（科技日报 潘宇菲 操秀英 何沛苁）责任编辑：李梦一 相关稿件： 光伏组件、汽车轻量化部件刚需的POE（聚烯烃弹性体），进口依存度曾超95%。这一局面终于被打 破！中国石油旗下独山子石化公司2025年共生产POE产品近6万吨，标志着国内首创POE气相聚合工艺 实现工业化量产与规模化供应。详情点击视频！ ...

1月21日，国际计算机学会（Association for Computing Machinery，ACM）公布2025年度ACM Fellow名 单，智象未来创始人兼首席执行官梅涛入选。 免责声明：以上内容为本网站转自其他媒体，相关信息仅为传递更多信息之目的，不代表本网观点，亦 不代表本网站赞同其观点或证实其内容的真实性。如稿件版权单位或个人不想在本网发布，可与本网联 系，本网视情况可立即将其撤除。 自2023年创立智象未来以来，梅涛带领团队持续深耕多模态对齐、跨模态语义理解、多模态生成基础模 型等关键技术，推动生成式AI实现从算法创新到产业应用的全链路突破。 素材源：吕怡蕾 编辑：康书源 审核：王明月 2025年4月，智象未来自主研发的开源图像生成大模型HiDream-I1在国际权威榜单Artificial Analysis中24 小时内登顶。同年7月，其研发的开源图像编辑模型HiDream-E1.1跻身Artificial Analysis图像编辑智能体 榜单第一梯队。 作为国际性计算机科学专业组织，ACM于1947年创立，旨在促进计算机科学和信息技术的发展。ACM Fellow是该组织授予会员的最高荣 ...

绿色低碳是高质量发展的必由之路。"十四五"期间，国家能源集团着力推动传统能源低碳转型和清洁能 源规模集约发展，火电产业升级脱硫脱硝设备、优化燃烧系统、推广超低排放技术，供电煤耗下降4.7 克/千瓦时，实现高效低碳运行；电力清洁可再生能源装机规模达到"十三五"末的3倍，水电装机增长达 16%，风电装机突破7200万千瓦，持续保持全球第一。 此外，在产业布局方面，"十四五"期间，国家能源集团服务国家重大战略，推动一批标志性工程先后落 地，牵头开发宁夏腾格里等"沙戈荒"大型风光基地，投运全球最大的开放式百万千瓦级海上光伏项 目，"国能共享号"深远海漂浮式风电项目开创"绿色能源+蓝色粮仓"融合发展新模式，建成黄河上游最 大的玛尔挡水电站，打造千万千瓦级"水光蓄储"一体化基地。（陆成宽） 责任编辑：闫弘旭 记者从1月23日举行的国家能源集团2026年工作会新闻发布会上获悉，2025年，国家能源集团多项生产 经营指标再创历史最好水平，自产煤量站稳6亿吨大关，单日发电量多次刷新纪录，供热量保持全国第 一，电煤中长协高比例签约履约，以实际行动诠释了能源保供"主力军"的责任与担当。 企业是科技创新的主体，是提供高质量科技供给的 ...

科技日报记者 吴叶凡 箱梁是现代大型桥梁、高架桥中最主要、最常见的桥梁结构，为一座座"飞虹"跨越天堑提供着坚实 支撑。 从传统的箱梁预制，到如今的智慧拼装，智能化正深刻重塑着我国梁场的制造模式，提升着桥梁的 建造效率与品质。 搭建信息"高速公路" 近期国新办举行的新闻发布会上，交通运输部综合规划司负责人金敬东介绍，"十四五"时期，我国 重大交通基础设施项目总体进展顺利。在这一进程中，桥梁建设尤为关键。 在位于湖北省黄冈市麻城市的麻城梁场，一榀榀箱梁在智能化流水线上完成制造，井然有序"排队 出场"，为合武高铁建设贡献"钢筋铁骨"。 中铁一局合武高铁麻城梁场总工程师董贺介绍，麻城梁场围绕"数智化"转型升级目标，以AI智能化 管理系统为基座，通过运用建筑信息模型（BIM）、物联网、AI智能识别、智能终端等先进技术，对生 产、人员、质量、安全、技术、设备、工序、物资8个方面进行全过程智能化管控，实现N个智能化工 装设备的深度链接。 "形象地说，箱梁制造过程中，有上百条看不见、摸不着的信息通道。这些通道将调度中心与各个 生产单元链接起来。"董贺进一步解释，"我们要做的，就是打通各个信息通道间的壁垒，搭建信息'高 速公路 ...

当传统锂电技术遭遇"天花板"，固态电池正以颠覆性姿态掀起能源革命的巨浪。固态锂电池是突破现有 液态锂离子电池能量密度与安全瓶颈的关键路径。其中，固态电池核心材料稳定性、规模化量产等世界 性难题，严重制约其产业化进程。近日，记者从江西省科技重大专项"2030先锋工程"专项"高比能固态 锂电池开发与产业化"的项目承担团队获悉，高比能固态锂电池目前已迎来关键进展。 近日，记者在有研新能源材料（江西）有限公司（以下简称有研新能源）的富锂锰基材料全自动生产线 上看到，其各项指标正在稳步运行，前不久还实现首批次量产交付，正式发货至行业头部客户。固态电 池核心材料——富锂锰基正极材料的研发与攻关一直是行业堵点。 "目前，我们已在连续化前驱体合成与表界面一体化调控方面取得阶段性突破，初步形成了具有自主知 识产权的材料体系。""固态锂电池用富锂锰基正极材料开发"专项负责人、有研新能源总经理张向军 说，项目围绕专项任务要求，系统开展了富锂锰基正极材料的前驱体合成、烧结工艺、界面优化及全电 池应用验证等研发工作。 目前，项目总体进展顺利，已达到中期既定目标。"富锂锰基材料在固态体系中的界面构建与稳定化是 我们团队研究的重点。"项目 ...

Group 1 - The core viewpoint of the articles highlights that the State Energy Group achieved record production and operational indicators in 2025, including a stable coal production of 600 million tons and multiple daily power generation records, demonstrating its commitment to energy supply security [1] - The company is actively promoting the transformation of scientific research achievements into practical productivity, with notable accomplishments such as winning the China Patent Gold Award for low-temperature Fischer-Tropsch synthesis catalysts and the successful operation of the world's largest "coal power + molten salt" energy storage project [1] - During the 14th Five-Year Plan period, the company has intensified technological innovation efforts, resulting in the establishment of nine national-level intelligent coal mine demonstrations and breakthroughs in technologies such as secondary reheating of coal-fired power and high-ratio ammonia combustion [1] Group 2 - The company emphasizes that green and low-carbon development is essential for high-quality growth, focusing on the low-carbon transformation of traditional energy and the large-scale development of clean energy during the 14th Five-Year Plan [2] - The company has upgraded its thermal power industry by implementing desulfurization and denitrification equipment, optimizing combustion systems, and promoting ultra-low emission technologies, achieving a reduction in coal consumption for power generation by 4.7 grams per kilowatt-hour [2] - The installed capacity of clean renewable energy has reached three times that of the end of the 13th Five-Year Plan, with hydropower capacity increasing by 16% and wind power capacity exceeding 72 million kilowatts, maintaining a global leadership position [2]

Group 1 - The two 14-meter diameter shield machines "Xiyi" and "Xiwang" have successfully rolled off the production line and will be used for the Wuxi to Yixing Intercity Rail Transit Project, marking the first ultra-large diameter shield tunnel project crossing Taihu Lake in China [1] - The total length of the Wuxi to Yixing S2 line is approximately 61.48 kilometers, with the Taihu section being a core control node of 21.6 kilometers, requiring four shield machines to excavate in five sections [1] - The first two shield machines will undertake tunnel construction for approximately 7.6 kilometers and 5.5 kilometers in designated work areas, integrating multiple advanced domestic innovative technologies [1] Group 2 - The design speed of the S2 line is 160 kilometers per hour, featuring nine stations and will facilitate a "half-hour commuting circle" in Wuxi, promoting integration between Wuxi and Yixing as well as the Yangtze River Delta [2]

Core Viewpoint - The commercial space tourism industry is evolving from concept to reality, with companies like Beijing Chuan Yue Zhe and others planning to offer manned space flights by 2028, as evidenced by the recent global launch event for the "Chuan Yue Zhe No. 1" spacecraft [1] Group 1: Forms of Space Tourism - Current international space tourism activities are categorized into three forms: suborbital, orbital, and deep space tourism [2] - Suborbital tourism allows passengers to experience a brief weightlessness at altitudes around 100 kilometers, with companies like Virgin Galactic and Blue Origin leading in this area [2] - Orbital tourism involves longer stays in low Earth orbit, with SpaceX conducting missions that include commercial spacewalks [2][3] Group 2: Technical Challenges for Space Tourism - Several domestic commercial space companies are targeting the space tourism sector, with Chuan Yue Zhe having completed key design and testing phases for its reusable suborbital spacecraft [4] - The company plans to conduct critical technology validations by 2026 and aims for manned flights in 2028 [4] - Other companies like Ziwei Technology and Zhongke Aerospace are also making progress in suborbital flight tests and technology validations [4] Group 3: Requirements for Ordinary People to Enter Space - Ordinary individuals do not need to be "superhuman" but must meet certain health criteria, including no severe underlying conditions and the ability to withstand 3 to 4 Gs of force [7] - The primary physiological challenges include the physical overload during launch and re-entry, as well as potential motion sickness in weightlessness [7] - The cost of space tourism remains a significant barrier, with prices for tickets currently exceeding millions, but there is potential for price reductions in the future as the industry matures [8]

一年前，恰是这个周一，深度求索（DeepSeek）公司R1大模型以高性能和低训练成本震撼世界， 外媒称之为"DeepSeek时刻"。 DeepSeek的横空出世，可谓2025年中国科技创新的"开场戏"。近期，国际机构密集发文盘点过去一 年全球科技发展，来自中国的成果格外耀眼。 这源于实现高水平科技自立自强的决心与努力。在美国先进芯片的封锁禁运下，DeepSeek以算法 创新突破算力限制，华为、阿里巴巴、百度、寒武纪等中国企业则纷纷踏上芯片技术攻关之路。随着 DeepSeek-R1大模型和"昇腾"等国产芯片问世，美国战略界终于意识到，出口管制管不住中国，只会使 中国加速建设一套独立的技术体系。 这倚赖高水平科研人才提供的核心支撑。2025年11月，科睿唯安发布《2025年度全球高被引科学 家》榜单，中国内地共有1406人次入选，占全球总数的19.7%。在AI领域，据全球知名科研技术公司 Digital Science估算，中国拥有超3万名研究人员，是美国的3倍，中国仅博士后和博士生数量已是美国 研究人员的2倍以上，在长期创新方面具有得天独厚的优势。 这离不开全社会对中国科技发展的坚定信心。2025年以来，中国A ...

Core Viewpoint - The recent wave of IPOs by domestic AI companies in China signifies a transition from a technology-driven exploration phase to a capital-driven industrialization phase, characterized by rapid revenue growth, significant R&D investment, and ongoing losses [1][2]. Group 1: Transition from Exploration to Growth Phase - The current IPO wave indicates that China's AI industry has surpassed the foundational development phase and is entering a growth phase marked by large-scale investment, industry chain collaboration, and accelerated application penetration [1][2]. - AI development inherently features high fixed costs and low marginal costs, with substantial initial investments in computing power, data management, and talent, but decreasing unit costs as scale increases [1]. Group 2: Role of Capital Markets - The IPOs suggest that capital markets are increasingly taking over the role of major internet companies in allocating computing power and R&D resources, shifting from short-term funding to long-term financing [2]. - The listed companies cover critical areas of AI development, including computing power, foundational models, and vertical applications, indicating a shift from isolated innovations to a collaborative ecosystem [2]. Group 3: Market Challenges Post-IPO - Post-IPO, AI companies face stricter public scrutiny and immediate profitability pressures, necessitating a balance between long-term R&D investments and short-term market expectations [3]. - Companies are advised to establish verifiable R&D milestones and present measurable indicators of success to align with market expectations and reduce the impact of short-term fluctuations on R&D [3]. Group 4: Commercialization Strategies - Companies should develop a layered cash flow strategy to secure funding for long-term R&D, diversifying revenue streams through subscriptions, API calls, and service deployments [4]. - The goal is to use more stable revenue sources to buffer against the uncertainties of cutting-edge R&D, ensuring that innovation does not compromise profitability [4]. Group 5: New Competitive Landscape - Continuous R&D investment and capital expenditure are essential for both model and chip layers, with many companies planning to allocate IPO proceeds towards foundational model development and platform capabilities [5]. - The financing boom is expected to accelerate innovation cycles and shift the focus from isolated breakthroughs to project-based and product-oriented developments, raising industry entry barriers [5]. Group 6: Macro Trends in AI - The global AI market is projected to reach $900 billion by 2026, with China expected to see a growth rate exceeding 30% during the same period, highlighting its position as a rapidly growing market [6]. - The effectiveness of this financing wave in fostering sustainable innovation will depend on directing capital towards long-term research and maintaining a balance between efficiency, ecosystem boundaries, and compliance [6].