital twin technologies, contributing to the digital economy and national technological self-reliance [10].

Core Viewpoint - Moores Threads has officially listed on the Shanghai Stock Exchange's Sci-Tech Innovation Board, becoming China's first fully functional GPU company to enter the capital market, which signifies a significant milestone in the domestic GPU industry [2][3]. Group 1: Company Overview - Moores Threads, established in 2020, is a leading domestic full-function GPU enterprise that focuses on independent innovation and has aligned its development with national strategies for technological self-reliance and a secure computing power system [4]. - The company has achieved significant technological breakthroughs with its MUSA unified system architecture, enabling a single chip to support AI computing acceleration, graphics rendering, physical simulation, scientific computing, and ultra-high-definition video encoding and decoding [6]. Group 2: Financial Performance - In the first half of this year, Moores Threads reported a revenue of 702 million yuan, surpassing the total revenue from 2022 to 2024, which was approximately 608 million yuan [9]. - The company's gross margin has dramatically improved from -70.08% in 2022 to 69.14% in the first half of 2025, indicating a significant enhancement in product competitiveness and market recognition [9]. Group 3: Research and Development - Moores Threads has invested over 4.3 billion yuan in R&D from 2022 to June 2025, with over 77% of its workforce dedicated to research, leading to substantial innovation outcomes [10]. - The company has obtained 514 authorized patents as of June 2025, positioning it among the top GPU enterprises in China, which strengthens its technological moat based on continuous independent innovation [10]. Group 4: Strategic Positioning - The listing on the Sci-Tech Innovation Board coincides with a historical opportunity, as the raised funds will focus on the development of next-generation AI training and inference chips, graphics chips, and AI SoC chips, aiming to solidify and expand the company's technological leadership in AI computing and graphics [11]. - Moores Threads is committed to advancing the computing industry towards more general and intelligent directions, contributing to the national digital economy and the mission of technological self-reliance [11].

Core Insights - The launch of the IPO for Moer Thread Intelligent Technology marks a significant event in China's chip industry, with an issue price of 114.28 yuan per share, making it the highest-priced new stock of the year and raising 8 billion yuan, setting a record for the Sci-Tech Innovation Board [1] - The company is set to become the first listed firm in the A-share market focusing on "full-function GPUs," indicating a critical revaluation of the domestic high-end GPU sector [1] Company Overview - Founded in 2020, Moer Thread's founding team has a background in NVIDIA, with founder Zhang Jianzhong previously serving as NVIDIA's global vice president [2] - The company has rapidly iterated four generations of GPU architecture within four years, covering high-performance computing fields such as AI intelligence, cloud computing, and personal computing [2] Financial Performance - Revenue surged from 46 million yuan in 2022 to 438 million yuan in 2024, with a compound annual growth rate exceeding 200%, and revenue for the first half of 2025 reached 702 million yuan, nearly 1.6 times that of the previous year [2] - The gross margin has remained stable at around 70% from 2024 to the first half of 2025, indicating strong product competitiveness [2] Market Valuation - The IPO pricing reflects the market's enthusiasm for scarce technological assets, with a total market capitalization of 53.7 billion yuan and a price-to-sales ratio of 122.51 times, alongside a price-to-earnings ratio of 60.12 times, both significantly above industry averages [3] - Such high valuations imply that the market has incorporated extremely optimistic future growth expectations, posing a risk of significant valuation corrections if revenue growth or gross margins do not meet expectations [3] Competitive Landscape - Moer Thread's strategy focuses on full-function GPUs, which integrate multiple computing engines, catering to diverse computational needs, and is supported by the MUSA unified system architecture [3] - The company has transitioned from a hardware supplier to a computing infrastructure platform with its "KUA" intelligent computing cluster, capable of supporting advanced model pre-training [4] Challenges Ahead - Despite the broad application scenarios for full-function GPUs, the company faces intense competition from specialized manufacturers in each segment [5] - The company has not yet achieved profitability and relies on continued financing to support technology development and market expansion [5] - Being listed on the "entity list" may restrict access to advanced process chips and EDA tools, necessitating a more resilient supply chain [5] - The competitive landscape includes domestic rivals such as Biran and Muxi, requiring Moer Thread to maintain technological leadership and market competitiveness [5] Industry Significance - The record IPO price for Moer Thread represents a milestone in China's hard technology sector, reflecting strong market confidence in domestic high-end GPU solutions, particularly in the full-function GPU pathway [6] - The company's post-IPO performance will serve as a barometer for the development of the entire domestic high-end chip industry [6]

Core Viewpoint - The article highlights that Moer Technology is set to become the first A-share listed company with "full-function GPU" as its core asset, marking a significant milestone in the domestic high-end AI chip sector [2][3]. Company Overview - Founded in 2020, Moer Technology focuses on the independent research and design of full-function GPUs, positioning itself as a leading player in the domestic high-end AI chip market [4]. - The company's strategic choice of "full-function GPU" distinguishes it from competitors that focus on AI computing, such as GPGPU or ASIC chips, by integrating multiple functionalities into a single chip [5]. Product and Technology - The full-function GPU integrates four major engines: AI computing acceleration, graphics rendering, physical simulation, and ultra-high-definition video encoding/decoding, catering to diverse computing needs [6]. - The MUSA unified system architecture, developed by Moer Technology, serves as the technical foundation for its full-function GPUs, ensuring compatibility with mainstream ecosystems and allowing developers to migrate existing models without rewriting code [6]. - Moer Technology has launched four generations of GPU architecture chips since its inception, with the latest "Pinghu" architecture featuring 8,192 shading cores and 512 tensor cores, supporting FP8 precision and a memory bandwidth of 800 GB/s [7]. Financial Performance - In the first half of 2023, Moer Technology achieved a revenue of 702 million yuan, nearly 1.6 times its projected full-year revenue for 2024, and surpassing the total revenue from 2022 to 2024 [7]. - The company has shown a trend of narrowing losses from 2023 to 2024, indicating improved commercial conversion efficiency and initial scale effects [8]. - The revenue from AI computing products has rapidly increased, with a total of 1 billion yuan generated from 2022 to 2024, reflecting a strategic focus on the high-growth AI computing sector [8]. Market Position and Ecosystem - Moer Technology's KUAE intelligent computing cluster, based on its self-developed full-function GPU, supports interconnectivity of up to 10,240 GPUs, showcasing its capability to provide large-scale computing infrastructure [9][10]. - The company has initiated the "Moer Academy" developer program and completed compatibility certification for over a hundred domestic software solutions, enhancing its ecosystem and attracting more developers [10]. - The initiation of Moer Technology's IPO signifies a systematic valuation assessment of domestic GPU enterprises in the capital market, amidst increasing global competition for AI computing resources [11].

Core Viewpoint - Moer Technology has officially launched its IPO application on the Sci-Tech Innovation Board, potentially becoming the first A-share company focused on "full-function GPUs" as a core asset in the Chinese market [1] Group 1: Industry Context - The topic of computing power has gained significant attention due to the explosion of large model training, the establishment of large-scale intelligent computing centers, and the rapid development of policies, with GPUs at the center of these transformations [1][3] - High-performance GPU computing power is now considered the "oil" of the digital economy era, highlighting the need for China to develop its own equivalent to NVIDIA [3] Group 2: Company Strategy and Technology - Moer Technology's unique strategic positioning lies in its focus on "full-function GPUs," which integrate AI computing acceleration, graphics rendering, physical simulation, and scientific computing into a single chip, catering to diverse computing needs [3][4] - The MUSA unified system architecture, developed by Moer Technology, serves as the technical foundation for its full-function GPUs, allowing for compatibility with mainstream ecosystems and enabling developers to migrate existing models and applications without rewriting code [4] Group 3: Financial Performance and Growth - Moer Technology has demonstrated a steep revenue growth trajectory, with revenues increasing from 46.09 million yuan in 2022 to 123.98 million yuan in 2023, and projected to reach 438.46 million yuan in 2024 [7] - The company's revenue from AI computing products has been rapidly increasing, indicating a strategic focus on the high-growth potential of AI computing, with total revenue from this segment reaching 1 billion yuan from 2022 to mid-2025 [8] Group 4: Ecosystem Development - The company has initiated the "Moer Academy" developer program and has completed compatibility certification for over a hundred domestic software solutions, enhancing its ecosystem and driving technological iteration [9] - The launch of the IPO marks a significant milestone for Moer Technology and the Chinese GPU industry, representing a historic opportunity for investors to engage with breakthroughs in core computing power [9]

Core Viewpoint - Moer Technology has officially launched its IPO application on the Sci-Tech Innovation Board, marking the first listing of a company focused on "full-function GPU" assets in the A-share market. This event signifies a systematic valuation assessment of domestic "full-function GPU" enterprises amid the growing importance of computing power in AI and digital economy [1][14]. Group 1: Industry Context - The global tech competition is reshaping, with high-performance GPU computing power becoming the "oil" of the digital economy era. The U.S. has imposed export controls on high-end chips to China, indicating Moer Technology's significant technological position [2][14]. - The GPU sector is characterized by high barriers to entry, long development cycles, and a limited number of players, making it a challenging field for domestic companies to compete against established giants like NVIDIA [2][3]. Group 2: Company Strategy and Product Offering - Moer Technology's strategy focuses on developing "full-function GPUs" that integrate AI computing acceleration, graphics rendering, physical simulation, and scientific computing, catering to diverse computing needs [3][4]. - The company has developed the MUSA unified system architecture, which enhances compatibility with existing CUDA ecosystems, reducing migration costs for developers and facilitating faster commercialization [4][5]. Group 3: Financial Performance and Growth - Moer Technology has demonstrated a steep revenue growth trajectory, with revenues increasing from 46.09 million yuan in 2022 to 123.98 million yuan in 2023, and projected to reach 438.46 million yuan in 2024. The first half of 2025 has already seen revenues of 701.76 million yuan, indicating a significant acceleration in commercial delivery [8][9]. - The company's AI computing products have shown a rapid increase in revenue contribution, aligning with industry growth trends and enhancing product line maturity [9][10]. Group 4: Market Position and Future Outlook - As the first domestic GPU company to go public, Moer Technology's valuation will be complex, balancing short-term financial metrics with long-term expectations for self-sufficiency in computing power [14]. - The company is positioned to provide large-scale computing infrastructure, evidenced by its KUAE computing cluster supporting over 10,000 GPUs, showcasing its system-level capabilities beyond chip design [12][14].