T-SQL Tuesday is a monthly blogging event in the SQL Server community, started by Adam Machanic in 2009 and now maintained by Steve Jones. Each month, a different community member hosts the event by proposing a topic, inviting SQL Server professionals and enthusiasts to share their experiences, knowledge, and thoughts through blog posts.
It’s a great way to learn from others, discover new voices in the community, and keep the spirit of knowledge sharing alive.
This month’s topic is: Are blogs still relevant?
It’s hard to tell whether people in the technology community still read blogs as much as they used to. With AI tools becoming more accessible, someone can simply ask ChatGPT, Claude, or another AI assistant a question and get an answer in seconds.
Why spend time searching through blog posts when you can just ask AI?
If we’re talking about popularity, I don’t think blogs have the same presence they had back in the 2010s. There are so many other ways to consume information now. YouTube, podcasts, social media, newsletters, and AI assistants have all changed how we learn and discover things.
But if the question is whether blogs are still relevant, I would say yes.
There’s Still Value in Reading Someone Else’s Experience
One thing I appreciate about technical blogs is that they often come from someone who actually encountered a problem, spent time figuring it out, and decided to share what they learned.
Take SQL Server, for example.
You can ask AI how to troubleshoot a slow-running query, how to interpret wait statistics, or how to configure a particular SQL Server feature. It will probably give you a reasonable starting point.
But there’s a difference between getting an answer and understanding how someone arrived at that answer.
A blog post can explain what the author tried, what worked, what didn’t, and why a particular approach made sense in their environment. Sometimes, the most useful part isn’t even the final solution. It’s the troubleshooting process that got them there.
And that’s something I still enjoy reading.
I’ve come across plenty of situations where I thought I understood a particular SQL Server behavior, only to discover something different when I started testing it myself. Reading about someone else’s experience can give you a different perspective and sometimes save you hours of investigation.
Of course, not every blog post is accurate, and not every solution applies to every environment. That’s part of being a technology professional. We still need to validate what we read and understand how it applies to our own circumstances.
AI Has Changed How We Write
Here’s another thing I’ve been thinking about.
I keep writing blog posts whenever time permits. And honestly, there are fewer excuses now for not putting something out there.
One of the hurdles for many people who want to blog is writing itself. They might know the technology, understand a particular problem, and have something valuable to share, but putting those thoughts into a well-written article can be difficult.
Grammar, sentence construction, organization, and simply finding the right words can get in the way.
I’ve been there myself.
Now we have AI tools that can help organize our thoughts, improve sentence construction, check grammar, and make an article easier to read.
That doesn’t mean we should let AI do all the thinking for us. We still need to bring our own experience, technical knowledge, research, and judgment to the table.
I’ve started using AI in my own writing and technical experiments.
For example, in my recent post about When Do You Reach For A #temp Table?, I used Claude Code to help generate test scenarios. That saved me time setting up the experiments so I could concentrate on examining the behavior and explaining what I found.
AI helped with the process, but I still needed to understand the SQL Server behavior, review the results, and decide what was worth sharing.
I think that’s a pretty good way to use these tools.
If you know something worth sharing but struggle to put it into words, there’s no reason to let grammar or writing skills stop you. Let AI help you communicate what you know.
We Should Be Seeing More Quality Technical Content
If AI makes writing easier, shouldn’t we be seeing more blog posts from the technology community?
I think so.
There are plenty of SQL Server professionals out there with years of experience. They have dealt with production outages, performance problems, migrations, database corruption, security issues, and all sorts of situations that don’t always have a straightforward answer.
Those experiences are worth sharing.
Imagine how much more we could learn if more people took the time to document what they encountered and what they learned along the way.
We don’t all have to write comprehensive technical guides. Sometimes a short post explaining a particular problem, a useful script, a troubleshooting tip, or a lesson learned from a production incident is already valuable.
And you don’t have to be a well-known name in the community to have something worth sharing.
Someone might be troubleshooting the exact same problem months or even years from now. Your post could be the one that helps them move forward.
That’s one of the things I like about blogging. A post can continue to be useful long after you publish it.
So, Are Blogs Still Relevant?
For me, absolutely.
Are blogs as popular as they were in the 2010s? Probably not. Do I expect people to spend hours browsing through technical blogs when they can ask an AI assistant for an answer? Probably not either.
But I don’t think those things make blogging any less valuable.
AI is another tool we can use to learn, research, and communicate. It can make finding information and putting our thoughts into words easier. But there’s still value in reading about the experiences and discoveries of people who work with the technology every day.
As for me, I’ll keep writing whenever time permits. I don’t expect every post to get hundreds of views or generate a lot of discussion. If someone finds an article useful, learns something new, or approaches a problem differently because of something I wrote, that’s already a good reason to keep doing it.
And if AI can help me spend less time worrying about grammar and more time sharing what I’ve learned, I’ll gladly take that help.
Now, I’d like to hear from the rest of the community.
Are you still reading technical blogs? Have you changed how you learn because of AI? And if you’ve stopped blogging, what’s holding you back?
Organizations join the cloud native community to strengthen security, visibility, and control for production AI
Key Highlights
CNCF welcomes 12 new Silver members, expanding its community across sovereign cloud, AI agent infrastructure, security and financial services.
As organizations move AI agents into production, new members bring expertise in open infrastructure that supports data sovereignty, operational visibility, security and control.
Enterprises, developers and platform teams gain access to a broader ecosystem of cloud native technologies and expertise to securely build, deploy and manage AI workloads across environments.
KUBECON + CLOUDNATIVECON NORTH AMERICA, SALT LAKE CITY – November 10, 2026 –The Cloud Native Computing Foundation® (CNCF®), which builds sustainable ecosystems for cloud native software, today announced the addition of twelve new Silver members, reinforcing the community’s focus on open, vendor-neutral infrastructure for running AI agents in production. This cohort, spanning sovereign cloud and hosting, AI agent infrastructure, container security, and financial services, broadens the ecosystem of organizations building and operating production AI.
“Deploying AI agents at scale requires the same core guarantees the cloud native ecosystem has spent a decade building: choice, portability, and the ability to control your own infrastructure,” said Jonathan Bryce, executive director, CNCF. “From sovereign cloud operators to specialized security and telemetry maintainers, this group of new silver members highlights how the community is expanding to support complex, autonomous AI workloads everywhere.”
As organizations move AI agents from pilot to production, platform teams face rising demands for data control, operational visibility, and software supply chain trust. CNCF’s newest members reflect that shift, bringing domain expertise and technology built to run agentic workloads on open, portable infrastructure.
New Silver Members
The following organizations have recently joined CNCF as Silver Members:
Centron is a German cloud provider on a mission to make the cloud as simple and accessible as possible – for everyone. Its ccloud³ platform offers managed Kubernetes, virtual machines, S3 object storage, and GPU compute, operated entirely from centron’s own ISO 27001- and C5-certified, GDPR-compliant data center in Germany.
CIQ is the founding support and services partner for Rocky Linux and a leading provider of Enterprise Linux infrastructure. CIQ delivers secure, performant software infrastructure for the full range of modern workloads. Global enterprises, government agencies, research institutions and supercomputing centers trust CIQ to run their most advanced workloads in production, from mission-critical environments to the systems behind major scientific and AI breakthroughs. CIQ’s products include the Rocky Linux from CIQ (RLC Pro) family of operating systems, CIQ Stratum, an open source Kubernetes engine, Ascender Pro for IT automation, Fuzzball job-based container orchestration, Warewulf Pro cluster provisioning, and Apptainer, the container system built for high-performance computing, with Apptainer Pro coming soon.
Cloudpunks GmbH is a cloud native IT consultancy from Germany, specializing in Kubernetes platform engineering, DevOps automation, and cloud infrastructure on AWS and Azure — covering both public cloud and on-premises environments. As a Kubernetes Certified Service Provider (KCSP) and certified AWS and Microsoft Azure partner with Kubestronaut-certified engineers, cloudpunks works embedded in customer teams to build and operate cloud native platforms, not just advise from the sideline.
Emphere fixes vulnerabilities in container images at the source, inside the Dockerfiles teams already maintain. Based in Seattle, the company works to close the gap between CVE disclosure and a deployed fix across base OS packages, language runtimes, and open source libraries.
FluidCloud Limited is a Cloud Cloning platform that uses proprietary Large Language Model to enable discovery, migration, and operations of entire cloud infrastructure in minutes. FluidCloud’s multi-cloud DevOps platform works across AWS, Azure, Oracle, Google Cloud, VMware, Nutanix, Vultr, and OVH to create a universally portable graph of the entire infrastructure, powering instant disaster recovery, resiliency, migration, and multi-cloud interoperability.
ImageEngine Kube is a cloud native, self-hosted image optimization engine that deploys to your Kubernetes cluster via a signed Helm chart. Working alongside your existing CDN, it delivers device-aware resizing, AVIF/WebP/JPEG XL conversion, and intelligent compression—giving platform teams faster image delivery, full control over their data, security auditability, and predictable costs.
Ilert GmbH is an incident management, on-call alerting, and status page platform with a built-in AI SRE, made and hosted in Germany.
Paper Compute is an open intelligence platform for teams building AI agents. Its agent gateway handles model routing, inference policy, and zero-instrumentation telemetry with tapes, deployed in your own cloud.
Radar is an open source Kubernetes interface for humans and AI agents that shows how everything in a live cluster connects, not just what’s in it. Teams use it through a UI built for people and an MCP server built for agents, so both can understand a cluster the way it actually works.
Open-Xchange AG(OX) is the world’s largest independent email provider, and its portfolio includes ubiquitous open source software like Dovecot and PowerDNS, alongside OX App Suite and OX Cloud. OX’s philosophy is clear: keep the internet free, open, and competitive. OX achieves this by building innovative, secure, and open solutions that empower its community, partners, and customers to offer trusted alternatives to global hyperscalers.
Varnish Software AB accelerates and secures the delivery of content, data, and software, from web and streaming traffic to the packages and container images that power software supply chains.
About the Newest Silver End User Member
CNCF also welcomes its newest Silver End User Member, an organization that actively uses cloud native technologies to drive digital transformation:
Deutsche Kreditbank AG(DKB) is one of Germany’s leading digital banks, serving millions of customers as well as businesses and institutions in future-oriented industries across the country.
Cloud native computing empowers organizations to build and run scalable applications with an open source software stack in public, private, and hybrid clouds. The Cloud Native Computing Foundation (CNCF) hosts critical components of the global technology infrastructure, including Kubernetes, Prometheus, and Envoy. CNCF brings together the industry’s top developers, end users, and vendors and runs the largest open source developer conferences in the world. Supported by nearly 800 members, including the world’s largest cloud computing and software companies, as well as over 200 innovative startups, CNCF is part of the nonprofit Linux Foundation. For more information, please visit www.cncf.io.
New research released finds 75% of cloud native developers use AI coding tools, compared to 55% of non-cloud native developer peers
Key Highlights:
CNCF and SlashData released the State of Cloud Native Development Q3 2026 report during KubeCon + CloudNativeCon North America 2026.
The report notes that the global cloud native developer population reached an estimated 21.2 million, representing approximately 41% of developers worldwide.
Cloud native developers are significantly more likely to use AI coding tools than their non-cloud native counterparts (75% vs. 55%) and integrate AI into their services (67% vs. 38%), reinforcing the relationship between cloud native maturity and AI adoption.
The report finds North America is now home to approximately 4 million cloud native developers.
SALT LAKE CITY – KubeCon + CloudNativeCon North America – Nov. 10, 2026 – The Cloud Native Computing Foundation® (CNCF®), which builds sustainable ecosystems for cloud native software, today released new findings from its State of Cloud Native Development Q3 2026 report, developed in collaboration with SlashData. As AI adoption accelerates, the report findings highlight the growing relationship between cloud native infrastructure, platform engineering and the development of production-ready AI and agentic systems.
“Cloud native is foundational to how the industry builds and operates modern software, and AI is accelerating that evolution,” said Chris Aniszczyk, CTO, CNCF. “This report reinforces how cloud native technologies are supporting AI workloads. As organizations move to scale AI workloads, the need for reliable, scalable and observable infrastructure will grow. Open source technologies like Kubernetes and OpenTelemetry are helping provide the foundation for that transition.”
Cloud native developers lead AI adoption as agentic systems move toward production
As AI becomes deeply integrated into software development, the report finds that cloud native developers are significantly more likely to adopt AI tools and integrate AI capabilities into their applications. However, the distinction is less about how developers use AI and more about whether they adopt it in the first place.
75% of cloud native developers use AI coding tools, compared to 55% of non-cloud native developers.
67% of cloud native developers are integrating AI into their services, compared to 38% of non-cloud native developers.
49% of developers integrating AI into their services report having agentic systems in production, including internal and external-facing applications.
55% of developers building agentic services are cloud native, while the remaining 45% demonstrate that developers can integrate agentic capabilities without directly managing cloud native infrastructure.
These findings suggest that cloud native developers are embracing AI while maintaining established engineering practices around oversight, reliability and control.
In North America, this preference for supervised AI development is pronounced. Among cloud native developers in the region, 21% use AI to draft code before reviewing and editing it themselves, while just 12% report using autonomous agents for coding tasks.
“Our research shows that cloud native maturity is becoming an important indicator of how actively developers engage with AI,” said Liam Bollmann-Dodd, principal market research consultant at SlashData. “Cloud native developers are more likely to adopt AI tools and integrate AI into their services, but once developers begin using these technologies, their approaches are remarkably similar. This highlights the importance of cloud native infrastructure in supporting AI at scale, even when the developers building AI-powered applications aren’t directly interacting with the underlying infrastructure.”
Global cloud native developer community surpasses 21 million
The report estimates that the global cloud native developer population reached 21.2 million in Q3 2026, up from 15.6 million in Q3 2025 and 19.9 million in Q1 2026. Cloud native developers now account for approximately 41% of the global developer population, reflecting the continued expansion of cloud native technologies beyond traditional backend infrastructure and into a wider range of development disciplines.
Key findings include:
21.2 million developers worldwide are cloud native, marking the first time the population has surpassed 20 million.
North America is home to approximately 4 million cloud native developers, reinforcing the region’s vital role in the global cloud native ecosystem.
More than 80% of professional backend services and web backend developers engage in cloud native practices and approaches.
Approximately 7.5 million AI developers are cloud native, underscoring the growing intersection between AI development and cloud native technologies.
While the pace of cloud native developer growth has moderated over the past six months, the report suggests that increased adoption of managed services and internal developer platforms may be making underlying cloud native technologies less visible to developers. As organizations standardize their infrastructure, developers may rely on Kubernetes and other cloud native technologies without directly interacting with or identifying the underlying platforms.
Platform engineering becomes standard as AI infrastructure matures
Platform engineering and DevOps standardization continue to reshape how organizations build, deploy and manage applications. The report finds that 93% of developers now work at organizations using some form of DevOps standardization or internal developer platform, up from 77% in Q3 2025.
This shift reflects the growing maturity of cloud native infrastructure as organizations centralize tooling, improve developer productivity and streamline operations. At the same time, greater abstraction can make the underlying infrastructure less visible to developers, even as its importance to production applications increases.
The report also examines the adoption of infrastructure technologies among backend developers:
40% use API gateways, making them the most widely adopted technology surveyed.
30% use microservices, reflecting their continued importance in modern application architectures.
21% use observability tools, highlighting the role of visibility into complex systems.
19% report using Kubernetes, although platform abstraction may contribute to underreporting of underlying infrastructure technologies.
8% use immutable infrastructure practices and 7% use chaos engineering, suggesting continued opportunities for advanced infrastructure practices to mature.
As AI workloads grow in complexity, these underlying capabilities are highly relevant to production environments. AI inference and agentic systems introduce new demands around compute scheduling, resource efficiency, observability and reliable distributed operations, areas where cloud native technologies already provide established approaches.
The report also identifies cost, data privacy and security as leading concerns among developers who have not yet integrated AI functionality into their services. As a result, infrastructure capable of supporting reliable operations, cost management and scalability will likely become more important.
The findings reinforce cloud native’s expanding role in the AI ecosystem, not only for developers building and operating infrastructure directly, but also for the growing number of developers consuming managed AI services built on cloud native foundations.
Cloud native computing empowers organizations to build and run scalable applications with an open source software stack in public, private and hybrid clouds. The Cloud Native Computing Foundation (CNCF) hosts critical components of the global technology infrastructure, including Kubernetes, Prometheus and Envoy. CNCF brings together the industry’s top developers, end users and vendors and runs the largest open source developer conferences in the world. Supported by nearly 800 members, including the world’s largest cloud computing and software companies, as well as over 200 innovative startups, CNCF is part of the nonprofit Linux Foundation. For more information, please visit www.cncf.io.
About SlashData
SlashData helps product, platform, and engineering leaders understand how software development is changing in the AI era. For over 20 years, they have helped many of the world’s leading technology companies make sense of developers, developer ecosystems, and the tools and workflows shaping the future of software. Their work shows clients how the developer ecosystem and software engineering teams are evolving, so they can build better products, platforms, and go-to-market strategies for the AI era and make smarter decisions about the tools, practices, and investments ahead.
What happens when a piece of mail or a package doesn't arrive when and where it's supposed to? In this episode of Mailin’ It!, hosts Karla and Jeff explore the fascinating world of the USPS Mail Recovery Center, where dedicated employees work to reunite undeliverable mail with its intended recipients or senders.
Joining the conversation is the Executive Manager of the Mail Recovery Center, who explains their mission to “complete the delivery cycle.” From damaged labels and incomplete addresses to packages separated from their contents, the team carefully examines items, researches clues, and uses specialized technology to identify their rightful owners.
The discussion also explores the center’s history, tracing its origins to the Dead Letter Office and the early days of the Postal Service. They explain how the recovery process has evolved, including computerized matching systems, missing mail search requests, and auctions for items that cannot be reunited with their owners.
Listeners will hear remarkable recovery stories, including a package containing valuable items that was successfully delivered after being missing for 20 years. They also share practical advice for preventing undeliverable mail, such as using clear addresses, including contact information inside packages, and choosing appropriate tracking and insurance options.
Above all, this episode highlights the dedication and investigative work of the Mail Recovery Center’s employees, whose mission is to complete the delivery cycle and give missing mail another chance to reach its destination.
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
The November 10, 2026 servicing updates for .NET 10, .NET 9, and .NET 8 include security and non-security fixes. This post brings together the security advisories, release notes, and download links for this month’s updates, along with resources for .NET Framework servicing.
Security improvements
The following table covers vulnerabilities affecting .NET 8, 9, 10, and 11, including associated tooling. Follow each advisory for the full list of affected products and remediation guidance.
Use the download links above to update your .NET SDK and runtimes. Review the release notes and security advisories for guidance relevant to your applications.
Today, we’re excited to announce the release of .NET 11! This year’s release makes building modern applications faster, more productive, and has improvements across C# 15, ASP.NET Core, .NET MAUI, and Windows development. With hundreds of runtime optimizations, more responsive SDK tooling, and less overhead in everyday commands, you can spend less time waiting and more time creating. Build richer web and cross-platform applications, take advantage of new C# language features, and explore AI-powered development with improved tools and agentic capabilities.
.NET 11 is a Standard Term Support (STS) release with two years of support under the .NET support policy. STS and Long Term Support (LTS) releases have the same quality and servicing policies; the difference is the length of support. .NET 10 remains a supported LTS option. Stay current with servicing updates whichever release you choose.
Join us at .NET Conf, November 10-13
Explore .NET 11 with the product teams and community at .NET Conf 2026. The conference includes product team sessions on November 10 and 11, followed by two days of community sessions.
Some of the changes we’re most excited for you to try:
C# 15 union types and closed hierarchies help you express the possible states of your application and handle them with compiler-checked patterns.
A Native AOT CLI host and a warm MSBuild server reduce overhead in everyday developer workflows.
ASP.NET Core brings asynchronous validation, OpenAPI 3.2, and new capabilities for Blazor and SignalR applications. Experimental Blazor AI components integrate with the Agent User Interaction Protocol (AG-UI) for advanced agentic web experiences.
Build AI-powered, agentic applications with Microsoft.Extensions.AI (MEAI) for model access, MCP for tool integration, and Microsoft Agent Framework (MAF) for agents and workflows.
Put agentic development tools to work with .NET Skills for .NET-specific guidance, .NET MAUI DevFlow for inspecting and testing running apps, and the .NET Upgrade canvas for reviewing modernization progress.
.NET MAUI adopts CoreCLR as its default runtime and improves XAML iteration and native controls.
The libraries add built-in Zstandard compression, simpler process execution, richer JSON support, and new ways to work with in-memory data without copying it first.
The what’s new in .NET 11 overview links to the full details. Here are the changes behind those headlines and how to start using them.
Languages
C# 15
C# 15 is the default language version for projects targeting .NET 11. Union types are one of its biggest additions: a value can be one of a declared set of case types, with conversions from those types and exhaustiveness checking in switch expressions.
Consider a product lookup that either finds a product or reports that it wasn’t found:
ProductLookup result = new Product(42, "Mechanical keyboard");
string message = result switch
{
Product product => $"Found {product.Name}",
ProductNotFound missing => $"No product with ID {missing.Id}",
};
Console.WriteLine(message);
public sealed record Product(int Id, string Name);
public sealed record ProductNotFound(int Id);
public union ProductLookup(Product, ProductNotFound);
The result type tells a caller which outcomes it needs to handle. There’s no need to use object or an unrelated base class just to put those outcomes behind one return type.
C# 15 also adds closed hierarchies. A closed class limits direct derivation to its declaring assembly, giving the compiler the information it needs to check patterns over the hierarchy. Unions are useful when you want to combine existing types; closed hierarchies fit models where you own the inheritance relationship.
Smaller language features help with everyday code, too. Collection expression arguments let you pass constructor or factory arguments without giving up collection expression syntax:
This prints 2: the comparer controls equality, so "Alice" and "ALICE" are one entry.
Other additions include extension indexers and labeled break and continue. These features don’t require the preview language version. The separate memory safety work remains in preview; see the C# 15 feature documentation for its opt-in requirements.
F# 11
F# 11 improves compiler memory use and reduces allocations in FSharp.Core collection operations. It also adds ways to bound concurrent asynchronous work, useful when a collection operation would otherwise start more work than a service or machine can handle comfortably.
The F# release notes distinguish the default-language updates from preview capabilities, including Runtime Async support for custom computation-expression builders.
Platform
SDK
The SDK now enables a Native AOT CLI host by default. Commands such as help and version queries can run without starting the managed CLI, and tool launches can avoid that startup path too. Commands that need managed execution still use it.
The MSBuild server also stays warm between CLI invocations by default, reducing repeated startup work. SDK installers on Linux and macOS are smaller through assembly deduplication.
File-based apps can grow across source files with #:include, and solution filters can be created and edited from the CLI. dotnet watch adds Aspire app-host integration, crash recovery, and device selection for mobile projects.
For teams using Microsoft.Testing.Platform, dotnet test gains run-level timeouts, failure limits, clearer progress reporting, and per-module results directories. The test runner matters: these switches aren’t interchangeable with the VSTest command-line options.
Container publishing adds reproducible image digests through SOURCE_DATE_EPOCH and avoids redundant layer uploads when the target manifest already exists. NuGet source configuration gains a minimum package publication age setting, which lets teams wait before newly published package versions become eligible for selection. This is an additional policy control, not a replacement for vulnerability checks or reviewing dependencies.
Performance work in .NET 11 reaches both the code you deploy and the tools you use to develop it. Stephen Toub’s Performance Improvements in .NET 11 takes a detailed tour through hundreds of runtime and library optimizations, with benchmarks and code examples.
The just-in-time (JIT) compiler improvements remove more redundant bounds checks and branches, improve devirtualization, and fold more operations into constants. For example, the JIT can optimize additional SequenceEqual calls when it knows the inputs at compile time. These changes improve generated code without requiring you to rewrite the corresponding application logic.
Native AOT interface dispatch uses a shared helper that reduces the code needed at interface call sites and improves throughput for interface-heavy workloads. Hardware support also expands, including FP16 instructions for Half operations on supported x64 and Arm64 hardware and new vector construction and composition APIs.
There’s useful work on diagnostics, too. In-process crash report logging can capture a managed stack trace, module information, and runtime state before a process exits. Previously available on mobile platforms, this reporting path now also supports Linux and macOS.
As always, measure the effect on your own application. Throughput, startup time, memory use, and deployment size depend on the workload and publishing options you choose.
Runtime Async has been a recurring theme throughout the .NET 11 previews. It moves asynchronous suspension and resumption into the runtime instead of relying on compiler-generated state machines.
One especially useful result is a clearer live call stack. Profilers and debuggers can show the asynchronous methods themselves instead of the infrastructure used to execute them. This is a distinction from exception stack traces, which already hide much of that infrastructure.
Runtime Async remains a preview feature. To try it in a project targeting net11.0, add:
You don’t need to enable EnablePreviewFeatures for this scenario. The .NET runtime libraries themselves are built with Runtime Async, but opting your application’s code into the new compilation model is a separate choice. See the Runtime Async documentation for its behavior and current guidance, and test your application’s dependencies and diagnostics before adopting it.
AI libraries and tools
The .NET 11 SDK includes an MCP server template. MCP gives AI applications a common way to discover and call tools and services. The template uses the official MCP C# SDK and includes a working sample tool, so you can start exposing your own functionality without first installing a separate template package:
dotnet new mcpserver -o MyMcpServer
For applications that consume AI services, Microsoft.Extensions.AI (MEAI) provides abstractions such as IChatClient and IEmbeddingGenerator. These libraries let you compose model access with dependency injection, caching, and OpenTelemetry using familiar .NET patterns. They’re available as NuGet packages, rather than features that require you to move all of your application code to a new framework version.
Microsoft Agent Framework (MAF) builds on these abstractions with agents that call tools and maintain conversation state, plus workflows that coordinate agents and application code. Connect agents to MCP tools to give them access to your application’s capabilities, and expose an agent through an ASP.NET Core AG-UI endpoint to build an interactive Blazor experience. The AG-UI hosting integration is currently prerelease.
The .NET libraries add capabilities that previously required extra setup, intermediate buffers, or custom helpers.
The Process API expansion includes one-shot methods to run a process and capture its output and exit status. More advanced APIs give you control over inherited handles, suspended starts, and process signaling. Build tools and developer utilities can spend less code on process plumbing while still inspecting how a child process finished. See Process API Improvements in .NET 11 for examples.
Zstandard compression joins the existing compression formats in System.IO.Compression. New span-based compression APIs support callers that manage their own buffers, and ZIP improvements include CRC32 validation when reading entries.
Four new in-memory stream adapters bridge existing data to APIs that accept a Stream: ReadOnlyMemoryStream, WritableMemoryStream, ReadOnlySequenceStream, and StringStream. For example, a pipeline’s segmented ReadOnlySequence<byte> can be read as a stream without first flattening it into a new contiguous buffer.
System.Text.Json adds support for C# unions, F# discriminated unions, and JSON Lines output from asynchronous sequences. F# union support uses reflection and doesn’t support source generation, trimming, or Native AOT, so check the documented limits before choosing it for a trimmed application.
LINQ adds FullJoin, which preserves unmatched elements from both inputs. This is useful for reconciliation work, such as comparing two inventories where you need to see missing items on either side.
The libraries also expand numeric types, DNS resolution, cryptography, and tracing configuration. The library overview covers the complete set of changes and marks APIs that remain experimental.
Entity Framework Core 11
Entity Framework Core 11 improves how your model becomes a query and how that query becomes SQL.
EF Core removes unnecessary joins in more reference-navigation and split-query scenarios, improves GroupBy translation, and translates the new LINQ FullJoin operator on relational providers that support it. These changes can simplify generated SQL while letting more queries stay in LINQ.
Complex types and JSON columns can now be used on entities mapped with table-per-type or table-per-concrete-type inheritance. Other improvements include migrations tooling and additional provider-specific JSON capabilities.
For SQL Server applications, EF Core adds support for vector indexes and approximate vector search. The SQL Server features and corresponding EF APIs are experimental. See the vector search documentation and limitations before evaluating them.
EF Core 11 requires .NET 11. When upgrading, check your database provider’s compatibility and the EF Core 11 breaking changes as well as the runtime changes.
Developer tools
The new dotnetup toolchain manager lets you install and update .NET SDKs and runtimes in your user account. It’s currently in public preview and helps you manage side-by-side versions and keep up with servicing updates.
.NET Skills gives coding agents task-specific guidance from the .NET team. Our .NET Skills introduction explains how to add these skills to your coding agent.
The experimental .NET MAUI DevFlow toolkit exposes a running app’s visual tree, screenshots, and interaction tools, so an agent can inspect and test the UI rather than work only from source code.
For modernization, the .NET Upgrade canvas gives you a visual view of a GitHub Copilot upgrade session. Review assessment findings and dependency compatibility, follow implementation tasks, and monitor failures and progress as the upgrade agent works. Our modernization walkthrough in the GitHub Copilot app covers the workflow from assessment through execution.
Workloads
Web: ASP.NET Core and Blazor
Application types are most useful when the rest of the stack understands them. ASP.NET Core supports C# unions wherever it uses System.Text.Json, including JSON request and response bodies in Minimal APIs and MVC. Generated OpenAPI documents describe the union’s alternatives with an anyOf schema.
This support applies to JSON bodies, not to route, query, header, or form binding. When multiple union cases have the same JSON shape, configure a classifier so deserialization can distinguish them. The ASP.NET Core union documentation explains those boundaries.
Validation also becomes more practical for rules that need to do I/O. Asynchronous validation works with the new asynchronous DataAnnotations APIs, so a validator can await a database lookup or a remote call instead of blocking a thread. Minimal APIs and Blazor forms can use this support.
OpenAPI 3.2 is now the default for generated documents. Among other improvements, it lets the document describe the per-event payload of a server-sent event stream. If a client generator or another part of your toolchain doesn’t yet support 3.2, you can explicitly select an earlier document version.
For long-lived connections, SignalR authentication refresh lets a connection refresh authentication without reconnecting. Blazor Interactive Server components can receive the updated identity and react when claims change.
Blazor adds capabilities for both statically rendered pages and interactive applications.
The new CacheView component caches the rendered output of a component subtree during static server-side rendering. On a cache hit, Blazor replays the markup without creating those child components or running their lifecycle methods. That gives you a smaller unit of caching for expensive, mostly static sections of a page.
Interactive Server apps can opt into automatic circuit pause when a browser tab becomes inactive. After a configurable delay, pausing frees the server memory and SignalR connection held by an inactive user. The feature requires an additional package and explicit configuration; upgrading doesn’t silently change your application’s circuit behavior.
Forms gain client-side validation for static server rendering and asynchronous validation support. QuickGrid and Virtualize add ways to move to a particular item, useful when returning a user to their place in a large collection.
Blazor also adds components for agentic user interfaces. The new Microsoft.AspNetCore.Components.AI package includes components for streaming chat, tool rendering, approval flows, and shared UI state. Basic chat works with any IChatClient. For advanced experiences with a remote agent, the AG-UI .NET integration carries streamed responses, tool events, approval requests, and shared-state updates between an ASP.NET Core agent endpoint and the Blazor UI.
The Blazor AI components package is experimental and remains prerelease throughout .NET 11. It’s an opportunity to try the direction and give feedback, not a stable part of the ASP.NET Core shared framework. See the Blazor AI components documentation.
See what’s new in ASP.NET Core for the Blazor features, configuration examples, and compatibility details.
Client: .NET MAUI
.NET MAUI in .NET 11 makes CoreCLR the default runtime across its platforms. The move to CoreCLR brings mobile and desktop apps onto the same runtime foundation as other .NET workloads. On Apple platforms that don’t allow JIT compilation, the runtime uses precompiled code and the interpreter rather than generating native code at runtime.
Native AOT is an opt-in publishing alternative; its Android support remains experimental. Review the platform-specific runtime guidance and test your app’s libraries when choosing a deployment model.
The UI development loop gets attention, too:
XAML Incremental Hot Reload is a preview feature in .NET 11. It updates live instances of pages and controls, including supported structural edits, without requiring you to navigate away and back. It’s enabled by default in Debug builds.
Implicit XAML namespace declarations remove the need to repeat the standard namespace declarations at the root of every file.
Windows uses the new CollectionView2 handler by default, while Android Shell adopts a handler-based architecture.
New mobile test templates and dotnet test device selection bring testing into the CLI workflow.
These are meaningful runtime and handler changes, so exercise navigation, custom handlers, and trimmed publishing as part of your upgrade. The optional Microsoft.Maui.Controls.Compatibility package no longer ships; apps that still reference it need to migrate off it.
.NET MAUI follows its own support policy, separate from the .NET runtime’s STS and LTS lifecycle.
Windows: WinUI, Windows Forms, and WPF
For new Windows desktop apps, WinUI 3 is the recommended native UI framework, pairing C# and XAML with modern controls and the Fluent Design System. It’s delivered through the Windows App SDK, which also makes modern Windows APIs available to existing WPF and Windows Forms apps.
Windows Forms in .NET 11 adds an opt-in modern visual styles mode that you can select at the application or control level. It also adds ways to react to system visual and accessibility settings, defer revealing a form until it’s ready, and suspend painting during bulk updates.
These controls let you introduce refreshed rendering where it makes sense for your application instead of treating a framework upgrade as an automatic visual redesign.
WPF updates focus on reliability, including fixes for Fluent window backdrops, window manager failures, and weak-event cleanup. Self-contained Windows publishing also includes a native dependency needed by apps that load mixed-mode WPF assemblies.
dotnet new console -n HelloDotnet11 -f net11.0
dotnet run --project HelloDotnet11
You can replace the generated program with either of the C# examples above to try the new language features.
For an existing application, plan the upgrade across your development and deployment environments:
Update the SDK selection in global.json if your repository pins it, and install the matching SDK on development machines and build agents.
Update target frameworks to .NET 11, preserving platform-specific target framework suffixes where applicable. Update packages and workloads to compatible versions.
Run your tests and measure the application under a representative workload. Validate your actual publishing mode, container base images, native dependencies, and deployment hosts.
Upgrading from .NET 8 or .NET 9
.NET 8 and .NET 9 reach end of support on November 10, 2026. Move to a supported release to continue receiving fixes. Choose .NET 11 for the latest capabilities, or .NET 10 if its LTS lifecycle fits your requirements better. See the .NET support policy for dates and servicing requirements.
If you’re upgrading from a .NET 11 release candidate, update to the final SDK, runtime, and compatible package and workload versions rather than staying on a prerelease build.
Thank you for building .NET with us
.NET 11 is the work of over 7,000 contributors. Thank you to everyone who contributed code, tested previews, reported issues, and helped us make better design decisions throughout the year. Your feedback across the previews and release candidates helped shape what we’re shipping today.
The .NET ecosystem continues to grow, with more than 589,000 unique packages on NuGet and over 1.1 trillion package downloads.
Download .NET 11, try it with an application you know well, and join us at .NET Conf to learn directly from the people building it. Share your results and questions in the .NET GitHub discussions.
Log in
