新年伊始，百度就正式发布公告，启动昆仑芯科技于港交所主板独立上市的计划。 这不仅是百度2026年的开年大戏，也被视为国产AI算力领域的一个标志性事件。 分拆上市背后的动机 对于当前的百度来说，虽然传统广告搜索业务在萎缩，但它依然还是一头现金奶牛，百度当下并不缺 钱。那么，百度要分拆昆仑芯上市的背后动机究竟是什么？ 如今的昆仑芯，不仅是百度All in AI的"硅基底座"，更是目前国内极少数经过大规模生产验证、且在特 定场景下能与国际巨头"掰手腕"的AI芯片。2025年，百度昆仑芯的整体估值已经达到了约210亿元。 那么，昆仑芯到底有多硬核呢？ 这当然还是离不开昆仑芯的实战性能。 这第一个战略动机应该是价值重估。一直以来，百度在资本市场的眼中都是一家"互联网广告"平台，这 在很大程度上压制了百度在AI领域的估值。此番百度把昆仑芯作为独立的硬科技公司分拆上市，就能 享受到更高的科技股估值溢价，而百度自身的资本价值也能得到更好的释放。 这第二个战略动机应该是独立发展。分拆上市之后，昆仑芯将成为一家独立面向市场的AI芯片供应 商，这将大大有助于昆仑芯拓展百度生态之外的客户，从而打破昆仑芯只有内部订单的质疑。 此外，昆 ...

1月2日，壁仞科技在港交所上市，成为港股"国产GPU第一股"，同时也是港股2026年上市的第一只股票。这家由前商汤总裁张文2019年创立的GPU企业，正 式填上了中国算力版图的关键一环。 当日，壁仞科技盘初快速拉升涨幅一度超过100%，市值超940亿港元。壁仞科技发行价为19.6港元，发行2.85亿股，募资总额为55.83亿港元。 壁仞科技在国产AI芯片版图中占据着重要的一席，其在通用GPU领域拥有多项技术专利，技术先进性可比肩国际巨头。截至2025年12月15日，壁仞科技在 全球多个国家和地区累计申请专利1500余项，位列中国通用GPU公司第一，获得专利授权600余项，位列中国通用GPU公司前列；发明专利授权率达100%， 位列国内企业发明专利授权率榜首。 但是从业绩来看，2022年—2024年及2025年上半年，公司收入分别为49.9万元、6203万元、3.37亿元、0.59亿元，年内亏损分别为14.74亿元、17.44亿元、 15.38亿元、16.01亿元。 由此计算，壁仞科技三年半以来产生亏损达63.57亿元。公司解释称，较大幅度亏损主要系若干前股东的赎回权相关的"赎回负债账面值变动"，该变动为非 现 ...

Group 1 - The H200 chip from Nvidia is expected to attract Chinese cloud vendors for procurement after its sales restrictions are lifted, despite its performance being lower than the current Blackwell series chips [2] - The H200 chip shows a significant performance improvement over the H20 chip, with a peak theoretical computing power of 1979 TFLOPS compared to H20's 148 TFLOPS, and it features 141 GB of HBM3e memory versus H20's 96 GB of HBM3 [2] - The procurement of the H200 chip will depend on formal policies, and cloud vendors are likely to test domestic AI chips to mitigate risks [3] Group 2 - The entry of the H200 chip into the Chinese market may intensify competition in the high-end chip sector, with local companies focusing on performance, compliance, and cost-effectiveness to strengthen their market position [4] - If Nvidia's H200 and AMD's MI325 can be exported to China, they could capture about 30% of the high-end AI chip market by 2026, while domestic ASIC chips are expected to hold 20% [5] - The overall supply of high-end AI chips in China is projected to grow by over 60% by 2026 [5]

美国司法部近日侦破一起高科技芯片走私大案，两名华裔男子因涉嫌违反出口管制法规，非法向中国转 运价值逾1.6亿美元（约12.4亿港元）的英伟达高端AI芯片而被捕。 案件详情 涉案的两名嫌犯分别是43岁居住在纽约的中国公民Fanyue Gong，以及58岁的加拿大籍中国公民 Benlin Yuan。他们被指控与一家香港物流公司及一家中国AI科技公司的员工共谋，规避美国出口管制政策。 政策背景与市场影响 值得注意的是，就在美国司法部公布此案的当天，美国总统特朗普宣布将允许英伟达向中国"获准客 户"出售H200芯片，但条件是美国政府将从销售额中抽取25%的分成。这一决定被视为英伟达的重大游说胜 利，可能帮助其重新夺回在中国这一关键市场中损失的数十亿美元业务。 英伟达CEO黄仁勋此前曾表示，由于美国出口管制，英伟达100%退出了中国市场，中国市场份额从 95%降到了0%。 执法力度持续加强 美国司法部检察官尼古拉斯·甘杰在声明中表示："'守门人行动'揭露了一个复杂的走私网络，该网络通过 将尖端AI芯片输送给那些试图利用它对抗美国利益的人，这威胁着美国的安全。这些芯片是人工智能优势的基 石，是现代军事应用的核心。" 该犯 ...

当地时间12月8日，美国总统特朗普宣布，将允许英伟达对华出口其H200人工智能（AI）芯片。 "我已经通知中方，美国将允许英伟达向中国及其他国家客户出口其H200芯片，"特朗普当天在自创社交 媒体平台"真实社交"发文称，"中方对此表示积极回应"。 特朗普还称，此决定有附带条件，包括保障美国所谓"国家安全"，且美国政府将获得25%的分成。他 说，美国商务部正在敲定协议细节，同样的方式将适用于其他AI芯片公司，如AMD和英特尔。 他写道："该政策将支持美国就业，强化美国制造业，并惠及美国纳税人……我们将保护国家安全，创 造美国就业机会，保持美国在AI领域的领先地位。" 特朗普同时表示，英伟达的最新Blackwell芯片以及即将发布的Rubin芯片，不在获批名单中。 特朗普在白宫会晤黄仁勋 视频截图 英国《金融时报》称，这一决定标志着与拜登政府的做法截然不同，拜登政府曾对芯片相关技术实施全 面的出口管制，试图阻碍中国半导体现代化进程。 8日稍晚时候，英伟达发言人表示，此举"对美国来说，是非常有利的、深思熟虑的平衡"。发言人指 出："我们赞赏特朗普允许美国芯片产业参与竞争，以支持美国的高薪就业和制造业发展的决定。" ...

国芯网[原:中国半导体论坛] 振兴国产半导体产业！ 不拘中国、 放眼世界 ！ 关注 世界半导体论坛 ↓ ↓ ↓ 11月24日消息，近日，据外媒最新报道称，美国政府正考虑允许英伟达对华出售H200芯片。 报道援引知情人士消息称，负责监管美国出口管制的美国商务部正就改变对华出口限制一事进行审查，并称相关计划可能会发生变动。 目前，美国商务部暂未对此作出回应，英伟达尚未就此直接置评。 两年前发布的H200芯片相比于其前代产品H100芯片拥有更多的高带宽内存，使其能够更快速地处理数据。据估计，H200芯片的性能是英 伟达H20芯片的两倍。 对于黄仁勋来说，他们迫切希望重新能够对中国厂商出手英伟达芯片，因为除了能够获得利润外，还能够在一定程度上抑制华为等国产厂 商的发展势头。 半导体公众号推荐 半导体论坛百万微信群 加群步骤： 第一步：扫描下方二维码，关注国芯网微信公众号。 第二步：在公众号里面回复"加群"，按照提示操作即可。 爆料｜投稿｜合作｜社群 文章内容整理自网络，如有侵权请联系沟通 "由于美国出口限制，英伟达对华芯片销售陷入停滞，预计未来两个季度在华销售额将为零。伤害中国的事情，往往也可能伤害美国，甚至 会更严 ...

Group 1 - The Trump administration is considering approving the export of NVIDIA's H200 AI chips to China, which have significantly improved performance compared to the previous H100 chips, with H200 estimated to be twice as powerful as H100 [1] - The H200 chip features HBM3e memory, providing a memory speed of 4.8TB per second and a memory capacity that is approximately double that of the A100, with a bandwidth increase of 2.4 times [1] - NVIDIA's H200 NVL, based on the Hopper architecture, offers a 1.5 times increase in memory capacity and a 1.2 times increase in bandwidth compared to H100 NVL, enhancing performance for large language model fine-tuning [1] Group 2 - Google’s TPU is considered the only AI accelerator that can compete with NVIDIA's GPUs, leveraging frameworks like TensorFlow and OpenXLA to build a comprehensive AI ecosystem [2] - Google is increasing its capital expenditure to meet strong demand for AI infrastructure, with a projected Capex of approximately $91-93 billion for 2025 and significant increases expected in 2026 [2] - Google has established a leading position in the industry with top-tier capabilities in reasoning, multimodal abilities, agent tool usage, multilingual performance, and long context handling [2] Group 3 - Zhongji Xuchuang is a main supplier of optical modules for Google, with products like silicon photonics and 1.6T already in mass production, and a 3.2T product currently under development [3] - TeraHop, a subsidiary of Zhongji Xuchuang, has launched the first silicon photonics-based 64x64 OCS switch, which reduces power consumption for AI clusters and aids in network architecture [3] - Dahong Technology has developed spatial intelligence technology similar to Google's nano banana technology, utilizing optimized Gaussian splashing techniques for 3D modeling from multi-angle images [3]

Core Viewpoint - The article discusses the potential easing of U.S. export restrictions on advanced technology to China, particularly focusing on NVIDIA's H200 AI chip, which may soon be approved for sale to the Chinese market due to a thaw in U.S.-China relations [3][4][6]. Group 1: U.S.-China Relations and Export Policies - Recent discussions between U.S. and Chinese leaders have led to a consensus on easing trade and technology disputes, indicating a shift towards a more moderate U.S. stance on China [6]. - The U.S. Department of Commerce is currently evaluating adjustments to export control policies that have previously prohibited the sale of advanced chips to China [4][6]. Group 2: NVIDIA's H200 Chip - The H200 chip, released two years ago, boasts significantly improved specifications compared to its predecessor, the H100, with higher bandwidth memory and faster data processing capabilities [6]. - Estimates suggest that the performance of the H200 chip is approximately double that of the NVIDIA H20 chip, making it a highly competitive product in the AI chip market [7]. Group 3: Geopolitical Implications and Market Dynamics - The selective approval of chip exports to allied countries, such as Saudi Arabia and the UAE, highlights a differentiated approach by the U.S. government in response to varying geopolitical contexts [7]. - The potential approval of the H200 chip for sale to China could allow NVIDIA to regain market share in a rapidly growing sector that has been increasingly dominated by foreign competitors [4][6].

Group 1: Optical Modules - Coherent, a leading company in the optical module sector, reported Q1 revenue of $1.58 billion, exceeding expectations of $1.54 billion, with adjusted EPS of $1.16 against an estimate of $1.05, driven by strong demand in AIDC and communications [1] - The growth in demand and continuous capacity expansion are expected to sustain strong growth throughout the fiscal year, leading to a significant pre-market stock price increase [1] - The evolution of AI data center architecture is expected to drive high-end optical module demand, with a projected global Ethernet optical module market growth rate of around 50% by 2025, stabilizing at a CAGR of 15% to 18% over the following five years [1] Group 2: Semiconductors - AMD's CEO revealed that the Instinct MI308 AI chip has received export approval to China, positioning it as a strong competitor to Nvidia's H20 chip, although future revenue from China is not yet included in Q4 forecasts [2] - AMD's advancements in AI, particularly with the MI350 series accelerators, show industry-leading memory bandwidth and capacity, achieving competitive performance in key workloads while reducing costs and complexity [2] - The rapid development of AI chips is creating new opportunities in semiconductor equipment and related sectors [2] Group 3: Nuclear Energy - The 2025 Fourth China Nuclear Energy High-Quality Development Conference will be held from November 12-14, focusing on global nuclear energy industry collaboration [3] - Asia has been the primary driver of global nuclear power generation growth over the past decade, with 56 out of 68 new operational units located in Asia [3] - Nuclear power plants maintain high annual utilization hours, exceeding 7000 hours, and are expected to play a larger role in seawater desalination, hydrogen production, and medical protection in the future [3] Group 4: 5G - The International Electrotechnical Commission (IEC) has released the first global industrial 5G international standard, addressing the integration of 5G technology in industrial applications [4] - This standard is a significant milestone for China's contributions to the "5G + Industry" integration, with projections indicating a global industrial 5G market size of approximately 345.75 billion yuan by 2025, expanding at a CAGR of 41.32% from 2025 to 2032 [4] Group 5: Commercial Aerospace - The 27th China International High-Tech Achievements Fair will take place from November 14-16, featuring the global launch of China's space tourism project [5] - The fair will cover various sectors, including satellite communication and space tourism, with predictions that the global space economy could exceed $1 trillion by 2030 [5] Group 6: OLED - Three high-end equipment for the mass production line of 8.6-generation large-size OLED screens were officially launched in Chengdu, marking a breakthrough in display equipment [6] - OLED technology is expanding from mobile devices to larger applications, including computers and automotive displays [6]

Group 1 - A-shares and Hong Kong stocks showed strong performance, particularly in the semiconductor sector, with Haiguang Information rising over 10% and Cambrian Technology increasing by over 7% [1] - The surge in domestic chips is attributed to AMD's announcement of receiving export licenses for its Instinct MI308 AI chip to China, which is a strong competitor to NVIDIA's H20 chip [1] - Analysts believe that technology security remains a key theme amid geopolitical tensions, with AI and domestic substitution having long-term certainty [1] Group 2 - On November 6, over 14.2 billion yuan of main funds flowed into the electronics sector, making it the second-highest inflow among 31 Shenwan primary industries [2] - The electronic ETF (515260) saw an intraday increase of 3.15%, recovering both the 5-day and 20-day moving averages, indicating a potential positive trend reversal [2] - Key stocks in the sector included Dongshan Precision rising over 10%, Lens Technology increasing by over 8%, and several others like Cambrian Technology and Haiguang Information also showing significant gains [2] Group 3 - The electronic ETF (515260) and its linked funds passively track the electronic 50 index, focusing on semiconductor and consumer electronics industries, including AI chips and 5G [4] - The external environment is pushing China to achieve self-sufficiency in the semiconductor supply chain, with AI reshaping consumer electronics and enhancing user experience [4] - National policies and industry support are expected to drive the electronics sector's growth [4]