Catz表示，预计未来几个月还将有几家数十亿美元的客户签约。 如果您希望可以时常见面，欢迎标星收藏哦~ 来 源 ： 内容来 自 wsj 。 甲骨文作为云计算领域后进者的日子已经一去不复返了。该公司似乎在周二的财报电话会议上提醒 了听众这一点，接连宣布了多项数十亿美元的大单。所有这些交易的核心都是：人工智能。 据首席执行官Safra Catz称，这家拥有近 50 年历史的数据库软件巨头在截至 8 月 31 日的季度 中，与三家不同的客户签署了四份数十亿美元的合同。尽管本季度业绩未达到分析师预期，但甲骨 文报告称，其总剩余履约义务（预计将收取的合同收入）达到 4550 亿美元。 Ellison 表示，甲骨文在运行 AI 模型（即 AI 推理）方面的独特优势在于，它通过其数据库掌握 着许多企业的私有数据。Ellison 说，公司的新型 AI 数据库将使这些企业能够轻松地查询或询问 其私有数据，所有这些都将增加甲骨文 AI 推理服务的账单。 Newman 表示："推理是增长最快的工作负载，也将创造最大的机遇。"Newman 补充说，凭借其 庞大的数据和基础设施，甲骨文在 AI 推理方面拥有引人注目的差异化优势，能够很 ...

Core Insights - The rise of AI agents is transforming AI from a tool to an assistant and even an agent, with 2025 being termed the "Year of AI Agents" [1] - The AI agent marketplace is emerging as a platform for developers to publish, distribute, integrate, and trade AI agents, allowing businesses to select and deploy agents for complex automation tasks [1][2] - Major companies like Salesforce, Moveworks, and Siemens are launching their own AI agent marketplaces, each targeting different segments of the enterprise market [3][5][6] Group 1: AI Agent Marketplace Development - Salesforce launched AgentExchange in March 2025, providing a channel for users to explore, trial, and purchase AI agents, thus accelerating the adoption of AI applications [3] - AgentExchange differs from Salesforce's existing AppExchange by offering "skills" and "capabilities" modules specifically for AI agents, aiming to create a marketplace for digital labor [4] - Moveworks introduced its AI agent marketplace in April 2025, featuring over 100 pre-built agents for various enterprise applications, simplifying the integration of AI into existing systems [5] Group 2: Industry-Specific AI Agents - Siemens announced its Industrial AI agents at the Automate2025 conference, promising up to a 50% increase in production efficiency through autonomous AI systems that can coordinate various AI agents [6] - Siemens aims to create an industrial AI agent marketplace within its Xcelerator Marketplace, allowing access to both Siemens' and third-party AI agents [6][7] Group 3: Business Models of AI Agent Marketplaces - The AI agent industry can be segmented into four layers: large language models, intermediary agent frameworks, AI agent construction/deployment platforms, and enterprise/consumer application layers [8] - Various business models are emerging, including Agent-as-a-Service, which offers specific AI agents through subscription or usage-based pricing [10] - Additional models include value-added services through agent plugins, platform operation models with revenue sharing, and custom agent development for large enterprises [12][13][14] Group 4: Future Competition and Ecosystem - The future competition in the AI agent market will focus on the quality of agents, plugins, and integration capabilities, as well as the ability to foster a thriving developer and user ecosystem [15]