Group 1 - The Zhongzheng Information Technology Application Innovation Industry Index (300832) has decreased by 2.73%, with component stocks showing mixed performance [1] - The top gainers include Borui Data up by 6.74%, Geer Software up by 5.47%, and Yingshisheng up by 4.97%, while the biggest losers are Fanwei Network down by 9.99%, Zhuoyi Information down by 9.41%, and Foxit Software down by 7.44% [1] - The Xinchang ETF (562570) has dropped by 2.62%, with the latest price at 1.49 yuan, and has seen an active trading volume with a turnover of 10.4% and a transaction value of 44.996 million yuan [1] Group 2 - The Xinchang ETF has experienced a significant increase in shares, growing by 9 million shares over the past week, ranking in the top third among comparable funds [1] - The latest net inflow of funds into the Xinchang ETF is 605.38 million yuan, with a total net inflow of 15.7309 million yuan over the last five trading days, averaging 3.1462 million yuan per day [1] - Yuntian Lifei has announced its strategic layout for large-scale AI inference chips over the next three years, focusing on reducing the cost barriers for large model implementation, aiming for a 100-fold reduction in inference costs using the GPNPU architecture [1] Group 3 - Galaxy Securities anticipates that demand catalysts and intensive bidding will lead to a new cycle for domestic computing power [2] - The Nvidia H200 chip is expected to conditionally enter the Chinese market, which will benefit the development of domestic computing power chips and ecosystems in the long term [2] - The Xinchang ETF closely tracks the Zhongzheng Xinchang Index, emphasizing domestic full-stack substitution and integrating key areas such as storage chips, CPUs, and AIPC, while actively participating in the DeepSeek ecosystem (72%) and AI applications (58%) [2]

Core Insights - The company announced its strategic focus on large-scale AI inference chips, aiming to reduce the cost of inference for million tokens by over 100 times within the next three years [2][6] - The global computing power industry is shifting towards inference capabilities, with major players like Google and NVIDIA emphasizing system optimization for efficiency and cost reduction [4][5] Group 1: Company Strategy - The company has established the GPNPU technology route, defined as GPGPU + NPU + 3D stacked storage, to address the challenges of portability, deployability, and sustainable cost reduction [5] - The CEO highlighted five key elements of the company's competitive advantage: technology, production capacity, ecosystem, market, and capital, which collectively support the company's strategic goals [5] - The company is one of the few in China with sufficient domestic production capacity, ensuring high certainty for large-scale chip production and delivery [5] Group 2: Industry Trends - The competition in the inference era is shifting from merely enhancing model parameters to improving application efficiency, focusing on lower inference costs and delivery efficiency [4] - The roadmap aims to align with international mainstream platforms, optimizing key inference stages like long context pre-filling and low-latency decoding to achieve cheaper, more stable, and easier deployment [6] - The essence of competition in the inference era is the cost per inference unit, which must be made affordable and stable for AI to transition from a visible capability to an accessible productivity tool [6]

这些行业信号共同指向一个趋势：推理侧竞争已不再单纯是"把模型做得更强"的参数竞赛，而是"让应用跑得更久、更稳、更便宜"的效能竞赛，单位推理成 本与交付效率已成为规模化落地的最大门槛。 2月3日，云天励飞正式举办"大算力芯片战略前瞻会"，首次对外公布未来三年的大算力 AI 推理芯片战略布局。面对人工智能从"基础模型构建"迈向"规模化 应用落地"的重要转折点，公司宣布将核心研发资源集中于攻克大模型落地的"成本壁垒"，致力于通过底层架构创新，力争实现百万 Tokens 推理成本降低 100 倍以上的目标，推动 AI 从技术尝鲜走向普惠生产力。 一、 产业变局：推理竞速，从"参数内卷"转向"效能为王" 过去一年，全球算力产业的风向标已发生显著偏转，重心正加速向推理侧倾斜。谷歌在 2025 年 4 月发布第七代 TPU "Ironwood"时，明确将其定位为"面向 推理时代"的基石，强调在大规模推理与能效上的系统化优化。 与此同时，围绕"更低时延、更低成本"的推理芯片与系统能力，产业整合动作也在加速。2025 年 12 月，英伟达与 Groq 达成非独占许可安排，并吸纳其核 心工程人才团队加入，此举被视为强化推理与实时 ...