News

.NET 10 Arrives with AI Integration, Performance Boosts, and New Tools

Microsoft announced .NET 10 at this week's .NET Conf 2025, calling it the most modern, secure, and performant version of the platform yet.

The Long Term Support (LTS) release adds built-in AI integration through the new Microsoft Agent Framework, major performance gains across the runtime and languages, and updated tooling in Visual Studio 2026 and Visual Studio Code. Microsoft said the release unifies all .NET workloads under one platform and extends the framework's support through November 2028.

As AI continues to transform Microsoft's developer tooling and the .NET ecosystem itself, the centerpiece of the new .NET 10 release is its built-in AI support. Microsoft describes .NET 10 as a unified, intelligent platform that lets developers move seamlessly from simple AI integrations to sophisticated multi-agent systems. This marks one of the most comprehensive expansions of AI capabilities ever introduced in .NET.

 AI in .NET 10
[Click on image for larger view.] AI in .NET 10 (source: Microsoft).

AI Integration and Multi-Agent Framework
At the core of this evolution is the Microsoft Agent Framework, designed to make it easier to build intelligent, agent-based applications. The framework combines prior technologies like Semantic Kernel and AutoGen into a single experience for creating and orchestrating multiple cooperating AI agents. It supports sequential, concurrent, and handoff workflows, along with a new group chat model where multiple agents can collaborate in real time. Developers can host these agents using ASP.NET Core and visualize them through an integrated Dev UI.

For user interfaces, the release introduces AG-UI, a lightweight, event-based protocol for building interactive, streaming agent interfaces. The protocol supports shared state management, tool invocation, and front-end collaboration. It can be implemented using the Microsoft.Agents.AI.Hosting.AGUI.AspNetCore package and connected to existing AG-UI client frameworks or custom .NET UI solutions built with MAUI or Blazor.

The Microsoft.Extensions.AI library further expands this foundation, offering standardized abstractions for AI services. The unified IChatClient interface enables developers to use AI models from providers such as OpenAI, Azure OpenAI, GitHub Models, or Ollama through a consistent API. Middleware, dependency injection, and telemetry features are built in, aligning with .NET's existing design patterns for scalability and observability.

Another key advancement is first-class support for the Model Context Protocol (MCP), which allows agents to securely access external APIs, databases, and services. Developers can create MCP servers directly with new .NET templates and publish them as NuGet packages, enabling reusable, composable AI capabilities across projects and organizations.

Performance and Language Enhancements
While AI integration headlines the release, Microsoft also calls .NET 10 its fastest version yet. Performance gains span runtime, JIT compilation, and memory management. Hardware acceleration now includes Intel AVX10.2 and Arm64 SVE optimizations, with improved loop inversion and stack allocation reducing garbage collection pauses. NativeAOT apps compile smaller and run faster.

C# 14 and F# 10 bring expanded language features aimed at writing more concise, expressive code. C# adds field-backed and extension properties, enhanced overload resolution, and new ref struct interfaces for allocation-free patterns. F# introduces scoped warning control, struct-based optional parameters, and a preview of parallel compilation to improve build speed and IDE responsiveness.

Libraries, Security, and Web Development
Microsoft expanded post-quantum cryptography support with ML-DSA and ML-KEM algorithms, added TLS 1.3 on macOS, and introduced a simplified WebSocketStream API. JSON serialization now enforces safer defaults and higher throughput. Distributed applications benefit from Aspire 13, which integrates service discovery, telemetry, and container orchestration with support for multiple languages and cloud workflows.

ASP.NET Core introduces automatic memory pool eviction, passkey-based authentication, and improved ahead-of-time compilation for APIs. Blazor improves performance, state persistence, and automated validation, while the new QuickGrid features and JavaScript interop APIs enhance component-based development.

Cross-Platform, Data, and Tooling Improvements
.NET MAUI adds global XAML namespaces, compile-time XAML generation, updated platform bindings, and extended control capabilities. Entity Framework Core 10 delivers AI-ready vector search, hybrid semantic and full-text search, and native JSON data types in Azure SQL and SQL Server. The SDK itself introduces new CLI features, SLNX project files, and enhanced dependency auditing through NuGet integration.

Visual Studio 2026, released alongside .NET 10, brings deeper Copilot integration, adaptive paste assistance, and Profiler Agents that generate performance recommendations automatically. The IDE also adds Fluent UI theming, improved debugging, and native support for Aspire projects. The C# Dev Kit for Visual Studio Code expands Razor editing and introduces full test coverage visualization and orchestration for distributed applications.

About the Author

David Ramel is an editor and writer at Converge 360.

comments powered by Disqus

Featured

  • Hands On with GitHub Copilot App Technical Preview: Turning a Blazor Issue into a PR

    GitHub's brand-new Copilot desktop app, in technical preview, handled a small Blazor issue from planning through pull request creation, but the hands-on test also showed why developers still need to verify agent work in the running app before merging.

  • At Build 2026, Microsoft Sets Up Windows as an OS for AI Agents

    Microsoft's Build 2026 Windows developer announcements point to a broader platform strategy for agentic AI, spanning terminal workflows, local models, app-building skills, Cloud PCs and operating system-level containment.

  • Slammed by Copilot Usage-Based Billing on Day 1, Facing $180 Bill for June

    A journalist using GitHub Copilot Pro details how a broken editorial workflow on day one of usage-based billing led to runaway token consumption, a projected $180 monthly bill, and practical tactics for cutting AI credit burn.

  • AdaBoost.R2 Regression Using C#

    AdaBoost.R2 regression works by building an ensemble of decision trees, training them on reweighted data, and combining their predictions with a weighted median, while also showing how parameter choices affect accuracy and overfitting.

Subscribe on YouTube