React Concurrent Mode - Revisited

Mar 5, 2020

React Concurrent Mode - Revisited

Understanding React Concurrent Mode & The Problems It Solves

Author

Aashish Tiwari
Aashish TiwariSenior Software Engineer - II

Concurrent Mode was announced about a year ago and is still an experimental feature that lets our app remain responsive in different network conditions and based on different device capabilities. Without any further ado, let's discuss the new features and concepts introduced in Concurrent Mode. We will discuss these features in the context of React and React apps:

1. Interruptible Rendering

Normally in React, if rendering begins and a user event occurs (like clicking a button), React will first complete the previous render and then process the user event. So, React uses blocking rendering, which can create some performance issues.

Concurrent Mode introduces the concept of interruptible rendering, so if rendering has started and a user event happens in between, React will stop the previous rendering, process the user event first, then continue the remaining part of the previous rendering. Concurrent Mode uses Time Slicing for interruptible rendering.

2. Data Fetching With Suspense

It works on a Render as you Fetch approach, which means that React starts loading the data without waiting for the component to load. Both the data and component start loading in parallel.

Here, we wrap our MySuspenseImage component in a Suspense component, which provides the fallback component until the image is loaded. We try to render MySuspenseImage even when imageData is not yet loaded. This makes MySuspenseImage suspend as a promise is thrown. The Suspense component catches the promise and shows the fallback component in the meantime. As soon as the promise resolves, it re-renders the image component. If any errors occur during data loading, they can be handled using the ErrorBoundary component.

3. SuspenseList

Suppose a page has multiple API calls. It’s likely their responses won’t come in the order required by the UI, which affects user experience. We can chain promises to maintain order, but that adds complexity.

SuspenseList solves this by providing a controlled loading sequence for different parts of the page. It wraps multiple Suspense components and uses the revealOrder property to control the order of reveal: 'forwards', 'backwards', or 'together'.

We can also control the number of loading indicators using the tail property, which can be 'hidden' or 'collapsed'. When tail is 'collapsed', only one loader is shown at a time.

In the example (shown in the GIF), with revealOrder='forwards' and tail='collapsed', first the user image is rendered, then user details, and finally user comments. This ensures smooth, sequenced UI loading. All async processes begin together, but the reveal order is controlled visually via SuspenseList.

4. useTransition Hook

Suppose you're on a fast connection and click your profile. You may transition instantly and see a loader for just 200ms, causing UI flicker. This can be fixed with complex logic, but the useTransition hook simplifies it.

It lets you wait on the current screen while new data loads, then transition smoothly.

The useTransition hook returns two values:

  • startTransition() – Handles the deferred state update.
  • isPending – A boolean that indicates if the data is still loading.

When a user clicks to navigate, React prepares the update in a separate branch while remaining on the current screen. If data loads quickly, it transitions early; otherwise, it waits for the defined timeout, then transitions. This reduces jarring screen transitions and flickering loaders.

You can see that the loader is shown on the previous screen while userDetails are fetched, then transitions smoothly to the new screen, loading the remaining data.

GitHub project link: https://github.com/Aashishgtbit/concurrent-react-example

References: https://reactjs.org/docs/concurrent-mode-intro.html

Thank you for reading!

Subscribe to Our Newsletter

More from the engineering frontline.

Dive deep into our research and insights on design, development, and the impact of various trends to businesses.
Insight
Why Legacy Systems Make Business Growth More Expensive: Navigate A Smarter Path to Legacy Modernization
Sep 23, 2026

Why Legacy Systems Make Business Growth More Expensive: Navigate A Smarter Path to Legacy Modernization

Learn how legacy systems make business growth more expensive and how edge-first modernization can remove constraints without replacing the existing system.

Insight
SSO, Audit Logs and RBAC: The Enterprise Features AI Prototyping Tools Do Not Cover | Sarika Gautam
Sep 23, 2026

SSO, Audit Logs and RBAC: The Enterprise Features AI Prototyping Tools Do Not Cover | Sarika Gautam

Why AI-generated prototypes fail enterprise review: the context behind SSO, the cost of skipping audit logs, and how role explosion makes RBAC a product of its own.

Insight
Feature Flags as Technical Debt: The Cleanup Nobody Schedules
Sep 21, 2026

Feature Flags as Technical Debt: The Cleanup Nobody Schedules

This blog explains how unmanaged feature flags create technical debt and how teams can detect, manage, and remove them safely.

Insight
Building AI-First Enterprises: Why System Design Matters More Than AI Adoption
Sep 21, 2026

Building AI-First Enterprises: Why System Design Matters More Than AI Adoption

This blog explores how system design, architecture, and validation shape AI-first enterprises, while examining AI’s impact on software engineering and human decision-making.

Insight
The Product Studio in the AI Era: What Actually Changes | Sarika Gautam
Sep 21, 2026

The Product Studio in the AI Era: What Actually Changes | Sarika Gautam

What changes in product development when AI writes the code: the shift to architecture, the token cost of unplanned builds, and why juniors still matter.

Insight
AI and the Future of Digital Customer Experience: Where Technology Meets Human Creativity
Sep 18, 2026

AI and the Future of Digital Customer Experience: Where Technology Meets Human Creativity

A discussion on how AI, human creativity, research, and cross-functional collaboration are shaping the future of digital customer experience.

Insight
Scaling Down Before Scaling Up: Why Bigger Servers Don’t Fix Bad Architecture
Sep 17, 2026

Scaling Down Before Scaling Up: Why Bigger Servers Don’t Fix Bad Architecture

This blog explores how inefficient backend architecture can cause performance issues even under low traffic, covering practical ways to reduce database load, API latency, and resource usage before scaling infrastructure.

Footer

The Right Conversation Can

Save You Six Months.

Book a Call