2026年的春节，中国互联网的底色是人工智能。"春节AI大战"已上演半个月，阿里巴巴发起了一场堪称教科书的AI"闪电战"，完成了从基础大模型到AI应 用的双线进击。 2月17日大年初一凌晨，千问发布数据显示，春节活动期间，全国超过1.3亿人第一次体验AI购物，说了50亿次"千问帮我"，千问一跃成为国民级AI助手。 可以说，阿里AI在这个春节获得了两条战线的全面胜利：千问跻身成为"国民AI"，Qwen3.5则登顶全球最强开源模型。这场胜利，是阿里十年磨一剑的战 略兑现。"通云哥"组成的"黄金三角"协同作战，构筑了全球罕见的"自研芯片+云平台+大模型"全栈垂直整合体系。拥有自研AI芯片"真武"，意味着阿里可 为千问量身定制硬件架构；而阿里云则如同"最懂模型的调度员"，实现资源效率最大化。这种软硬一体的"全栈AI"模式，正是谷歌TPU与Gemini成功逻辑 的翻版，也是阿里掌握核心算力定义权的关键一步。 放眼全球，唯有阿里同时具备顶尖模型能力与庞大实体生态，真正补齐了"从意图到执行"的最后一块拼图。这场由千问点燃的AI生活化浪潮，或将重塑整 个行业的竞争格局——技术为骨，生态为翼，谁能在真实世界中让AI"落地生根" ...

Core Insights - Huang Renxun, CEO of Nvidia, emphasizes the importance of the Chinese market during his recent visits, highlighting the company's strategic focus on China [1][2] - Nvidia has not yet received orders for its H200 AI chips from Chinese customers due to pending import decisions from the Chinese government [1] - The H200 chip is positioned as a transitional power support for large model inference and training, with significant specifications [5][6] Company Strategy - Huang's travel itinerary indicates a strong emphasis on the Asia-Pacific region, particularly China, Japan, and India, reflecting Nvidia's strategic priorities [1] - The company employs approximately 36,000 people globally, with significant representation in China (11%) and India (14%) [1] - Huang's approach in China includes a mix of business and cultural engagement, showcasing a more public relations-focused strategy compared to other regions [4][5] Market Dynamics - The Chinese AI chip market is becoming increasingly competitive, with local companies like Huawei, Alibaba, and Baidu developing their own AI chips [6] - TrendForce predicts that the proportion of AI servers in China using imported chips will decrease from 63% in 2024 to 42% in 2025, indicating a rise in domestic chip suppliers [6] - The competition between the US and China in the computing power sector highlights each country's strengths, with the US excelling in chip computing and AI models, while China benefits from low-cost energy and manufacturing capabilities [6][7] Ecosystem Development - The AI ecosystem is complex, involving data, energy, chip manufacturing, and AI models, necessitating global collaboration for advancement [7] - Huang's background and Nvidia's role position the company as a key player in building this ecosystem, with China capable of being a significant contributor [7]

Core Insights - Huang Renxun, CEO of Nvidia, emphasizes the importance of the Chinese market during his annual visit, highlighting the urgency to revitalize business in China [1][2] - Nvidia has not yet received orders for the H200 AI chips from Chinese customers due to pending import decisions by the Chinese government, although the chip is deemed suitable for the market [1][3] - The global AI chip landscape is competitive, with Chinese companies like Huawei, Alibaba, and Baidu emerging as significant players, potentially reducing reliance on foreign chips [4][5] Group 1: Nvidia's Strategic Focus - Huang Renxun's travel itinerary reflects a strategic focus on the Chinese market, with significant visits to China, Japan, and India, indicating a prioritization of the Asia-Pacific region [2] - Nvidia employs approximately 36,000 people globally, with 11% in China, suggesting a substantial commitment to the Chinese market despite geopolitical tensions [2] - The H200 chip, which is currently under U.S. export license approval, is positioned as a transitional power support for large model inference and training, featuring 141GB HBM3e and 4.8TB/s bandwidth [3] Group 2: Competitive Landscape and Market Dynamics - The Chinese AI server market is projected to see a decrease in reliance on foreign chips, with local suppliers expected to increase their market share from 2024 to 2025 [4] - The competitive advantages in the AI field are split between the U.S. and China, with the U.S. excelling in chip computing power and AI models, while China benefits from lower energy costs and extensive application scenarios [5] - The AI ecosystem is complex, involving data, energy, chip manufacturing, and AI models, indicating that collaboration across global players is essential for advancement [5]

Group 1 - Nvidia, Microsoft, and Amazon are in talks to invest approximately $60 billion in OpenAI [6] - A new artificial intelligence industry fund has been established in Kunshan with a total scale of 5 billion yuan [1] - Alibaba's T-head has launched its self-developed AI chip "Zhenwu" [2] Group 2 - Beijing has officially launched its first humanoid robot pilot verification platform, expected to have an annual production capacity of 5,000 units [3] - 360 has announced the public testing of its industrial-grade AI comic intelligent agent production platform, named "Nano Comic Production Line" [7] - Kunlun Wanwei has released the Mureka V8 music large model [8] Group 3 - MiniMax Music 2.5 has been officially released [4] - Yushu has open-sourced UnifoLM-VLA-0 [5] - Songyan Power has announced its partnership as the humanoid robot collaborator for the 2026 CCTV Spring Festival Gala [9]

Core Viewpoint - The semiconductor sector is experiencing a short-term pullback, with the Huabao ETF (589190) showing a decline of 1.75% after an initial rise, indicating active buying interest during dips [1][11]. Group 1: Market Performance - The semiconductor sector shows mixed performance, with leading companies like Zhongwei, Lanke Technology, and SMIC experiencing declines of over 4% and 3% respectively [3][11]. - The Huabao ETF tracks the Shanghai Stock Exchange's semiconductor index, which includes 50 companies across various semiconductor sectors, maintaining a high technology barrier and significant weight in core areas [4][14]. Group 2: Price Adjustments - A new wave of price increases in the semiconductor industry has been reported, with Samsung and SK Hynix raising prices for LPDDR memory used in iPhones by over 80% and nearly 100% respectively [3][13]. - Domestic companies like Zhongwei Semiconductor and Guokewai have announced price hikes for MCU and Norflash products ranging from 15% to 50%, and KGD products by 40% to 80% [3][13]. Group 3: Future Outlook - CITIC Securities forecasts a high certainty in computing power development by 2026, highlighting opportunities in super-node technology and the competitive strength of domestic computing power manufacturers [3][13]. - Donghai Securities anticipates significant growth in the performance of domestic A-share companies related to AI by 2025, recommending a focus on computing power chips, semiconductor equipment, and storage [3][13]. Group 4: Index Performance - As of the end of 2025, the annualized return of the Shanghai Stock Exchange's semiconductor index is 17.93%, outperforming similar indices and showing a better risk-return profile [6][16]. - The index has shown varying annual performance, with a notable increase of 61.33% projected for 2025 [6][17].