Google is preparing a security-minded change for Chrome on Android that would turn off the WebGPU API when Advanced Protection Mode is enabled, shrinking a high-value attack surface for people at elevated risk. References to a new option labeled “Turn off WebGPU to help protect against security threats” have appeared in recent Google Play Services builds, suggesting the toggle will sit alongside other hardened Chrome settings inside Android’s one-click protection suite.
What Disabling WebGPU Changes for Chrome on Android
WebGPU is the modern graphics and compute API for the web, designed to succeed WebGL with far better access to today’s GPUs. It unlocks high‑performance 3D rendering, video effects, and general‑purpose GPU compute for web apps, from rich editors and games to on‑device machine learning.

On Android, WebGPU has been enabled by default in Chrome 121 and newer on devices running Android 12 or later with Qualcomm or ARM GPUs. That covers a broad slice of current phones and tablets, which is precisely why Google is eyeing it under an opt‑in security mode: the larger the feature’s reach, the more enticing it becomes to attackers.
If WebGPU is switched off, sites that rely on it should fall back to WebGL or CPU paths. Most core browsing remains unaffected, but graphically intensive web apps, WebGPU‑accelerated ML demos, and some game engines could see reduced performance or features until a user exits Advanced Protection Mode.
Why Google Is Targeting The GPU Attack Surface
GPU APIs provide low‑level access to powerful hardware through complex driver stacks and shader compilers. Historically, that complexity has spawned security issues. Chrome’s release notes frequently include fixes in the GPU process and graphics pipeline, and security researchers have demonstrated bugs capable of code execution or sandbox escape when paired with other flaws.
Academic teams have also warned about potential side channels exposed by web‑accessible GPU features, which can leak sensitive signals or amplify timing attacks. While Google’s multi‑process architecture and sandboxing significantly blunt the blast radius, reducing exposure for high‑risk users is a prudent layer of defense‑in‑depth—especially in a browser used by billions.
The stakes are large. Chrome maintains roughly 65% of global browser share according to StatCounter, and Android itself now runs on well over three billion active devices by Google’s own disclosures. With that scale, even niche exploitation paths can be lucrative for well‑resourced adversaries.

Advanced Protection Mode’s Expanding Toolkit
Advanced Protection Mode debuted with Android 16 as a one‑tap way to aggregate the platform’s most stringent safeguards for journalists, activists, executives, and other high‑risk users. Recent code paths indicate Google is tightening more levers under this umbrella, including newly tested limits on the AccessibilityService API unless an app is explicitly classified as an accessibility tool.
The forthcoming WebGPU toggle appears to join an existing set of Chrome protections surfaced in Advanced Protection Mode, such as Android Safe Browsing, HTTPS‑first browsing, and stricter JavaScript defenses. The change is controlled by Play Services and server‑side flags, a common pattern for rolling out security features without waiting on full OS updates.
What It Means for Users and Web Developers
For high‑risk users, the trade‑off is straightforward: accepting a potential performance dip in cutting‑edge web apps in exchange for a materially smaller attack surface. Most day‑to‑day browsing, media, and productivity sites will continue to work as expected, and users can always toggle Advanced Protection Mode off when they need peak graphics performance.
Web developers should ensure robust feature detection and graceful fallback paths. If navigator.gpu is unavailable, applications ought to switch to WebGL or CPU execution with clear user messaging. This aligns with guidance from the Chromium security team and OWASP’s emphasis on reducing optional attack surfaces for sensitive contexts.
Google has not formally announced timing, and features spotted in teardown builds can change before launch. Still, the direction is consistent with broader industry moves to harden default experiences and offer “high assurance” modes. Expect a gradual rollout, documentation for enterprise mobility managers to standardize policies, and continued tuning as feedback arrives from Chrome and Android security programs, including Project Zero and external researchers coordinated through CISA’s disclosure process.
