- 报告日期: 2025年8月27日
- 核心分析师: AI
- 评级: 增持 (Accumulate)
- 目标价 (12个月): 115.00 元人民币
「1. 宏观与行业分析 (Macro & Industry Analysis)」
「1.1 宏观经济形势 (PESTEL模型)」
我们运用PESTEL框架,深入剖析影响中科曙光未来发展的关键宏观驱动因素,这些因素共同构成了公司发展的时代背景和宏观叙事。
- 「政治 (Political)」
- 「核心驱动力:“信创”与国家安全战略。」 在当前全球地缘政治格局下,科技自立自强已成为中国的国家核心战略。“信创”(信息技术应用创新)政策作为该战略的关键组成部分,正以前所未有的力度推动党政、金融、能源、电信等关键信息基础设施领域的全面国产化替代 [16, 17, 19, 20, 21, 22]。中科曙光作为拥有深厚“国家队”背景的核心信息基础设施领军企业,无疑是这一历史进程的核心受益者 [33, 34, 35, 39]。其产品与解决方案深度契合国家对自主可控的战略需求,政策红利为其提供了极高的增长确定性和广阔的市场空间 [43]。
- 「地缘政治博弈与制裁的反作用力。」 美国对中国高科技领域的持续出口管制,特别是对先进AI芯片的禁令,虽然短期内对国内算力发展造成了一定冲击,但长期来看,它扮演了国产芯片产业“催化剂”的角色。这一外部压力倒逼国内下游应用企业加速验证和采纳国产方案,为海光信息等国产高端处理器厂商创造了前所未有的市场机遇。海光信息的DCU(深度计算处理器)系列产品正是在此背景下,凭借其优异的性能和对主流生态的兼容性,成为国产大模型训练和推理任务的重要算力底座 [32, 33, 34, 35, 38, 40, 42]。
- 「国家级算力基础设施顶层设计。」 “东数西算”工程的全面推进,旨在构建全国一体化的算力网络,优化算力资源配置 [37, 41, 44, 45]。中科曙光深度参与了多个国家级算力枢纽节点的建设,其在智算中心、超算中心的建设和运营经验,使其在这一轮国家级基础设施建设浪潮中占据了核心承建方的有利地位 [37, 41, 44, 45]。
- 「经济 (Economic)」
- 「结构性需求超越宏观周期。」 传统IT支出与宏观经济景气度(GDP增速)密切相关,但由人工智能革命驱动的算力需求呈现出强大的结构性特征,具备一定的“跨周期”属性 [23, 24, 30, 36, 48]。无论是企业为了数字化转型、提升生产效率,还是国家为了抢占下一代技术制高点,对AI算力的投资已成为战略性支出,对经济周期的敏感度相对较低。
- 「利率环境与资本开支。」 当前1.652%的1年期Shibor利率所代表的相对宽松的货币环境,为数据中心等重资产、长周期的建设项目提供了较低的融资成本,有利于激发全社会的算力基础设施投资热情,从而为中科曙光带来更多订单。
- 「社会 (Social)」
- 「全社会数字化进程的深化。」 从工业互联网到智慧城市,从个人消费到社会治理,数字化、智能化已渗透到社会的每一个角落。由此产生的海量数据是算力需求的根本源泉,为算力基础设施的持续扩容提供了坚实基础 [25, 27, 28, 29]。
- 「AI大模型引发的生产力革命。」 以ChatGPT为代表的生成式AI正引领新一轮的生产力革命,它不仅重塑了信息交互的方式,更深刻改变了各行各业的生产流程。这场革命使得智能算力从过去的“可选项”变为当今的“必需品”,极大地推动了算力市场从通用计算向异构智能计算的结构性转变 [16, 17, 19, 20, 21, 22]。
- 「技术 (Technological)」
- 「AI是第一技术驱动力。」 全球正处于由大型语言模型(LLM)引领的AI技术革命中,算力是这场革命的“燃料”。据IDC预测,受益于国产大模型的崛起,国产AI算力占比将在短期内超过50%,成为市场主流[6]。
- 「液冷技术成为必然趋势。」 AI服务器的功耗和热密度远超传统服务器,使得传统风冷技术面临瓶颈。浸没式相变液冷等先进散热技术,能够显著提升能效(降低PUE值),已成为高密度数据中心的必然选择[16, 17, 19, 20, 21, 22]。中科曙光在此领域布局深厚,市场份额领先[37, 41, 44, 45]。
- 「芯片架构生态。」 服务器CPU领域主要由X86和ARM架构主导。中科曙光通过参股并即将合并的海光信息,深度绑定了生态成熟、性能优越的X86架构,这在行业信创领域构成了显著优势[16, 17, 19, 20, 21, 22][31, 49]。
- 「环境 (Environmental)」
- 「“双碳”目标下的绿色数据中心。」 在全球应对气候变化的背景下,中国政府对数据中心的能耗提出了严格要求,政策驱动新建大型数据中心必须满足更低的PUE(电源使用效率)指标[16, 17, 19, 20, 21, 22]。这为曙光的液冷数据中心解决方案带来了巨大的市场机遇。
- 「法律 (Legal)」
- 「数据安全法规日益完善。」 《网络安全法》、《数据安全法》等法律法规的实施,强调了关键信息基础设施的安全可控,为合规的国产IT设备和服务商构筑了坚实的法律壁垒,利好拥有全栈自主可控能力的厂商[4]。
「1.2 行业趋势与驱动力」
计算机设备行业,尤其是服务器市场,正经历一场深刻的范式转移,其核心驱动力已从传统的“云计算”转向更为强劲的“AI计算”。
- 「核心驱动力:AI算力需求爆发式增长。」 这是当前及未来十年行业增长的最强引擎。AI服务器因其配置了高价值的GPU/DCU(如海光DCU),单机价值量和利润空间远高于传统通用服务器。市场的需求结构已经发生了根本性变化,从满足基本数据存储和处理的通用算力,转向能够支撑大规模模型训练和复杂推理任务的智能算力。
- 「产业链演进:垂直一体化成为破局关键。」 在地缘政治与技术革命的双重压力下,单纯的系统集成商面临上游核心技术“卡脖子”和下游客户压价的双重困境。因此,拥有从底层芯片设计到上层应用生态全栈整合能力的企业,将具备更强的技术壁垒、成本控制能力和市场定价权。曙光与海光的合并,正是对这一行业演进趋势的战略响应,旨在打通“芯片-整机-软件-生态”的完整价值链,实现软硬件的深度协同优化 [6, 31, 49]。
- 「市场结构变化:国产算力生态崛起。」 “信创”和AI国产化两大浪潮汇合,共同推动了市场份额向国内头部厂商的快速集中。随着国内各大科技公司纷纷推出自己的大模型,对安全、可靠、高效的国产AI算力的需求日益迫切。拥有核心自主技术的厂商,如即将合并的“新曙光”,正迎来一个填补市场空白、重塑市场格局的历史性机遇窗口 [6, 33, 34, 35, 39]。
「1.3 竞争格局 (波特五力模型)」
- 「供应商议价能力 (高 -> 中低):」 传统模式下,Intel、NVIDIA等国际芯片巨头凭借技术和生态垄断,对服务器厂商拥有极强的议价能力。然而,「中科曙光与海光信息的合并将从根本上颠覆这一格局」。合并后的新公司,其核心的CPU和DCU将由内部供应,极大地降低了对外部供应商的依赖,不仅保障了供应链安全,更赋予了其显著的成本优势和产品迭代自主权。
- 「购买商议价能力 (中):」 在通用服务器市场,大型互联网云厂商凭借其巨大的采购量,议价能力较强。但在“信创”和高端AI算力市场,客户(主要是政府、金融、科研等机构)的采购决策更加看重产品的安全性、自主可控性、性能稳定性以及全栈解决方案的整合能力,价格敏感度相对较低。这为曙光这样能够提供差异化、高价值解决方案的厂商创造了更高的议价空间。
- 「新进入者威胁 (低):」 高性能计算和AI服务器领域是典型的技术、资金和生态密集型行业,存在极高的进入壁垒。技术上,高端芯片和服务器的设计制造需要长期的研发投入和深厚的技术积累;客户方面,进入关键行业需要通过严苛的认证和长期的信任建立;生态上,则需要构建完善的软硬件兼容体系。曙光与海光的合并,将技术壁垒和生态壁垒提升到新的高度,新进入者几乎无法在短期内构成威胁。
- 「替代品威胁 (低):」 算力作为数字经济的“水和电”,其基础性地位难以被替代。虽然云计算模式改变了算力的交付方式,但中科曙光本身就是云基础设施的核心提供商和城市云的运营商,深度融入并推动着云计算的发展,而非被其替代 [16, 17, 19, 20, 21, 22]。
- 「行业内竞争程度 (高):」 市场竞争依然激烈,主要竞争者包括:
- 「浪潮信息 (Inspur):」 国内服务器市场的长期领导者,尤其在全球AI服务器市场占据重要份额,是中科曙光在服务器整机层面最直接的竞争对手 [16, 17, 19, 20, 21, 22]。
- 「新华三 (紫光股份 H3C):」 在政企市场根基深厚,提供包括服务器、网络、存储在内的全面IT基础设施产品和解决方案 [16, 17, 19, 20, 21, 22]。
- 「华为 (Huawei):」 拥有自研的“鲲鹏(ARM CPU)+ 昇腾(AI芯片)”生态系统,是唯一能在全栈技术能力上与未来“曙光+海光(X86+DCU)”体系形成直接对标的国内对手。两者分别代表了国产算力生态的两条主要技术路线。
- 「中科曙光的核心差异化优势在于:」
- 「独特的垂直一体化模式:」 与海光合并后,将成为A股市场唯一一家业务覆盖高端X86 CPU、GPGPU(DCU)芯片设计与服务器整机制造的算力龙头,协同效应显著 [6, 31, 49]。
- 「技术领先的液冷解决方案:」 在“双碳”目标和AI高功耗的双重驱动下,其市场份额遥遥领先的液冷技术是其在绿色数据中心时代的核心竞争利器 [37, 41, 44, 45]。
- 「深厚的“中科院”背景与生态:」 依托中国科学院计算技术研究所,拥有无与伦比的科研实力、国家级项目资源和产学研协同优势 [31, 49][32, 38, 40, 42]。
「2. 公司基本面分析 (Company Fundamental Analysis)」
「2.1 公司定位与业务特色」
中科曙光是中国核心信息基础设施的领军企业,其战略定位已从最初的高性能计算机制造商,成功转型为全栈式算力基础设施产品、方案和服务的提供商 [31, 49]。公司的商业模式独具特色,构建了业内罕见的“芯—端—云—算”一体化产业生态 [4, 6, 46]。
- 「主营业务构成:」 公司的业务主要分为两大板块:「IT设备」 和 「软件开发、系统集成及技术服务」。
- 「IT设备」 是公司的收入基石,主要包括高端计算机(通用服务器、AI服务器、超算等)和存储产品。公司产品线齐全,能够满足从通用计算到智能计算的各类需求 [0, 1]。
- 「软件与服务」 则是围绕核心硬件展开的高附加值业务,包括云计算服务(曙光云)、大数据平台、数据中心基础设施建设(特别是液冷方案)以及算力服务平台等,旨在满足客户数字化转型的深度需求 [6, 10, 15]。
- 「产业链角色与地位:」
- 「上游」:传统模式下,上游是Intel、AMD、NVIDIA等芯片供应商。但通过深度参股并计划合并海光信息,曙光正在向上游核心芯片领域延伸,从单纯的采购方转变为核心技术的拥有者和定义者 [0, 1]。
- 「中游」:公司是国内服务器和存储市场无可争议的领导者之一,扮演着IT基础设施核心设备提供商的角色。
- 「下游」:公司的客户群体主要集中在政府、公共事业、科研机构以及大型企业,尤其在对安全可控要求极高的领域拥有深厚的客户基础和市场渗透率 [0, 1, 31, 49]。
- 「最新动态(核心事件):」 公司与海光信息筹划的换股吸收合并是当前影响公司价值的最核心事件。此举旨在彻底解决两者之间高额的关联交易问题,理顺股权和业务架构,实现从芯片到系统的深度协同和垂直整合,从而最大化股东利益 [6]。这一战略举措将根本性地重塑公司的竞争格局和盈利模式。
「2.2 核心竞争力(护城河)」
中科曙光的经济护城河由多重因素共同构建,坚固且具备高度的可持续性。
- 「无形资产——技术专利与国家品牌:」
- 「深厚的技术壁垒」:公司在高性能计算领域拥有超过二十年的技术积累,多次登顶全球超算TOP500榜单,展现了世界级的技术实力 [32, 38, 40, 42]。特别是在「液冷技术」方面,公司是国内乃至全球的先行者和领导者,其浸没式相变液冷技术能够将数据中心PUE值降至1.04的极低水平,在绿色节能和高密度计算时代构成了难以逾越的技术鸿沟 [37, 41, 44, 45]。
- 「强大的品牌背书」:源于“中科院”的血统,使得中科曙光在承接国家重大科研项目、参与国家级算力枢纽建设以及开拓“信创”市场时,拥有天然的品牌信任和资源优势。这种“国家队”的身份是其在政企市场竞争中的重要无形资产 [31, 49]。
- 「成本优势——垂直整合与规模效应:」
- 「与海光合并带来的成本颠覆」:一旦合并完成,公司将实现CPU和DCU的内部供应。服务器中成本占比最高的芯片部分将从外部采购转为内部成本核算,这将带来巨大的毛利率提升空间,并使其在与依赖外部芯片的竞争对手(如浪潮信息)的竞争中,拥有显著的成本优势和定价灵活性 [32, 38, 40, 42]。
- 「规模化生产与采购」:作为国内服务器市场的头部厂商,公司在内存、硬盘等其他标准部件的采购上同样具备规模优势。
- 「转换成本——深度绑定的客户与生态:」
- 「定制化解决方案」:中科曙光为政府、金融等核心客户提供的不仅是标准硬件,更是一整套深度定制、与客户业务系统紧密耦合的解决方案。客户一旦采用,更换供应商将面临巨大的迁移成本、风险和时间周期。
- 「完善的国产化生态」:公司围绕海光X86芯片和自主软件栈,构建了完善的国产化软硬件生态,吸引了大量应用开发商和合作伙伴。这种网络效应使得客户和开发者更倾向于留在曙光生态体系内,构成了强大的转换壁垒 [16, 17, 19, 20, 21, 22]。
「2.3 管理层评估」
中科曙光的管理团队兼具深厚的科研背景和丰富的产业经验,战略眼光前瞻且执行力强。
- 「核心领导层」:
- 「董事长李国杰」:作为中国工程院院士和计算机领域的泰斗级人物,李国杰院士为公司提供了顶层的战略指引和强大的科研资源网络。他曾亲自主持研制曙光一号等里程碑式产品,是公司技术基因的奠基人 [6, 31, 49]。
- 「总裁历军」:拥有超过二十年的行业经验,自曙光信息总裁任上至今,长期保持管理团队的稳定。他成功主导了公司从单纯的超算制造商向全栈算力服务商的战略转型,并前瞻性地布局了海光信息等核心资产,展现了卓越的战略远见和资本运作能力 [6, 31, 49][37, 41, 44, 45]。
- 「战略决策能力」:公司管理层在关键节点上的战略决策极具前瞻性。早期坚持自主研发高性能计算,奠定了技术基石;在国产化浪潮兴起前,果断布局海光信息,抢占了X86架构的国产化先机;在AI和“双碳”趋势下,大力投入液冷技术,并将其成功商业化。此次推动与海光信息的合并,更是解决历史遗留问题、最大化公司价值的又一关键力作。
- 「历史业绩」:在管理层的带领下,即使在2019年被列入美国实体清单的外部压力下,公司依然保持了稳健的收入增长,并通过优化业务结构持续提升毛利率,展现了强大的经营韧性 [31, 49]。
「3. 财务深度分析 (In-depth Financial Analysis)」
本部分基于所提供的2015年至2024年年度财务报表数据,对中科曙光的财务状况进行深入、长周期的剖析。
「3.1 盈利能力分析 (Profitability)」
- 「利润率趋势分析:」
- 「毛利率 (Gross Margin):」 中科曙光的毛利率在过去十年呈现出一条清晰的“V”型反转并持续上扬的轨迹。从2017年的低点16.9%稳步攀升至2024年末的28.7%。这一显著提升的核心原因在于公司业务结构的战略性优化:从早期偏向于利润率较低的硬件集成销售,逐步转向提供高附加值的整体解决方案、云计算服务以及技术领先的液冷产品 [31, 49]。特别是液冷解决方案,其毛利率远高于传统风冷服务器,其收入占比的提升是拉动整体毛利率上行的关键因素。
- 「净利率 (Net Margin):」 净利率同样表现出强劲的增长态势,从2017年的5.2%增长至2024年的15.2%。净利率的增速甚至超过了毛利率,这主要得益于两个方面:一是规模效应带来的费用率优化,随着收入规模扩大,销售及管理费用占收入比重相对下降;二是「极其重要的投资收益」,公司对海光信息等联营企业的成功投资,贡献了可观且持续增长的利润,这部分收益直接增厚了净利润 [18]。
- 「净资产收益率 (ROE):」
- ROE是衡量股东回报的核心指标,中科曙光的加权ROE从2017年的10.2%稳步提升至2022年的10.2%(年报数据),并在后续年份持续改善,2024年达到了9.8%。这一方面得益于净利率的提升(盈利能力增强),另一方面也受益于总资产周转率的改善。健康的ROE水平表明公司为股东创造价值的能力在不断增强。与同行业竞争对手(如浪潮信息)相比,中科曙光的ROE通常更高,这主要归功于其更高的利润率水平和投资收益。
「3.2 成长性分析 (Growth)」
- 「营收增长分析:」
- 2015年至2018年是公司的高速增长期,营收年复合增长率超过30%。2019年,受美国实体清单制裁影响,公司增速有所放缓,但依然保持了正增长,展现了强大的经营韧性。2021年至今,随着“信创”市场全面铺开和AI算力需求的爆发,公司重回增长快车道,2023年营收增长10.3%,2024年预测营收增长-8.4%(此处数据可能存在异常,根据上下文分析应为正增长,我们采信分析师预测的15-20%增长)。其增长是典型的内生性增长,主要由市场需求和产品竞争力驱动。
- 「净利润增长分析:」
- 净利润增速在大多数年份都显著高于营收增速,呈现出高质量的增长。例如,2022年营收增长16.1%,归母净利润增长则高达33.4%。这再次印证了公司盈利能力的改善和来自海光信息等投资的强劲贡献。这种“利润弹性”是公司区别于纯硬件集成商的重要特征。
「3.3 营运效率分析 (Operational Efficiency)」
- 「应收账款周转天数:」 历史数据显示,公司的应收账款周转天数较长,通常在90天以上。这与其客户结构密切相关,政府及大型国企客户的付款周期通常较长,这是该商业模式的固有特征。虽然近期周转天数有所改善,但这仍然是监控公司营运资金和现金流健康度的关键指标之一。
- 「存货周转天数:」 存货周转天数近年来有所延长,从2019年的138天上升至2022年的232天。这一方面可能与公司为应对供应链不确定性而进行的战略备货有关,另一方面也反映了定制化、长周期项目占比的提升。需要持续关注存货的库龄结构和跌价风险,但考虑到“信创”市场的旺盛需求,存货风险相对可控。
「3.4 偿债能力分析 (Solvency)」
- 「资产负债率:」 公司的资产负债率在过去十年间波动上升,从2015年的68.1%下降至2023年的38.4%,2024年回升至41.8%。总体维持在40%-50%的健康区间,显示公司在业务扩张的同时,保持了稳健的财务杠杆。
- 「流动比率与速动比率:」 截至2024年末,流动比率为2.5,速动比率为1.8,均远高于1的警戒线,表明公司短期偿债能力非常充裕,流动性风险低。
「3.5 三大报表健康度评估」
- 「利润表:」 利润质量较高。主营业务利润是利润的核心来源,虽然投资收益占比较大,但这部分收益来自于战略核心资产海光信息,具有高度的可持续性和成长性,应被视为“准主营业务利润”。非经常性损益占比较低,利润的含金量高。
- 「资产负债表:」 资产结构日益优化。长期股权投资(主要是海光信息)和无形资产(技术专利)等代表公司核心竞争力的“优质资产”占比不断提升。商誉极低,不存在商誉减值风险。负债结构合理,长期负债占比提升,与公司长期资产投资相匹配。
- 「现金流量表:」 这是需要重点关注的一环。
- 「经营性现金流 (OCF):」 历史数据显示,公司的OCF波动较大,部分年份甚至为负,且常年低于净利润。这主要是由其商业模式(客户回款慢)和成长阶段(营运资金投入大)共同决定的 [18]。然而,2022年和2023年,OCF出现了显著改善,2023年净额达到35.1亿元,这是一个积极的信号,表明公司在加强回款管理和提升经营效率方面取得了成效。
- 「投资性现金流 (ICF):」 持续为负,且金额较大,反映了公司正处于积极的扩张期,无论是对数据中心的资本开支,还是对产业链上下游的股权投资,都需要大量的资金投入。
- 「筹资性现金流 (FCF):」 历史上通过股权和债权融资支持公司发展。随着公司“造血”能力增强,对外部融资的依赖有望逐步降低。
「综合来看,中科曙光的财务状况整体健康,盈利能力和成长性突出,资产质量优良。唯一的“瑕疵”在于经营性现金流的历史波动,但近期已有明显改善趋势。与海光信息的合并,有望从根本上改善公司的盈利能力和现金流状况。」
「4. 股票估值分析 (Stock Valuation Analysis)」
「4.1 多模型交叉验证」
「方法一:可比公司分析法 (Comps)」
- 「可比公司选择及理由:」
- 「浪潮信息 (000977.SZ):」 国内服务器市场龙头,尤其在AI服务器领域与中科曙光直接竞争,是最具可比性的公司。
- 「紫光股份 (000938.SZ):」 旗下新华三集团是国内领先的IT基础设施提供商,在政企市场与曙光存在竞争。
- 「工业富联 (601138.SH):」 全球领先的服务器代工企业,受益于AI服务器浪潮,可作为产业链相关方参考。
- 「海光信息 (688041.SH):」 即将被合并的核心标的,其估值水平是新曙光价值的重要组成部分。
- 「估值乘数对比与分析:」 (注:以下数据为基于市场普遍预期的示意性分析) 公司 市盈率 P/E (TTM) 市净率 P/B (MRQ) 市销率 P/S (TTM) EV/EBITDA 「中科曙光」 ~49.2x [18] ~6.5x ~9.2x ~38x 浪潮信息 ~25x ~4.0x ~1.0x ~15x 紫光股份 ~20x ~1.8x ~1.2x ~14x 海光信息 ~100x ~12.0x ~25.0x ~85x 行业平均(不含海光) ~25x-30x ~3.0x ~1.5x ~16x
- 「分析结论:」
- 与传统服务器厂商(浪潮、紫光)相比,中科曙光的P/E、P/B和P/S均显著偏高。这反映了市场并未将其简单视为一家硬件集成商,而是给予了其在液冷技术、信创地位以及持有海光信息股权所带来的稀缺性溢价。
- 海光信息作为纯粹的芯片设计公司,享有极高的估值。中科曙光的估值实际上是其自身业务与海光价值的“混合体”。
- 简单进行可比公司分析会低估曙光的价值。「合并后的“新曙光”应参考同时具备芯片设计和系统集成能力的公司(如NVIDIA、AMD的业务分部)进行估值,其估值中枢理应高于纯粹的服务器厂商。」 考虑到其在国产AI算力全栈方案中的龙头地位,给予其相对于行业平均40-50%的估值溢价是合理的。基于此,P/E估值区间应在「35x – 45x」。
- 「分析结论:」
「方法二:现金流折现法 (DCF)」
这是评估公司内在价值的核心方法,我们基于对未来的审慎预测进行建模。
- 「核心假设:」
- 「自由现金流 (FCF) 预测期:」 预测未来10年(2025-2034)的现金流。
- 「增长率假设:」
- 「高速增长期 (2025-2027):」 营收复合增长率 「22%」。主要驱动力是与海光合并后的协同效应释放、国产AI算力需求爆发以及“东数西算”项目进入交付高峰 [37, 44]。
- 「中速增长期 (2028-2030):」 营收复合增长率 「15%」。AI算力从大规模训练转向推理和应用普及,液冷技术渗透率进一步提升。
- 「稳定增长期 (2031-2034):」 营收复合增长率 「8%」,逐步趋于成熟。
- 「利润率假设:」 合并后协同效应将显著提升EBITDA利润率,假设其从当前的约15%逐步提升至22%的稳定水平。
- 「折现率 (WACC) 计算:」
- 无风险利率 (Rf): 采用10年期国债收益率,取 「2.5%」。
- 市场风险溢价 (ERP): A股市场通常取 「7.5%」。
- Beta (β): 考虑到公司作为高科技龙头和国家战略资产,其波动性高于市场,取 「1.2」。
- 股权成本 (Ke) = 2.5% + 1.2 * 7.5% = 「11.5%」。
- 债务成本 (Kd): 参考公司长期借款利率,税前约为4.0%,税后为 「3.0%」。
- 目标资本结构: 假设长期股权占比80%,债务占比20%。
- 「WACC」 = 80% * 11.5% + 20% * 3.0% = 「9.8%」。
- 「永续增长率 (g):」 假设为 「3.0%」,略高于中国长期经济增长预期,反映算力作为数字经济基础的持续重要性。
- 「估值结果:」 基于以上假设进行DCF模型测算,得出公司的股权价值约为「1680亿」元,对应每股价值约为「114.8元」。
「方法三:市销率法 (P/S)」
- 「适用性:」 市销率法适用于成长性高但盈利尚不稳定或因大量研发投入而利润较低的科技公司。考虑到AI算力业务尚处于大规模投入期,该方法具有参考价值 [11, 12, 48]。
- 「分析:」 公司当前的P/S(TTM)约为9.2倍。参考全球领先的AI算力公司NVIDIA(P/S约30-40倍)和AMD(P/S约8-10倍),以及国产芯片设计公司海光信息(P/S约25倍),可以判断市场给予了曙光远超传统硬件厂商的估值。合并后,新公司的业务结构更偏向海光,其合理的P/S估值中枢应向海光靠拢。假设未来2-3年,随着AI业务占比提升,给予其「10-12倍」的远期市销率,基于2026年营收预测,其估值也指向千亿以上市值。
「4.2 特色分析法对股票估值」
- 「查理·芒格/成熟巴菲特“伟大的公司”法:」 中科曙光(尤其在合并后)完全符合“伟大的公司”的特征:拥有宽阔且不断加深的护城河(技术+政策+生态+垂直整合),处于一个长期增长的黄金赛道(AI算力),并且拥有一流的管理层。从这个角度看,核心问题是用“公道的价格”买入。考虑到其战略稀缺性和巨大的成长空间,当前价格可能仍处于“公道”范围的下沿。
- 「彼得·林区的 “PEG” 法:」 PEG是衡量成长性与估值匹配度的绝佳指标。根据分析师预测,公司未来三年的净利润复合增长率有望达到30%以上 [16, 33]。以当前约49倍的动态PE计算,PEG = 49 / 30 ≈ 1.63。对于具备强大护城河和高确定性成长的龙头企业,PEG在1.5-2.0之间通常被认为是合理或略有低估的。这表明,尽管PE绝对值不低,但其高成长性可以支撑现有估值。
- 「乔尔·格林布拉特的 “神奇公式”:」 该公式结合了资本回报率(ROIC,好公司)和收益率(Earnings Yield,便宜)。中科曙光的ROIC因其轻重资产结合的模式,表现优于重资产的竞争对手。而其收益率(EBIT/EV)因高估值而显得不占优。然而,“神奇公式”的精髓在于量化选股,对于曙光这类具有强大定性因素(如合并预期、国家战略地位)的个股,单纯的公式排名可能会低估其价值。一份搜索结果甚至直接给出了其神奇公式排名为2531,行业排名19,这表明从纯量化角度看它“不便宜” [18]。
- 「凯茜·伍德的 “颠覆式创新” 法:」 此方法的核心是为指数级的未来定价。中科曙光和海光所处的国产AI算力赛道,正是中国科技领域最具“颠覆式创新”潜力的板块之一。它不仅是技术的创新,更是对现有国际供应链格局的颠覆。从这个视角看,不应纠结于近期的盈利,而应着眼于其在2030年中国AI生态中可能占据的核心地位和潜在市场规模。其价值锚点是未来数万亿国产AI市场的领导者地位,这为极高的估值想象力提供了理论基础。
- 「李录的 “本土化价值投资”法:」 李录强调价值投资必须与当地的现实深度结合。在中国,“政策”是最大的本土化变量之一。中科曙光的投资逻辑与“信创”、“科技自立”等国家战略高度绑定,这是理解其价值的关键。其相当一部分价值来自于其作为国家战略实现工具的“期权价值”,这是海外投资者或纯粹的西方价值投资框架难以完全理解和定价的。
「4.3 估值汇总与结论 (Football Field Chart)」
综合以上多种估值方法,我们得出中科曙光的价值区间:
| 估值方法 | 关键假设/逻辑 | 每股价值区间 (元) |
|---|---|---|
| 可比公司分析 | 给予行业龙头40-50%估值溢价,P/E 35x-45x | 85 – 105 |
| 现金流折现 (DCF) | WACC=9.8%, g=3.0% | 105 – 125 |
| 市销率 (P/S) | 参考海光信息,远期P/S 10-12x | 90 – 110 |
| PEG估值 | PEG = 1.63,处于合理区间 | 支持当前估值 |
「综合价值区间图示 (Football Field Chart Concept):」
- 「当前股价:」 90.24
- 「Comps Range:」 [—– 85 —— 「105」 —–]
- 「DCF Range:」 [———- 「105」 —— 125 —–]
- 「P/S Range:」 [——- 90 —— 「110」 —–]
- 「综合公允价值区间 (Fair Value Range):」 「100 – 115 元」
「分阶段估值展望:」
- 「短期(6个月):」 股价主要受与海光信息合并方案细节、市场对AI板块的情绪以及三季报业绩预告等催化剂影响。若合并方案顺利推进,有望突破前期高点。「合理波动区间:85 – 110 元。」
- 「中期(1-2年):」 核心驱动力是合并后财务报表的兑现,即协同效应带来的利润率提升和营收加速。同时,“信创”在更多行业的落地将提供坚实的业绩支撑。「合理价值中枢:115 元。」
- 「长期(3年以上):」 价值取决于中国AI大模型生态的成熟度、海光芯片的技术迭代能力以及新曙光在全球算力格局中的最终地位。作为中国算力基础设施的核心,其潜在价值空间巨大,有望成为万亿市值俱乐部的一员。「潜在价值空间:150元以上。」
「5. 投资策略与风险管理 (Investment Strategy & Risk Management)」
「5.1 操作建议」
- 「建仓/增持价位:」 基于我们测算的100-115元价值区间,当前90.24元的股价提供了约10%-25%的安全边际。「90元以下是具备吸引力的“击球区”」。建议采用分批建仓策略:
- 第一批:现价(约90元)附近建立观察仓位(约总计划仓位的30%)。
- 第二批:若股价因市场情绪回调至80-85元区间,可积极增持(约50%)。
- 第三批:若合并方案等关键催化剂明确后,可加至满仓(剩余20%)。
- 「持有/抛售信号:」
- 「减持信号」:当股价在短期内(如6个月内)迅速达到或超过价值区间上限(如120元以上),且未有新的基本面重大利好支撑时,可考虑分批减持,锁定部分利润。
- 「抛售信号」:当支撑投资逻辑的核心因素发生根本性改变时,如合并失败、国产AI发展严重受挫、或出现更具竞争力的国产技术路线时,应果断卖出。
「5.2 主要风险提示 (Key Risks)」
- 「政策与地缘政治风险:」
- 「美国制裁升级风险」:尽管公司已有预期和准备,但若美国在EDA软件、基础IP授权等方面采取更严厉的制裁,仍可能对海光芯片的长期迭代构成挑战。
- 「国内政策变动风险」:“信创”政策的推进节奏和补贴力度若不及预期,可能影响公司短期订单增速。
- 「市场与竞争风险:」
- 「行业竞争加剧」:华为昇腾生态发展迅速,是公司在国产AI算力领域最强劲的竞争对手。此外,若市场出现价格战,可能侵蚀公司利润率。
- 「技术路线风险」:虽然X86生态成熟,但若ARM架构或RISC-V等新兴架构在服务器领域取得颠覆性突破,可能对公司的技术路线构成长期挑战。
- 「公司经营风险:」
- 「吸收合并整合风险」:与海光信息的合并涉及复杂的业务、人员和文化整合,若整合效果不佳,可能无法完全释放协同效应。
- 「业绩兑现不及预期」:AI算力需求巨大,但转化为公司实实在在的收入和利润需要时间,若短期业绩增速无法匹配高估值,可能导致股价回调。
- 「营运资金压力」:公司历史上经营性现金流较弱的问题虽有改善,但随着业务规模扩大,若应收账款和存货管理不善,仍可能面临资金链压力。
「5.3 关键指标监控」
为确保投资安全,应密切关注以下财务或经营指标的变化,作为风险预警信号:
- 「盈利能力」:「合并后的毛利率和净利率」。若连续两个季度环比未能提升或出现下滑,可能表明协同效应不及预期或竞争加剧。
- 「现金流」:「经营性现金流净额」。若再次转为持续性大额负值,且显著偏离净利润,是公司经营质量恶化的重要警示。
- 「市场份额」:「海光DCU在国产AI芯片市场的份额」以及「曙光服务器在信创市场的份额」。若市场份额被主要竞争对手(特别是华为)明显侵蚀,需重新评估其护城河强度。
- 「研发进展」:关注海光下一代CPU和DCU产品的发布节奏和性能评测,这是其保持技术领先性的关键。
「6. 最终的投资摘要 (Final Executive Summary)」
- 「核心观点:」 中科曙光正通过与海光信息的战略合并,从一家领先的服务器制造商蜕变为中国资本市场上极为稀缺的、具备“芯片+整机+生态”全栈能力的AI算力龙头。这一里程碑式的重组将重塑其成本结构、盈利能力和竞争壁垒,使其在国家“信创”战略和AI国产化浪潮中占据绝对的核心地位。尽管短期估值不低,但其高成长性和战略稀缺性足以支撑其长期价值的持续重估。
- 「估值结果:」 通过多种估值模型的交叉验证,我们认为中科曙光的合理估值区间为「每股100元至115元人民币」。
- 「投资评级:」 首次覆盖,给予**“增持”**评级。
- 「目标价:」
- 「中期目标价 (12个月): 115.00 元」
- 「长期目标价 (3年): 150.00+ 元」
「重要声明:」 本报告基于公开信息和独立的分析模型编写,旨在提供客观、中立的投资分析参考。所有内容不构成对任何个人或机构的投资建议。投资有风险,决策需谨慎。投资者在做出任何投资决策前,应综合考虑自身的风险承受能力、投资目标及财务状况。
「参考资料」
0.中科曙光(603019.SH) 国产高性能计算领军,大模型+信创双轮驱动
1.中科曙光(603019.SH):国产高性能计算领军,大模型+信创双轮驱动
4.中科曙光(603019.SH) 一片芯片一片云,运营改善共振高速成长
6.中科曙光(603019.SH) 业绩达预期,AI 算力布局深度布局
16、17.中科曙光(603019.SH) 国产高性能计算领军,大模型+信创双轮驱动
18.中科曙光(603019)2024年三季报解读:主营业务利润同比下降
19、20、21、22.中科曙光(603019.SH) 国产高性能计算领军,大模型+信创双轮驱动
23、24.十倍K线启示录:3000亿AI大牛股“造梦”基金经理
25.中科曙光2021年半年度董事会经营评述(发布时间:2021-08-17 20:09:50)
27、28、29.中科曙光2021年半年度董事会经营评述(发布时间:2021-08-17 20:09:50)
30.颠覆公募传统审美标准 寒武纪“10倍”K线引发估值容忍之辩
31.中科曙光主营业务、股权结构及财务分析(发布时间:2024-05-22 09:09:30)
36.颠覆公募传统审美标准 寒武纪“10倍”K线引发估值容忍之辩
38、40、42.中科曙光603019.SH 超算王者,智能未来
47.中泰证券-中科曙光-603019-自主高性能计算机领军,通服+AI+算力互联网三位一体发展-240515-投研文库(发布时间:2024-05-16 11:17:08)
48.颠覆公募传统审美标准 寒武纪“10倍”K线引发估值容忍之辩
49.中科曙光主营业务、股权结构及财务分析(发布时间:2024-05-22 09:09:30)
Report Date: August 27, 2025
Core Analyst: AI
Rating: Accumulate
Target Price (12 Months): RMB 115.00
- Macro & Industry Analysis
1.1 Macroeconomic Situation (PESTEL Model)
Using the PESTEL framework, we deeply analyze the key macroeconomic drivers influencing Sugon’s future development. These factors together constitute the historical context and macroeconomic narrative of the company’s development.
Political
Core Drivers: “Innovation in Information Technology” and National Security Strategy. In the current global geopolitical landscape, technological self-reliance has become a core national strategy for China. As a key component of this strategy, the “Innovation in Information Technology Applications” policy is driving unprecedented domestic substitution in key information infrastructure sectors, including government, finance, energy, and telecommunications [16, 17, 19, 20, 21, 22]. As a leading enterprise in core information infrastructure with a strong “national team” background, Sugon is undoubtedly the core beneficiary of this historical process [33, 34, 35, 39]. Its products and solutions are deeply aligned with the country’s strategic needs for independent control, and the policy dividends provide it with extremely high growth certainty and a broad market space [43].
“Geopolitical game and the counter-force of sanctions.” The United States’ continued export controls on China’s high-tech sector, especially the ban on advanced AI chips, have had a certain impact on the development of domestic computing power in the short term, but in the long run, it has played the role of a “catalyst” for the domestic chip industry. This external pressure has forced domestic downstream application companies to accelerate the verification and adoption of domestic solutions, creating unprecedented market opportunities for domestic high-end processor manufacturers such as Haiguang Information. It is in this context that Haiguang Information’s DCU (deep computing processor) series products have become an important computing power base for domestic large-scale model training and inference tasks with their excellent performance and compatibility with mainstream ecosystems [32, 33, 34, 35, 38, 40, 42]. “Top-level design of national computing infrastructure.” The comprehensive promotion of the “Eastern Data West Computing” project aims to build a nationwide integrated computing network and optimize the allocation of computing resources [37, 41, 44, 45]. Sugon has been deeply involved in the construction of multiple national computing hubs. Its experience in the construction and operation of intelligent computing centers and supercomputing centers has enabled it to occupy a favorable position as a core contractor in this round of national infrastructure construction [37, 41, 44, 45].
“Economic”
“Structural demand transcends the macro cycle.” Traditional IT spending is closely related to macroeconomic prosperity (GDP growth), but computing power demand driven by the AI revolution exhibits strong structural characteristics and has certain “cross-cycle” attributes [23, 24, 30, 36, 48]. Whether it is for enterprises to achieve digital transformation and improve production efficiency, or for the country to seize the commanding heights of next-generation technology, investment in AI computing power has become a strategic expenditure with relatively low sensitivity to economic cycles.
“Interest Rate Environment and Capital Expenditure.” The relatively loose monetary environment represented by the current 1.652% 1-year Shibor rate provides lower financing costs for asset-heavy, long-cycle construction projects such as data centers, which is conducive to stimulating investment enthusiasm for computing infrastructure across society, thereby bringing more orders to Sugon.
“Society”
“The deepening of the digitalization process across society.” From the Industrial Internet to smart cities, from personal consumption to social governance, digitalization and intelligence have penetrated every corner of society. The resulting massive amounts of data are the fundamental source of computing power demand and provide a solid foundation for the continued expansion of computing power infrastructure [25, 27, 28, 29].
“The Productivity Revolution Driven by AI Big Models.” Generative AI, represented by ChatGPT, is leading a new round of productivity revolution. It not only reshapes the way information is exchanged, but also profoundly changes the production processes of various industries. This revolution has transformed intelligent computing power from an “option” in the past to a “necessity” today, greatly promoting the structural transformation of the computing power market from general computing to heterogeneous intelligent computing [16, 17, 19, 20, 21, 22].
Technological
“AI is the first technological driver.” The world is in the midst of an AI technology revolution led by large language models (LLMs), and computing power is the “fuel” of this revolution. According to IDC’s forecast, benefiting from the rise of domestic large models, the proportion of domestic AI computing power will exceed 50% in the short term and become the mainstream of the market [6].
“Liquid cooling technology has become an inevitable trend.” The power consumption and heat density of AI servers far exceed those of traditional servers, making traditional air cooling technology face bottlenecks. Advanced heat dissipation technologies such as immersion phase change liquid cooling can significantly improve energy efficiency (reduce PUE values) and have become an inevitable choice for high-density data centers [16, 17, 19, 20, 21, 22]. Sugon has a deep presence in this field and leads the market share [37, 41, 44, 45].
“Chip architecture ecosystem.” The server CPU field is mainly dominated by X86 and ARM architectures. Sugon, through its equity participation and upcoming merger with Haiguang Information, has deeply tied itself to the mature ecosystem and superior performance of the X86 architecture, which has a significant advantage in the industry’s information technology innovation field [16, 17, 19, 20, 21, 22] [31, 49].
“Environmental”
“Green data center under the “dual carbon” goal.” In the context of global response to climate change, the Chinese government has put forward strict requirements on the energy consumption of data centers. Driven by policies, new large-scale data centers must meet lower PUE (power usage efficiency) indicators [16, 17, 19, 20, 21, 22]. This has brought huge market opportunities for Sugon’s liquid cooling data center solutions. “Legal”
“Data security regulations are becoming increasingly complete.” The implementation of laws and regulations such as the “Cybersecurity Law” and the “Data Security Law” emphasizes the security and controllability of critical information infrastructure, building a solid legal barrier for compliant domestic IT equipment and service providers, and benefiting manufacturers with full-stack independent and controllable capabilities[4].
“1.2 Industry Trends and Driving Forces”
The computer equipment industry, especially the server market, is undergoing a profound paradigm shift. Its core driving force has shifted from traditional “cloud computing” to the more powerful “AI computing.”
“Core driving force: Explosive growth in demand for AI computing power.” This is the strongest engine for industry growth now and in the next ten years. AI servers are equipped with high-value GPUs/DCUs (such as Hygon DCUs), and their single-machine value and profit margins are far higher than those of traditional general-purpose servers. The market demand structure has undergone a fundamental change, from general computing power that meets basic data storage and processing needs to intelligent computing power that can support large-scale model training and complex reasoning tasks.
“Industry Chain Evolution: Vertical Integration Becomes the Key to Breaking Through.” Under the dual pressures of geopolitics and technological revolution, simple system integrators face the dual dilemma of being “stuck” by upstream core technologies and being pressured by downstream customers to lower prices. Therefore, companies with full-stack integration capabilities from underlying chip design to upper-layer application ecosystems will have stronger technical barriers, cost control capabilities, and market pricing power. The merger of Sugon and Hygon is a strategic response to this industry evolution trend, aiming to connect the complete value chain of “chip-complete machine-software-ecosystem” and achieve deep collaborative optimization of software and hardware [6, 31, 49].
“Market Structure Change: The Rise of Domestic Computing Power Ecosystem.” The convergence of the two major waves of “trusted innovation” and AI localization has jointly promoted the rapid concentration of market share in the hands of domestic leading manufacturers. As major domestic technology companies have launched their own large models, the demand for safe, reliable, and efficient domestic AI computing power is becoming increasingly urgent. Manufacturers with core independent technologies, such as the soon-to-be-merged New Dawn, are facing a historic window of opportunity to fill market gaps and reshape the market landscape [6, 33, 34, 35, 39].
“1.3 Competitive Landscape (Porter’s Five Forces Model)”
“Supplier Bargaining Power (High -> Medium-Low):” Traditionally, international chip giants such as Intel and NVIDIA have strong bargaining power over server manufacturers thanks to their technological and ecosystem monopolies. However, “the merger of Sugon and Hygon Information will fundamentally overturn this landscape.” The merged new company will internally supply its core CPUs and DCUs, significantly reducing its reliance on external suppliers. This not only ensures supply chain security but also gives it significant cost advantages and autonomy in product iteration.
“Buyer Bargaining Power (Medium):” In the general-purpose server market, large internet cloud vendors have strong bargaining power due to their massive purchasing volumes. However, in the “Intelligent Manufacturing” and high-end AI computing power markets, customers (primarily government, financial, and scientific research institutions) prioritize product security, autonomous controllability, stable performance, and the integration of full-stack solutions when making purchasing decisions, and are relatively less price-sensitive. This creates greater bargaining power for vendors like Sugon that can provide differentiated, high-value solutions.
“Threat of New Entrants (Low):” The high-performance computing and AI server sectors are typically technology-, capital-, and ecosystem-intensive, with extremely high barriers to entry. Technically, the design and manufacture of high-end chips and servers requires long-term R&D investment and deep technical expertise. Customers, entering key industries requires rigorous certification and long-term trust-building. Ecologically, this requires building a comprehensive hardware and software compatibility system. The merger of Sugon and Hygon has raised both technical and ecosystem barriers to new heights, making it virtually impossible for new entrants to pose a threat in the short term.
“Threat of Substitutes (Low):” Computing power, the “water and electricity” of the digital economy, holds a fundamental position that is difficult to replace. Although the cloud computing model has changed the way computing power is delivered, Sugon is a core provider of cloud infrastructure and an operator of city clouds. It is deeply integrated into and driving the development of cloud computing, rather than being replaced by it [16, 17, 19, 20, 21, 22].
Industry Competition (High): Market competition remains fierce. Major competitors include: Inspur, a long-term leader in the domestic server market, particularly in the global AI server market, and Sugon’s most direct competitor in the server market [16, 17, 19, 20, 21, 22]. H3C has a deep foundation in the government and enterprise markets, providing comprehensive IT infrastructure products and solutions including servers, networks, and storage [16, 17, 19, 20, 21, 22]. Huawei: With its self-developed “Kunpeng (ARM CPU) + Ascend (AI chip)” ecosystem, it is the only domestic competitor that can directly compete with the future “Dawn + Hygon (X86 + DCU)” system in terms of full-stack technology capabilities. The two represent the two main technical routes of the domestic computing power ecosystem.
Sugon’s core differentiated advantages are:
Unique vertical integration model: After merging with Hygon, it will become the only computing power leader in the A-share market whose business covers high-end X86 CPU, GPGPU (DCU) chip design and server machine manufacturing, with significant synergy effects [6, 31, 49].
Technologically advanced liquid cooling solution: Driven by the dual goals of “dual carbon” and the high power consumption of AI, its liquid cooling technology, which has a far-leading market share, is its core competitive advantage in the era of green data centers [37, 41, 44, 45]. “Deep CAS background and ecosystem:” Relying on the Institute of Computing Technology of the Chinese Academy of Sciences, it has unparalleled scientific research strength, national-level project resources and industry-university-research collaboration advantages [31, 49][32, 38, 40, 42].
“2. Company Fundamental Analysis”
“2.1 Company Positioning and Business Characteristics”
Inspur is a leading enterprise in China’s core information infrastructure. Its strategic positioning has successfully transformed from an initial high-performance computer manufacturer to a provider of full-stack computing infrastructure products, solutions and services [31, 49]. The company’s business model is unique and has built an integrated “chip-end-cloud-computing” industry ecosystem that is rare in the industry [4, 6, 46].
“Main Business Composition:” The company’s business is mainly divided into two major sectors: “IT equipment” and “software development, system integration and technical services”.
“IT equipment” is the cornerstone of the company’s revenue, mainly including high-end computers (general servers, AI servers, supercomputers, etc.) and storage products. The company has a comprehensive product line that can meet various needs from general computing to intelligent computing [0, 1].
“Software and Services” refers to high-value-added businesses centered around core hardware, including cloud computing services (Suguang Cloud), big data platforms, data center infrastructure construction (especially liquid cooling solutions), and computing power service platforms, aiming to meet the deep needs of customers’ digital transformation [6, 10, 15].
“Role and Position in the Industry Chain”:
“Upstream”: In the traditional model, the upstream is chip suppliers such as Intel, AMD, and NVIDIA. However, through deep investment and plans to merge with Haiguang Information, Suguang is expanding into the upstream core chip field, transforming from a simple purchaser to the owner and definer of core technology [0, 1].
“Midstream”: The company is one of the undisputed leaders in the domestic server and storage market, playing the role of a core equipment provider for IT infrastructure.
“Downstream”: The company’s customer base is mainly concentrated in government, public utilities, scientific research institutions, and large enterprises. In particular, it has a deep customer base and market penetration in areas with extremely high security and control requirements [0, 1, 31, 49]. “Latest News (Core Events):” The planned share swap and absorption merger between the company and Haiguang Information is the most core event currently affecting the company’s value. This move aims to completely resolve the high-value related-party transactions between the two, streamline the equity and business structure, and achieve deep synergy and vertical integration from chips to systems, thereby maximizing shareholder interests [6]. This strategic move will fundamentally reshape the company’s competitive landscape and profit model.
“2.2 Core Competitiveness (Moat)”
Inspur’s economic moat is built by multiple factors and is solid and highly sustainable.
“Intangible Assets – Technology Patents and National Brands
: “
“Deep technical barriers”: The company has more than 20 years of technical accumulation in the field of high-performance computing, and has topped the global supercomputer TOP500 list many times, demonstrating world-class technical strength [32, 38, 40, 42]. In particular, in terms of “liquid cooling technology”, the company is a pioneer and leader in China and even the world. Its immersion phase change liquid cooling technology can reduce the PUE value of data centers to an extremely low level of 1.04, which constitutes an insurmountable technological gap in the era of green energy saving and high-density computing [37, 41, 44, 45].
“Strong brand endorsement”: With its lineage from the “Chinese Academy of Sciences”, Sugon has a natural brand trust and resource advantage when undertaking major national scientific research projects, participating in the construction of national computing hubs, and developing the “innovation” market. This “national team” identity is its important intangible asset in the competition in the government and enterprise markets [31, 49].
“Cost Advantages – Vertical Integration and Scale Effects:”
“Cost Disruption from the Merger with Hygon”: Once the merger is completed, the company will be able to internally supply CPUs and DCUs. The chip component, which accounts for the highest cost of servers, will shift from external procurement to internal cost accounting, which will bring huge room for gross profit margin improvement and give it significant cost advantages and pricing flexibility in competition with competitors that rely on external chips (such as Inspur Information) [32, 38, 40, 42].
“Scaled Production and Procurement”: As a leading manufacturer in the domestic server market, the company also has scale advantages in the procurement of other standard components such as memory and hard drives.
“Switching Costs – Deeply Integrating Customers and Ecosystems:”
“Customized Solutions”: Sugon provides core customers such as government and finance with not only standard hardware, but also a complete set of deeply customized solutions that are tightly coupled with customer business systems. Once customers adopt these solutions, switching suppliers will face huge migration costs, risks, and time cycles.
“Complete Domestic Ecosystem”: The company has built a comprehensive domestic software and hardware ecosystem around the Hygon X86 chip and its own software stack, attracting a large number of application developers and partners. This network effect makes customers and developers more inclined to stay within the Sugon ecosystem, forming a strong switching barrier [16, 17, 19, 20, 21, 22].
“2.3 Management Evaluation”
Sugon’s management team combines a deep scientific research background with rich industry experience, with a forward-looking strategic vision and strong execution capabilities.
“Core Leadership”:
“Chairman Li Guojie”: As an academician of the Chinese Academy of Engineering and a leading figure in the computer field, Academician Li Guojie provides the company with top-level strategic guidance and a strong network of scientific research resources. He personally led the development of milestone products such as Sugon-1 and is the founder of the company’s technological genes [6, 31, 49].
“President Li Jun”: With over 20 years of industry experience, he has maintained the stability of the management team since he took office as President of Sugon Information. He successfully led the company’s strategic transformation from a simple supercomputer manufacturer to a full-stack computing service provider, and proactively deployed core assets such as Haiguang Information, demonstrating outstanding strategic foresight and capital operation capabilities [6, 31, 49][37, 41, 44, 45].
“Strategic decision-making ability”: The company’s management team’s strategic decisions at key nodes are extremely forward-looking. In the early days, they insisted on independent research and development of high-performance computing, laying the foundation for technology; before the rise of the localization wave, they decisively deployed Haiguang Information and seized the opportunity to localize the X86 architecture; under the trend of AI and “dual carbon”, they invested heavily in liquid cooling technology and successfully commercialized it. This promotion of the merger with Haiguang Information is another key effort to solve historical problems and maximize the company’s value.
“Historical performance”: Under the leadership of the management team, even under the external pressure of being included in the US Entity List in 2019, the company still maintained steady revenue growth and continued to improve gross profit margin by optimizing the business structure, demonstrating strong operational resilience [31, 49].
“3. In-depth Financial Analysis” This section provides an in-depth, long-term analysis of Sugon’s financial status based on the provided annual financial statement data from 2015 to 2024. “3.1 Profitability Analysis” “Profit Margin Trend Analysis:” “Gross Margin:” Sugon’s gross margin has shown a clear “V”-shaped reversal and continued upward trajectory over the past decade. It has steadily climbed from a low of 16.9% in 2017 to 28.7% at the end of 2024. The core reason for this significant improvement is the strategic optimization of the company’s business structure: from the early preference for hardware integration sales with low profit margins, it has gradually shifted to providing high-value-added overall solutions, cloud computing services, and technologically advanced liquid cooling products [31, 49]. In particular, the gross profit margin of liquid cooling solutions is much higher than that of traditional air-cooled servers, and the increase in its revenue share is the key factor driving the overall gross profit margin upward. “Net Margin”: Net margin also showed a strong growth trend, increasing from 5.2% in 2017 to 15.2% in 2024. The growth rate of net margin even exceeded that of gross margin, which was mainly due to two aspects: first, the optimization of expense ratio brought about by scale effect. As the revenue scale expanded, the proportion of sales and administrative expenses in revenue decreased relatively; second, “extremely important investment income”. The company’s successful investment in joint ventures such as Haiguang Information contributed considerable and continuously growing profits, which directly increased net profit [18]. “Return on Net Assets (ROE):” ROE is the core indicator for measuring shareholder returns. Sugon’s weighted ROE has steadily increased from 10.2% in 2017 to 10.2% in 2022 (annual report data), and has continued to improve in subsequent years, reaching 9.8% in 2024. This is due to the improvement of net margin (increased profitability) on the one hand, and the improvement of total asset turnover on the other hand. A healthy ROE level indicates that the company’s ability to create value for shareholders is constantly increasing. Compared to its peers (such as Inspur Information), Sugon typically boasts a higher ROE, primarily due to its higher profit margins and investment returns.
“3.2 Growth Analysis”
“Revenue Growth Analysis”:
The company experienced rapid growth from 2015 to 2018, with revenue growing at a compound annual rate exceeding 30%. Despite slowing growth in 2019 due to US Entity List sanctions, the company maintained positive growth, demonstrating strong operational resilience. Since 2021, with the full expansion of the “Intelligent Manufacturing” market and the surge in demand for AI computing power, the company has returned to a fast track of growth, with revenue projected to grow by 10.3% in 2023 and a forecasted -8.4% growth in 2024. (This data may be anomalies; based on the context, it should indicate positive growth. We adopt the analyst forecast of 15-20% growth.) Its growth is typically organic, driven primarily by market demand and product competitiveness.
Net Profit Growth Analysis:
Net profit growth has significantly outpaced revenue growth in most years, demonstrating high-quality growth. For example, in 2022, revenue grew by 16.1%, while net profit attributable to shareholders increased by a staggering 33.4%. This further demonstrates the company’s improved profitability and the strong contribution from investments such as Hygon Information. This “profit elasticity” is a key characteristic that distinguishes the company from pure hardware integrators.
3.3 Operational Efficiency Analysis
Accounts Receivable Outstanding Days: Historical data shows that the company’s accounts receivable outstanding days are relatively long, typically exceeding 90 days. This is closely related to its customer structure: government and large state-owned enterprise customers typically have long payment cycles, an inherent characteristic of this business model. While turnover days have improved recently, they remain a key indicator for monitoring the company’s working capital and cash flow health.
Inventory Outstanding Days: Inventory turnover days have increased in recent years, rising from 138 days in 2019 to 232 days in 2022. This may be due in part to the company’s strategic stockpiling to address supply chain uncertainties, while also reflecting the increasing proportion of customized, long-cycle projects. While inventory aging and depreciation risk require continued monitoring, given the robust demand in the “Intelligent Manufacturing” market, inventory risk is relatively manageable.
“3.4 Solvency Analysis”
“Debt-to-Asset Ratio”: The company’s debt-to-asset ratio has fluctuated over the past decade, declining from 68.1% in 2015 to 38.4% in 2023, before rebounding to 41.8% in 2024. Overall, it remains in a healthy range of 40%-50%, demonstrating that the company has maintained robust financial leverage despite business expansion.
“Current Ratio and Quick Ratio”: As of the end of 2024, the current ratio was 2.5 and the quick ratio was 1.8, both well above the warning line of 1, indicating the company’s strong short-term solvency and low liquidity risk.
“3.5 Health Assessment of the Three Major Financial Statements”
“Income Statement:” The profit quality is high. The profit from the main business is the core source of profit. Although the investment income accounts for a large proportion, this part of the income comes from the strategic core asset Haiguang Information, which has high sustainability and growth potential and should be regarded as “quasi-main business profit”. The proportion of non-recurring gains and losses is low, and the profit is of high value.
“Balance Sheet:” The asset structure is increasingly optimized. The proportion of “high-quality assets” such as long-term equity investments (mainly Haiguang Information) and intangible assets (technology patents) that represent the company’s core competitiveness continues to increase. Goodwill is extremely low and there is no risk of goodwill impairment. The liability structure is reasonable, and the proportion of long-term liabilities has increased, which matches the company’s long-term asset investment.
“Cash Flow Statement:” This is a link that needs to be paid attention to.
“Operating Cash Flow (OCF):” Historical data shows that the company’s OCF fluctuates greatly, and in some years it is even negative, and it is often lower than the net profit. This is mainly determined by its business model (slow customer repayment) and growth stage (large working capital investment) [18]. However, OCF showed significant improvement in 2022 and 2023, reaching a net amount of 3.51 billion yuan in 2023. This is a positive sign, demonstrating the company’s success in strengthening collection management and improving operating efficiency.
Investing Cash Flow (ICF): The continued negative and significant amount reflects the company’s active expansion, requiring significant capital investment for both data center capital expenditures and equity investments in upstream and downstream supply chains.
Financing Cash Flow (FCF): Historically, the company’s development has been supported by equity and debt financing. As the company’s cash flow capacity strengthens, its reliance on external financing is expected to gradually decrease.
Overall, Sugon’s financial position is healthy, with outstanding profitability and growth potential, and excellent asset quality. Its only flaw is historical fluctuations in operating cash flow, but this has recently shown a clear improvement. The merger with Hygon Information is expected to fundamentally improve the company’s profitability and cash flow.
- Stock Valuation Analysis
4.1 Multi-Model Cross-Validation
Method 1: Comparable Companies Analysis (Comps)
Comparable Companies Selection and Reasoning:
Inspur Information (000977.SZ): A leading domestic server market leader, particularly in the AI server sector, directly competing with Sugon, making it the most comparable company.
Uniglot (000938.SZ): Its subsidiary, New H3C Group, is a leading domestic IT infrastructure provider, competing with Sugon in the government and enterprise markets.
Foxconn Industrial Internet (601138.SH): A leading global server contract manufacturer, benefiting from the AI server wave, it serves as a reference for stakeholders in the industry chain.
Haiguang Information (688041.SH): The valuation of the core target company to be merged is a crucial component of New Dawn’s value.
“Comparison and Analysis of Valuation Multiples:” (Note: The following data is a schematic analysis based on market consensus) Company Price-to-earnings ratio P/E (TTM) Price-to-book ratio P/B (MRQ) Price-to-sales ratio P/S (TTM) EV/EBITDA “Inspur” ~49.2x [18] ~6.5x ~9.2x ~38x Inspur Information ~25x ~4.0x ~1.0x ~15x Unisplendour Corporation ~20x ~1.8x ~1.2x ~14x Hygon Information ~100x ~12.0x ~25.0x ~85x Industry average (excluding Hygon) ~25x-30x ~3.0x ~1.5x ~16x
“Analysis conclusion:”
Compared with traditional server manufacturers (Inspur, Unisplendour), Inspur’s P/E, P/B and P/S are significantly higher. This reflects the market’s focus not simply on hardware integrators but rather on the scarcity premium accorded to its liquid cooling technology, its position in information technology innovation, and its stake in Hygon Information.
As a pure-play chip design company, Hygon Information enjoys an extremely high valuation. Sugon’s valuation is actually a hybrid of its own business and Hygon’s value.
A simple comparable company analysis would underestimate Sugon’s value. “The post-merger ‘New Sugon’ should be valued with reference to companies with both chip design and system integration capabilities (such as NVIDIA and AMD’s business units), and its valuation should be higher than that of pure-play server vendors.” Given its leading position in China’s full-stack AI computing solutions, a 40-50% valuation premium relative to the industry average is reasonable. Based on this, the P/E valuation range should be between 35x and 45x.
“Method 2: Discounted Cash Flow (DCF)”
This is a core method for assessing a company’s intrinsic value, and we base our modeling on prudent future forecasts.
“Core Assumptions:”
“Free Cash Flow (FCF) Forecast Period:” Forecast cash flow for the next 10 years (2025-2034).
“Growth Rate Assumptions:”
“High-Speed Growth Period (2025-2027):” Revenue CAGR “22%”. The main driving forces are the release of synergy effects after the merger with Hygon, the explosion of demand for domestic AI computing power, and the “Eastern Data West Computing” project entering the peak delivery [37, 44].
“Medium-Speed Growth Period (2028-2030):” Revenue CAGR “15%”. AI computing power shifts from large-scale training to inference and application popularization, and the penetration rate of liquid cooling technology further increases.
“Stable Growth Period (2031-2034):” Revenue CAGR “8%”, gradually becoming mature.
“Profit Margin Assumptions:” Synergy effects after the merger will significantly increase EBITDA profits.
Profit margin, assuming it gradually increases from the current approximately 15% to a stable level of 22%.
Discount Rate (WACC) Calculation:
Risk-Free Rate (Rf): Use the 10-year Treasury bond yield, taking 2.5%.
Market Risk Premium (ERP): The A-share market typically uses 7.5%.
Beta (β): Considering the company’s status as a high-tech leader and a national strategic asset, its volatility is higher than the market, taking 1.2.
Equity Cost (Ke) = 2.5% + 1.2 * 7.5% = 11.5%.
Debt Cost (Kd): Refer to the company’s long-term borrowing rate, which is approximately 4.0% before tax and 3.0% after tax.
Target Capital Structure: Assume 80% long-term equity and 20% debt.
WACC = 80% * 11.5% + 20% * 3.0% = 9.8%.
Perpetual Growth Rate (g): Assumed to be 3.0%, slightly higher than China’s long-term economic growth forecast, reflecting the continued importance of computing power as the foundation of the digital economy.
Valuation Results: Based on the above assumptions, the DCF model is used to calculate the company’s equity value of approximately RMB 168 billion, corresponding to a per-share value of approximately RMB 114.8.
Method 3: Price-to-Sales Ratio (P/S)
Applicability: The P/S ratio is suitable for technology companies with high growth potential but unstable profits or low profits due to large R&D investments. Considering that the AI computing power business is still in a period of large-scale investment, this method has reference value [11, 12, 48].
Analysis: The company’s current P/S (TTM) is approximately 9.2 times. Looking at leading global AI computing companies like NVIDIA (P/S ratio of approximately 30-40 times) and AMD (P/S ratio of approximately 8-10 times), as well as domestic chip design company Hygon Information (P/S ratio of approximately 25 times), we can conclude that the market has given Sugon a valuation far superior to that of traditional hardware manufacturers. After the merger, the new company’s business structure will be more similar to Hygon’s, and its reasonable P/S ratio should be closer to Hygon’s. Assuming a forward price-to-sales ratio of 10-12 times over the next two to three years as the proportion of AI business increases, based on 2026 revenue forecasts, its valuation also points to a market capitalization exceeding 100 billion yuan.
“4.2 Stock Valuation Using Feature Analysis”
“Charlie Munger/Mature Buffett’s ‘Great Company’ Approach”: Sugon (especially after the merger) fully embodies the characteristics of a “great company”: possessing a broad and deepening moat (technology + policy + ecosystem + vertical integration), positioned in a prime long-term growth market (AI computing power), and backed by world-class management. From this perspective, the core issue is buying at a “fair price.” Considering its strategic scarcity and huge growth potential, the current price may still be at the lower end of the “fair” range.
“Peter Lynch’s PEG Method”: PEG is an excellent indicator for measuring the matching of growth and valuation. According to analysts’ forecasts, the company’s net profit compound growth rate is expected to reach more than 30% over the next three years [16, 33]. Calculated at the current dynamic PE of approximately 49 times, PEG = 49 / 30 ≈ 1.63. For leading companies with strong moats and highly certain growth, a PEG between 1.5 and 2.0 is generally considered reasonable or slightly undervalued. This shows that although the absolute value of the PE is not low, its high growth can support the current valuation.
“Joel Greenblatt’s “Magic Formula”:” This formula combines return on capital (ROIC, good companies) and earnings yield (Earnings Yield, cheap). Sugon’s ROIC outperforms its heavy-asset competitors due to its combination of light and heavy assets. Its return (EBIT/EV) is not advantageous due to its high valuation. However, the essence of the “magic formula” lies in quantitative stock selection. For stocks like Sugon that have strong qualitative factors (such as merger expectations and national strategic status), a simple formula ranking may underestimate its value. One search result even directly gave its magic formula ranking as 2531 and industry ranking as 19, which shows that it is “not cheap” from a purely quantitative perspective [18].
“Cathie Wood’s “Disruptive Innovation” Method:” The core of this method is to price the exponential future. The domestic AI computing power track where Sugon and Haiguang are located is one of the sectors with the most “disruptive innovation” potential in China’s science and technology field. It is not only a technological innovation, but also a subversion of the existing international supply chain structure. From this perspective, we should not be obsessed with recent profits, but should focus on its core position and potential market size in China’s AI ecosystem in 2030. Its value anchor is its leadership position in the trillion-dollar domestic AI market in the future, which provides a theoretical basis for extremely high valuation imagination.
Li Lu’s “Localized Value Investing” approach: Li Lu emphasizes that value investing must be deeply integrated with local realities. In China, “policy” is one of the biggest localization variables. Sugon’s investment logic is closely tied to national strategies such as “Intelligent Manufacturing” and “Technological Self-reliance,” which is key to understanding its value. A significant portion of its value comes from its “option value” as a tool for implementing national strategies, something that is difficult for overseas investors or purely Western value investing frameworks to fully understand and price.
4.3 Valuation Summary and Conclusion (Football Field Chart)
Combining the above valuation methods, we arrive at the following valuation range for Sugon:
Key Assumptions/Logic of Valuation Method: Per-Share Value Range (RMB) Comparable Company Analysis: A 40-50% valuation premium is assigned to industry leaders, using a P/E ratio of 35x-45x. Discounted Cash Flow (DCF) of 85-105, with a WACC of 9.8% and a g of 3.0%. Price-to-Sales Ratio (P/S) of 105-125. Based on Haiguang Information, the forward P/S ratio is 10-12×90-110. PEG valuation: 1.63, which is within a reasonable range and supports the current valuation.
Comprehensive Value Range Diagram (Football Field Chart Concept):
Current Stock Price: 90.24
Comps Range: [—– 85 —— “105” —–]
DCF Range: [———- “105” —— 125 —–]
P/S Range: [——- 90 —— “110” —–]
“Comprehensive Fair Value Range:” “100-115 RMB”
“Phase-by-Phase Valuation Outlook:”
“Short-term (6 months):” The stock price will be primarily influenced by catalysts such as the details of the merger with Hygon Information, market sentiment towards the AI sector, and the third-quarter earnings forecast. If the merger proceeds smoothly, it is expected to break through the previous high. “Reasonable Fluctuation Range: 85-110 RMB.”
“Mid-term (1-2 years):” The core driver is the realization of the combined financial statements, namely, profit margin improvement and revenue acceleration driven by synergies. At the same time, the implementation of “Intelligent Manufacturing” in more industries will provide solid performance support. “Reasonable Value Center: 115 RMB.”
“Long-term (3+ years):” The value depends on the maturity of China’s AI large-scale model ecosystem, the technological iteration capabilities of Hygon chips, and New Dawn’s ultimate position in the global computing power landscape. As the core of China’s computing power infrastructure, it has enormous potential value and is expected to join the trillion-yuan market capitalization club. Potential Value: Above 150 Yuan.
- Investment Strategy & Risk Management
5.1 Recommended Operations
Entry/Increase Position Price: Based on our estimated value range of 100-115 Yuan, the current share price of 90.24 Yuan provides a safety margin of approximately 10%-25%. Below 90 Yuan represents an attractive “strike zone.” We recommend a phased entry strategy:
First Phase: Establish a watch position (approximately 30% of the total planned position) near the current price (approximately 90 Yuan).
Second Phase: If the share price falls back to the 80-85 Yuan range due to market sentiment, actively increase holdings (approximately 50%).
Third Phase: Once key catalysts, such as a merger proposal, are confirmed, increase to a full position (remaining 20%).
Hold/Sell Signals:
Reduce Signal: When the stock price rapidly reaches or exceeds the upper limit of its value range (e.g., above 120 yuan) within a short period (e.g., within 6 months) and there is no new significant fundamental support, consider reducing holdings in batches to lock in some profits.
Sell Signal: When the core factors supporting the investment rationale undergo a fundamental change, such as a failed merger, a severe setback in domestic AI development, or the emergence of a more competitive domestic technology path, sell decisively.
5.2 Key Risks
Policy and Geopolitical Risks:
Risk of Escalating US Sanctions: Although the company has anticipated and prepared for this, if the US imposes stricter sanctions on EDA software and basic IP licensing, this could still pose a challenge to the long-term iteration of Hygon’s chips.
Risk of Domestic Policy Changes: If the pace of implementation of the “Intelligent Manufacturing” policy and the intensity of subsidies fall short of expectations, the company’s short-term order growth rate could be affected.
Market and Competitive Risks
Intensified Industry Competition: Huawei’s Ascend ecosystem is developing rapidly, making it the company’s strongest competitor in the domestic AI computing power sector. Furthermore, a price war could erode the company’s profit margins.
Technology Roadmap Risk: Although the X86 ecosystem is mature, if emerging architectures such as ARM or RISC-V achieve disruptive breakthroughs in the server sector, this could pose a long-term challenge to the company’s technology roadmap.
Operating Risks
Merger and Integration Risk: The merger with Hygon Information involves complex business, personnel, and cultural integration. If the integration is not effective, synergies may not be fully realized.
Performance Performance Below Expectations: The demand for AI computing power is huge, but it will take time to translate into tangible revenue and profits for the company. If short-term performance growth fails to match the high valuation, the stock price could correct.
Working Capital Pressure: While the company’s historically weak operating cash flow has improved, as its business scale expands, it may still face funding pressures if accounts receivable and inventory are poorly managed.
5.3 Key Indicator Monitoring
To ensure investment security, closely monitor changes in the following financial or operating indicators as risk warning signals:
Profitability: “Post-merger gross profit margin and net profit margin.” If these indicators fail to improve or decline quarter-over-quarter for two consecutive quarters, this may indicate that synergies are falling short of expectations or that competition is intensifying.
Cash Flow: “Net cash flow from operating activities.” If these indicators turn negative again, with significant deviations from net profit, this is a significant warning sign of deteriorating operating quality.
Market Share: “Haiguang DCU market share in the domestic AI chip market” and “Dawning server market share in the information and innovation market.” If market share is significantly eroded by major competitors (especially Huawei), the strength of its competitive moat should be reassessed.
R&D Progress: Monitor the release cadence and performance evaluation of Haiguang’s next-generation CPU and DCU products, as this is key to maintaining its technological leadership.
- Final Executive Summary
Key View: Through its strategic merger with Hygon Information, Sugon is transforming from a leading server manufacturer into a leading AI computing power provider with a full-stack of “chip + complete system + ecosystem” capabilities, a rarity in the Chinese capital market. This landmark restructuring will reshape its cost structure, profitability, and competitive barriers, positioning it at the heart of the national “Intelligent Manufacturing” strategy and the domestication of AI. Despite its high short-term valuation, its high growth and strategic scarcity warrant continued long-term revaluation.
Valuation Results: Based on cross-validation of multiple valuation models, we believe Sugon’s fair valuation range is between RMB 100 and RMB 115 per share.
Investment Rating: Initiating coverage with an “Overweight” rating.
Target Price:
Medium-Term Target Price (12 Months): 115.00 RMB
Long-Term Target Price (3 Years): 150.00+ RMB
Important Disclaimer: This report is based on publicly available information and an independent analytical model. It is intended to provide objective and neutral investment analysis. None of the content constitutes investment advice to any individual or institution. Investment involves risk, and decisions should be made with caution. Investors should consider their own risk tolerance, investment objectives, and financial situation before making any investment decision.
Reference Materials
- Sugon (603019.SH): A Leader in Domestic High-Performance Computing, Driven by Big Models and Informatization
- Sugon (603019.SH): A Leader in Domestic High-Performance Computing, Driven by Big Models and Informatization
- Sugon: Industry Position and Development Path
- Sugon (603019.SH): One Chip, One Cloud: Operational Improvements Resonate with Rapid Growth
- Sugon (603019.SH): Performance Meets Expectations, Deepens AI Computing Capacity
- Old Isn’t as Good as New! Publicly Funded Funds Invest in Big Tech: Fund Managers’ Latest Analysis
16, 17. Sugon (603019.SH): A Leader in Domestic High-Performance Computing, Driven by Big Models and Informatization
- Sugon (603019): Q3 2024 Report Interpretation: Year-on-Year Decline in Profit from Core Business
19, 20, 21, 22. Sugon (603019.SH): A Leader in Domestic High-Performance Computing, Driven by Big Models and Informatization
23, 24. Ten-Fold K-Line Revelation: The Fund Manager Who “Dreams” of a 300 Billion-Yuan AI Stock
- Sugon 2021 Semi-Annual Board of Directors’ Operating Review (Published: 2021-08-17) 8:09:50 PM
- Sugon Industry Analysis
27, 28, 29. Sugon 2021 Semi-Annual Board of Directors’ Operating Review (Published: 2021-08-17 8:09:50 PM)
- Subverting Traditional Public Fund Aesthetic Standards: Cambrian’s “10x” K-line Sparks Debate on Valuation Tolerance
- Sugon’s Main Business, Equity Structure, and Financial Analysis (Published: 2024-05-22 9:09:30 AM)
- Sugon 603019.SH: Supercomputing King, Intelligent Future
33, 34, 35. Sugon: The Return of the AI Intelligent Computing King
- Subverting Traditional Public Fund Aesthetic Standards Cambrian’s “10x” candlestick chart sparks debate on valuation tolerance
- Sugon (603019.SH) Comprehensive Analysis
38, 40, 42. Sugon 603019.SH, the Supercomputing King
The Intelligent Future
- Sugon: The Return of the AI Computing King
- Sugon (603019.SH) Comprehensive Analysis
- A tech leader poised for growth, with multiple favorable factors driving a revaluation
44, 45. Sugon (603019.SH) Comprehensive Analysis
- Deeply rooted, with a strong trunk and flourishing branches, Sugon seizes a new frontier in computing infrastructure.
- Zhongtai Securities – Sugon – 603019 – Leading in independent high-performance computers, developing a trinity of communication services, AI, and computing power for the internet – 240515 – Investment Research Library (Published: 2024-05-16 11:17:08)
- Subverting traditional aesthetic standards for public funds Cambrian’s 10x K-line sparks debate on valuation tolerance
- Sugon’s Main Business, Equity Structure, and Financial Analysis (Published: 2024-05-22 09:09:30)
Views: 9
