Spring AI 2.0.0 has been released, providing significant improvements and a new baseline for developing AI applications. This release is built upon Spring Boot 4 and Spring Framework 7, featuring enhanced JSON serialization with Jackson 3 and full null-safety annotations improving code reliability. Options handling has been refactored for clarity and immutability, with builders replacing constructors. The core project now focuses on a refined set of well-supported chat model providers like OpenAI, Anthropic, and Google GenAI, leveraging vendor SDKs for better adaptability. Collaboration with the community has been strengthened through updated CONTRIBUTING.md guidelines. Agentic AI capabilities have been greatly improved, with tool execution loops now first-class composable components within the advisor chain. Unified tool calling, progressive tool disclosure for scaling to hundreds of tools, and structured self-correcting output mechanisms are new additions. The release also introduces Community Extensions for event-sourced conversation memory and agentic patterns. Model Context Protocol (MCP) integration is enhanced with an official Java SDK, annotation-driven programming, and new transport implementations including streamable HTTP. Enterprise features like Micrometer spans, OpenTelemetry metrics, and OAuth 2.0 security are also inherited. Spring AI 2.0 aims to significantly improve the developer experience and pave the way for future AI features.