AI应用收入两极分化：模型公司与系统集成商挤压独立应用层
英文摘要
A new article argues that AI application startups are being squeezed from above by model companies that absorb popular features, and from below by system integrators with direct model vendor backing. It highlights revenue data: OpenAI ARR reportedly hit $25B, Anthropic over $47B, while Cursor reached $500M and Glean $300M—only a few apps cross $100M. Early tools like Jasper were quickly replicated by ChatGPT, and AI search faces direct competition from model-native search. The author sees the biggest threat in the Anthropic-DXC alliance, which bypasses app developers for enterprise delivery. The future, it contends, lies in AI-powered physical services (robotaxi, AI hospitals) where value extends beyond software.
中文摘要
文章通过收入数据指出AI应用公司正被两头挤压：上层是模型厂商直接提供功能（如搜索、编码），下层是Anthropic与DXC等集成商联盟绕过独立应用。OpenAI年经常性收入约250亿美元，Anthropic超470亿美元，而Cursor、Glean等头部应用仅数亿，大量公司收入不足千万。Jasper等早期明星被模型自身消灭，AI搜索、视频也面临类似风险。作者认为纯软件应用生存空间锐减，未来真正的AI应用需转向Robotaxi、AI医院等端到端的实体服务，在物理世界构筑护城河。
关键要点
Top model providers dominate revenue: OpenAI ARR ~$25B, Anthropic ARR >$47B, while the best independent apps (Cursor, Glean) only reach $0.3–0.5B.
模型公司收入遥遥领先：OpenAI年经常性收入约250亿美元，Anthropic超470亿美元，而最成功的独立应用Cursor约5亿美元、Glean约3亿美元。
Many early AI apps like Jasper and AI search tools lost their edge as model companies added the same features natively, illustrating the "application innovation cycle limit".
Jasper等早期明星应用因ChatGPT原生支持写作而衰落，AI搜索也面临ChatGPT/Claude内置搜索的挤压，凸显“应用创新周期有限性”。
Anthropic's alliance with DXC (a team of tens of thousands of certified engineers) marks a shift: model companies and system integrators now bypass independent app vendors directly into enterprise systems.
Anthropic与IT服务巨头DXC结盟，组建数万名认证工程师团队，使模型厂商与集成商直接向企业交付，绕开独立应用层。
AI coding apps like Cursor face direct competition from model-native tools (Claude Code, OpenAI Codex), raising questions about long-term viability even for the most successful sector.
即使是AI编程这一最成功赛道，Cursor也面临Claude Code等模型原生工具的强烈竞争，说明应用层压力已扩散至核心领域。
The author foresees future AI applications moving beyond software into end-to-end physical services (robotaxi, AI hospitals) where value comes from real-world assets and business outcomes, not just model calls.
作者认为未来AI应用必须跳出纯软件，转向Robotaxi、AI医院等端到端实体服务，利润来自物理资产和完整商业闭环，而非仅调用模型。