Blog
Philip Fulcher
December 22, 2025

Wrapping Up 2025

Wrapping Up 2025

In 2025, we had a clear goal: help teams ship at AI velocity without the typical development interruptions. AI tools now let developers write more code than ever before, but what's the point if you can't actually get it merged and deployed?

This year, Nx grew into a Build Intelligence Platform combining build systems, CI orchestration, and AI integration all in one place — eliminating the things that most get in the way of pushing code to production. Self-Healing CI, enterprise analytics, and AI-powered workflows transformed how teams interact with their codebase. All this while keeping up with the latest tools and upgrades.

Unlocking AI Potential with Full Workspace Awareness

AI assistants are transforming development, but they're only as useful as the context you give them. LLMs need context from source code, CI pipeline, and terminal to unlock their full potential.

Ask an LLM to help with a bug in your monorepo, and it's working blind — it doesn't know your project structure, your build configuration, or your CI history.

We shipped the Nx MCP server to change this. It gives LLMs deep access to your workspace structure, project dependencies, and task configurations. When you ask for help, it can actually help, because the AI understands your entire codebase.

The terminal integration extends this context to include real-time terminal output. AI assistants can see errors as they happen and suggest fixes immediately.

Shipping at AI Velocity: The Self-Healing Platform

AI assistants help developers write code faster, but broken CI creates a bottleneck. When a developer pushes code, they context-switch to another task. Twenty minutes later, CI fails. Now they have to context-switch back, figure out what went wrong, fix it, push again, and wait. This cycle kills the productivity gains that AI promises. It kills the flow state.

Self-Healing CI eliminates this interruption. When a PR fails, Nx analyzes the failure, identifies the root cause, and posts a fix suggestion directly in the PR. In many cases, it can automatically commit the fix, letting CI continue without developer intervention. Even in large-scale monorepos, about 60% of these fixes are accepted.

The Nx MCP server takes this further. AI assistants can now read your terminal output in real-time, understanding errors as they happen and suggesting fixes before you even open your browser.

This is only possible because Nx has more context than other tools: your workspace structure, task graph, and its role as an intelligent build platform.

DevEx That Compounds

The daily tools you use should get better over time, and Nx has doubled down on a great developer experience this year.

The Terminal UI introduced in Nx 21 provides an elegant interface for viewing multiple running tasks simultaneously. Instead of scrolling through interleaved logs, you get a clear view of what's running and what needs attention. Continuous tasks enable task pipelines that don't wait for long-running processes. A single command can spin up the backend that your frontend needs to operate.

Nx Console has steadily brought more capabilities, improving the IDE experience for devs. With CI integration in VSCode and JetBrains, you get notified about CI failures right where you're working. One click opens the fix suggestion. Another click applies it. You stay in flow instead of bouncing between your IDE and GitHub. The Migrate UI in Nx Console makes upgrades visual and guided.

Enterprise Observability: Moving from Reactive to Proactive

You can't fix what you can't see. Most teams discover performance problems reactively — someone complains that CI is slow, or a flaky test wastes an afternoon. We shipped our most recent set of enterprise analytics to allow teams to be proactive and spot the root issues.

Task Analytics shows the performance of every task in your workspace over time. You can spot regressions the day they happen, not weeks later when someone finally notices. Sparklines show trends at a glance, making it easy to identify which tasks need attention.

Flaky Task Analytics goes deeper, identifying which flaky tasks cause the most reruns and waste the most time. Instead of treating all flaky tests equally, you can prioritize fixing the ones that actually hurt productivity.

Agent Resource Usage provides process-level visibility into CI resource consumption. When a task suddenly starts using twice as much memory, you'll know immediately and can investigate before it causes pipeline failures.

These analytics aren't just dashboards — they're intelligence that drives automated decisions across the platform.

The impact is measurable. PayFit reduced deployment time from 2-5 days to 20 minutes. Hetzner Cloud cut CI time from 45 minutes to 6 minutes with an 85% cache hit rate. Our benchmarks show the full platform delivers 3-5x productivity improvement compared to DIY cache solutions.

One Workspace, Every Language

Modern applications rarely fit into a single ecosystem. Your React frontend talks to a Java backend. Managing these in separate repositories means coordinating releases across teams and hoping nothing breaks.

Nx now handles many polyglot workspaces natively with support for Gradle, Maven, and .NET. Not only does this enable your team to configure tasks so that one command spins up any backend necessary for your frontend, it also means your AI agents have access to more context. Changes can be made across both frontend and backend all at once, without having to switch between workspaces.

Workspace Intelligence for Multi-Repo Organizations

Not every organization can consolidate everything into a single monorepo. Some teams manage multiple repositories for practical reasons: organizational boundaries, compliance requirements, or gradual migrations.

Conformance automates consistency checks across all your repositories, whether they use Nx or not. Define rules once, enforce them everywhere. Teams get feedback directly in their PRs, making it easy to stay compliant without creating bottlenecks.

The Workspace Graph visualizes dependencies across repos, helping you understand how changes in one repository might affect others. Custom Workflows let you run scheduled tasks beyond traditional CI — security scans, dependency updates, architecture validation.

These Enterprise plan features bring monorepo intelligence to organizations with multiple repositories. Platform teams get control and visibility without forcing everything into one place.

Ecosystem Updates

One of the tasks that Nx takes off your plate is regular updates of essential packages. Our migrate process has kept you up-to-date this year with new versions of major frameworks like Angular, React, and Next.js as we always have. But this year, we also expanded our reach.

  • TypeScript Project References - After introducing this new approach to TypeScript project configuration that greatly speeds up development, we’ve steadily improved it and made it the default configuration for new workspaces.
  • pnpm catalog support - We expanded support for this important pnpm feature, making sure you can manage dependencies throughout your workspace how you’d like.
  • Angular Rspack - We introduced Angular Rspack, maintained by the Nx team, ensuring that teams using Angular have access to the fastest bundlers available.

Coming in 2026

AI agents are already here and effective, but we will continue to enhance their effectiveness in Nx Workspaces. In Q1 2026, we're shipping code mode for better LLM context management, a Claude Code plugin that ships LLM metadata with Nx plugins, and specialized agents for migrations and code generation. We're also investing in larger repository support:

  • subgraph creation so you don't need to install Java when you're only touching JavaScript
  • better native toolchain management for polyglot workspaces
  • strict cache validation that provides hard guarantees that your cache is configured correctly.

Self-healing CI continues to improve as well: we're currently seeing about two-thirds of broken PRs receive effective fixes, and we're working to push that number higher through better dynamic configuration and an improved deflaker.

Thank you

Finally, we want to end with gratitude. Nx would not be where it is today without the support of our community. We recently celebrated our 9th anniversary, and we're looking forward to many more with you. Thank you for all of your contributions at conferences and meetups, your PRs and issues on GitHub, your videos, your articles, and countless other ways that you interact with Nx. Thank you for making this community what it is. We couldn't do it without you, and we're excited to build an even better platform with you in 2026.