Core Insights - The rapid development of artificial intelligence (AI) technology has intensified global strategic competition, with intelligent computing centers becoming a new infrastructure in tech competition [1] - Investment in intelligent computing centers in China has surged from hundreds of billions to trillions, with nearly 150 operational projects and close to 400 under construction or planning as of November 2024 [2][3] Investment Trends - As of August 2024, the total number of intelligent computing center projects in China exceeded 300, with a computing power scale surpassing 500,000 PFlops, and over 50 new projects launched in 2024 alone [3] - The total investment in 128 disclosed projects reached over 430 billion yuan, but only 16 projects are in operation or trial phases, indicating a significant gap between planned and actual deployment [3][9] Structural Issues - The market is experiencing a "structural mismatch" rather than an outright surplus of computing power, with significant demand growth in AI inference driving the need for more efficient resource allocation [5][9] - Some intelligent computing centers report utilization rates as low as 10-15%, with certain centers having GPU utilization below 30%, leading to concerns about resource wastage [9][11] Management and Operational Challenges - The supply chain uncertainties, particularly regarding NVIDIA chips, pose challenges for intelligent computing centers, prompting a shift towards domestic chip manufacturers providing comprehensive solutions [12][14] - The industry is moving towards "refined management," with traditional leasing models becoming less profitable, necessitating mergers and resource reallocation to enhance efficiency [14][15] Future Outlook - The ongoing demand for AI inference and the acceleration of domestic chip development are critical for the industry's future, with a focus on building a sustainable computing power system [14] - The integration of AI technology with industry applications requires a skilled workforce, emphasizing the need for professionals who understand both AI and industry-specific needs [14]

Financial Performance - In 2024, the company's total revenue was CNY 7.757 billion, a decrease of 4.71% year-on-year [3] - Data center revenue reached CNY 3.152 billion, an increase of 14.07% year-on-year, with a gross margin of 28.77% [3] - Smart energy revenue was CNY 922 million, down 9.01% year-on-year, with a gross margin of 41.06% [3] - New energy revenue was CNY 3.608 billion, a decline of 15.98% year-on-year, with a gross margin of 16.78% [3] - The net profit attributable to shareholders was CNY 315 million, impacted by credit and asset impairment losses totaling CNY 258 million [3] Business Development - The smart computing center business generated approximately CNY 3.152 billion in revenue, growing 14.07% year-on-year [4] - The company is actively engaging in AI, computing power, and autonomous driving markets, establishing a robust heterogeneous computing power platform [4] - The company has secured numerous product orders from major clients in the internet, telecommunications, and finance sectors, with significant growth expected in 2025 [5] Technological Advancements - The company has been recognized as a top 10 service provider in the third-party computing power center sector, ranking 8th in comprehensive development and 4th in technical capability [6] - In the smart energy sector, the company has launched innovative products such as the 1500V 5MW liquid-cooled photovoltaic inverter, enhancing its market position [8] - The company has achieved the highest shipment volume of energy storage PCS among Chinese enterprises globally, maintaining a strong presence in the renewable energy sector [9] Future Outlook - The company plans to intensify its focus on the smart computing center business, leveraging its core technology in power electronics for new product development [10] - A strategic customer department has been established to deepen collaborations with major internet companies and expand into various sectors [10] - The company is considering future financing plans based on strategic goals and market conditions, following the successful completion of a convertible bond project [11]

通信 分析师：李璐毅 登记编码：S0730524120001 lily2@ccnew.com 021-50586278 Manus 在 AI Agent 领域实现技术突破， AI 驱动智算中心升级 ——通信行业月报 证券研究报告-行业月报 强于大市(维持) 通信相对沪深 300 指数表现 相关报告 《通 信 行业 月 报： 三 大 运营 商 全面 接入 DeepSeek 开源大模型，数据要素产业化进程 加速》 2025-02-16 《通信行业月报：国家实施手机等数码产品购 新 补 贴 ， 智 能 手 机 市 场 平 稳 复 苏 》 2025-01-08 《通信行业月报：运营商大数据业务高速增 长，AI 手机渗透率逐步提升》 2024-12-11 联系人：李智 投资要点： 本报告版权属于中原证券股份有限公司 www.ccnew.com 请阅读最后一页各项声明 第1页 / 共28页 发布日期：2025 年 03 月 12 日 风险提示：国际贸易争端风险；供应链稳定性风险；云厂商或运营商资 本开支不及预期；数字中国建设不及预期；AI 发展不及预期；行业竞 争加剧。 本报告版权属于中原证券股份有限公司 www.ccn ...

Core Viewpoint - The article discusses the missed opportunity for JD.com to acquire Guanghuan Xinnet, a significant player in the IDC sector, and how this decision has impacted JD Cloud's growth and market position in the AI and cloud computing landscape [4][10][26]. Group 1: Historical Context - Guanghuan Xinnet's stock price has nearly doubled since the beginning of the year, highlighting its growing significance in the market [2]. - JD Cloud launched eight intelligent computing products, including its large model "Yansi," indicating its ongoing efforts to innovate [3]. - Several key executives at JD.com supported the acquisition of Guanghuan Xinnet, but the decision ultimately failed due to opposition from the then Chief Strategy Officer, Liao Jianwen [7][8]. Group 2: Strategic Importance of Guanghuan Xinnet - Guanghuan Xinnet's IDC infrastructure is crucial for cloud computing, and its acquisition would have strengthened JD Cloud's IaaS capabilities [10]. - The potential acquisition was seen as a strategic move to enhance JD's asset base and facilitate tax exchange mechanisms with local governments, which could have provided significant operational advantages [11][12]. Group 3: Impact on JD Cloud's Performance - Following the 2018 incident, JD Cloud's revenue declined sharply, and the company struggled to maintain its market position [20][21]. - JD Cloud's PS ratio is significantly lower than that of Alibaba Cloud, indicating a lack of market confidence in JD Cloud's growth potential [23]. - The article suggests that if JD had acquired Guanghuan Xinnet, JD Cloud might not be facing its current challenges and could have leveraged the growing demand for intelligent computing [26].