The honest answer: 'React is faster than Angular' is half-true and depends on what you measure.

What React did differently

Smaller runtime.

• React + ReactDOM: ~45 KB gzipped.
• Angular: framework + RxJS + Zone.js historically larger (Angular 16+ ivy is much leaner, but still bigger than React baseline).

Different change detection.

• Angular (pre-signals): Zone.js patches setTimeout / Promise / events; any async tick triggers a check across the component tree. OnPush opts a component into manual marks.
• React: re-renders only when state changes; the developer explicitly causes updates.

Virtual DOM diff vs template binding checks.

• Angular compiles templates to imperative DOM update functions. Each check walks bindings.
• React diffs an in-memory tree and patches only the differences.

Neither is obviously faster — but React's model is simpler to reason about, which makes optimization more direct.

Less framework runtime per component.

• Angular components carry decorators, DI metadata, change detection bookkeeping.
• React function components are plain functions called by the reconciler.

Tree-shaking.

• React: standard ES modules, leverages bundler tree-shaking aggressively.
• Angular: improved with Ivy, but historically more framework code shipped because of decorator metadata.

Where Angular pushes back

• Built-in everything: routing, forms, HTTP, animations, i18n. With React you're picking N libraries; if you choose badly, the bundle grows.
• AOT compilation: ahead-of-time template compilation makes runtime faster than dynamic JIT-style frameworks.
• Signals (Angular 16+): fine-grained reactivity that bypasses Zone.js — closing the gap on React's change detection model.
• Strict TypeScript first-class: integration is tighter out of the box than React + TS.

What actually drives real-world frontend perf

• Bundle size delivered to the browser — framework runtime is part of this, but app code usually dominates.
• First paint / hydration cost — SSR / streaming matters more than framework choice.
• Render frequency — too many state updates, missing memoization.
• DOM size — both frameworks suffer from huge component trees.
• Network — slow APIs make any framework look slow.

Framework choice is rarely the bottleneck once the app is non-trivial.

Honest framing

React's smaller runtime and simpler change detection model give it an edge by default. But a well-built Angular 17 app with signals and OnPush is comparable. The bigger differences are ecosystem and team productivity, not raw performance. For most apps the perf bottleneck is the developer's choices, not the framework.

A note on benchmarks

JS Framework Benchmark (krausest) shows React and Angular within 10–20% on most operations. Svelte and SolidJS lead. Differences this small rarely matter for product work.