当市场给智谱定价时，中国大模型正在被一并评估

Core Insights - The article discusses the evolving identity of Zhipu AI as it transitions from a traditional ToB service provider to a significant player in the global AI model competition, particularly in the context of its IPO and the rapid advancements in its model capabilities [2][3][16]. Group 1: Company Overview - Zhipu AI's revenue from localized deployment accounts for 85%, which has led to a narrative of being an AI outsourcing company focused on delivery rather than API services [3]. - The company has shifted its commercialization strategy from "model delivery" to "model invocation," aiming for a balanced revenue stream between localized deployment and model-as-a-service (MaaS) [2][8]. - CEO Zhang Peng indicated that the future revenue split between localized deployment and MaaS will be 50-50, reflecting a strategic pivot in their business model [2]. Group 2: Market Position and Competitiveness - Zhipu AI has been compared to OpenAI, not just in product form but in its foundational technology, positioning it within a global competitive framework for large models [2][16]. - The company has released multiple iterations of its GLM model (GLM-4.5, 4.6, and 4.7) in rapid succession, establishing itself as a key player in the developer ecosystem [3][5]. - Zhipu AI's models have gained traction among global developers, with significant adoption and integration into mainstream AI platforms, indicating its transition from a domestic alternative to a global competitor [8][10]. Group 3: Financial Performance and Growth - The company reported a significant increase in paid developer subscriptions, surpassing 150,000 within two months of launching its GLM Coding plan, leading to an annual recurring revenue exceeding 100 million [8]. - In November, Zhipu AI achieved a daily token invocation volume of 4.2 trillion, countering the narrative of being merely a delivery-focused company [9]. - The overall gross margin has shown improvement as model invocation scales up, indicating a positive trend in financial performance despite initial losses [13]. Group 4: Industry Context and Future Outlook - The AI market is expected to grow significantly, with estimates suggesting that the Chinese B-end AI market could reach 100 billion RMB by 2030, highlighting the potential for Zhipu AI's growth [13]. - The article emphasizes that the valuation of Zhipu AI should not be based solely on past financial data but rather on its potential as a foundational model provider in the evolving AI landscape [16][17]. - Zhipu AI's strategic focus on building trust with enterprise clients through localized deployments and its independent model development is seen as a competitive advantage in the global market [15][20].