在这片中国科技产业密度最高的区域之一，21世纪经济报道记者再次见到了云天励飞董事长陈宁。长达 1个半小时的深度交流中，陈宁侃侃而谈关于AI的一切。 五年前，记者对陈宁的专访，正值中国AI创业最为喧闹的阶段。新锐企业群雄并起，算法、场景与资 本交织，云天励飞也刚刚踏上上市之路。五年过去，陈宁依旧精神饱满，而云天励飞的坐标，已经从 AI解决方案，转向更底层，也更具长期价值的AI推理芯片赛道。 有观点认为，当下AI投入正在积聚泡沫。对此，陈宁有不同的看法："AI就像蒸汽机刚出现的时候。站 在一个村庄的视角，可能会觉得这是泡沫，但站在历史的角度看，这是一个时代的起点。AI一定会经 历泡沫和调整，但方向本身不会错。" 训练时代，英伟达当之无愧的王者，也是标准制定者。但在推理时代，陈宁认为，"所有人都站在同一 条新的起跑线上。谁能在成本、效率和系统能力上建立优势，谁就有机会。" 当前，推理已经进入到算力的中心舞台，毫无疑问，接下来AI赛场将构建起更加繁荣的推理芯片和应 用生态。在未来的硬件架构中，可能有更多异构组合，有专门用来做通用计算CPU，有专门做训练的 GPU、也有专门做推理的推理芯片。 深圳湾科技生态园中，AI弄 ...

Core Viewpoint - The AI training market in China is entering a competitive phase dominated by major companies, with a significant reliance on large orders from these firms to sustain market activity [2][3]. Group 1: AI Training Market Analysis - Tencent has sufficient training chip reserves and does not face chip shortage concerns, focusing on using the best available models from various suppliers [2]. - The training market is currently dominated by NVIDIA, with over 60% of training card demand driven by Alibaba, followed by ByteDance and Tencent [3]. - The "Six Little Dragons" are withdrawing from training resources, negatively impacting the overall training market, as these companies are still in the early stages of commercialization [3]. Group 2: Competition Among Major Players - The competition between Alibaba and ByteDance is intensifying, with both companies striving to excel in large model training, leading to a zero-sum game scenario [3]. - The demand for training resources is primarily concentrated among major companies, with Tencent continuing to invest in next-generation models despite the competitive landscape [3]. Group 3: Market Trends and Future Outlook - The demand for inference computing power has not seen the expected significant growth, despite initial optimism earlier in the year [4]. - The growth of AI applications, such as Yuanbao, has begun to slow down, with a modest increase in monthly active users and a significant drop in monthly downloads [4]. - The influx of second-hand A100 and H100 training devices into the domestic market is expected to lower prices significantly, impacting the compliance card market [4][5]. Group 4: Investment Allocation Among Companies - Alibaba allocates approximately 80% of its budget to training and 20% to inference, while ByteDance maintains a balanced 50:50 ratio [5][6]. - Tencent's investment distribution is approximately 20% for training and 80% for inference, indicating a product-oriented approach that has not yet yielded positive revenue [5][6].

Core Insights - The article emphasizes three key principles for success in business: selecting the right trends, having a long-term perspective, and maintaining a high frequency of iteration, with the latter being a counterintuitive insight [1][4]. Group 1: Key Principles - The first principle is the importance of accurately selecting major trends, such as in e-commerce and food delivery, which is a common understanding among successful entrepreneurs [2]. - The second principle highlights the necessity of a long-term approach, suggesting that significant results often take 15 to 20 years to materialize, contrasting with short-term gains that may attract more competition [2][4]. - The third principle, which is less commonly accepted, stresses the need for rapid iteration within a chosen long-term path, as seen in successful companies like Nvidia and Meituan, which adapt quickly based on market feedback [4][9]. Group 2: Examples and Comparisons - Companies like Nvidia exemplify the principle of high-frequency iteration, releasing new products annually compared to competitors who may take several years, thus maintaining a competitive edge [8][12]. - The article draws parallels between business iteration and reinforcement learning, where real market feedback is crucial for growth and improvement, emphasizing that practice and adaptation are essential [9][11]. - The discussion also notes that striving for perfection can hinder progress, and that successful companies often prioritize rapid iteration over perfection [11].