Google Chrome EUVD-2026-65829

Exploitation requires that the attacker has ALREADY achieved compromise of the Chrome renderer process through a separate, independent vulnerability before this flaw can be triggered - this CVE is a sandbox escape primitive and cannot be exploited standalone. The victim must also actively browse to and render a crafted HTML page (UI:R), providing the attacker-controlled input to the GPU subsystem. The race condition nature (AC:H) means exploitation requires precise timing coordination between concurrent GPU operations, introducing unreliability and likely requiring repeated attempts. What limits exploitation: the mandatory renderer pre-compromise prerequisite restricts this to sophisticated attackers capable of chaining two or more browser vulnerabilities; timing sensitivity of the race condition reduces reliability; no special Chrome configuration or non-default feature flag is required beyond a standard desktop Chrome installation running on any supported platform.

The CVSS base score of 8.3 (AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H) reflects scope-changed, full-impact exploitation, but the Chromium security team's own Medium rating is the more operationally relevant signal - it explicitly accounts for the renderer pre-compromise prerequisite that CVSS cannot fully express. EPSS at 0.20% (10th percentile) is consistent with this: sandbox escapes almost never circulate as standalone proof-of-concept exploits and instead appear as components of expensive, multi-stage exploit chains. There is no CISA KEV listing and no public exploit identified at time of analysis. Real-world risk is moderate-to-high for sophisticated targeted attackers (nation-state actors, advanced cybercrime groups) constructing full browser exploit chains, but low for opportunistic or commodity attackers who cannot supply the required renderer exploit. The AC:H metric captures the race condition timing complexity, which introduces inherent unreliability requiring repeated attempts. Organizations running unpatched Chrome in high-value or high-exposure environments should prioritize this update given the full-host-compromise potential of a successful chain.

An attacker first delivers a separate renderer exploit - for example, a JavaScript engine memory corruption bug - to compromise Chrome's renderer process while a victim visits a malicious or attacker-controlled website. With renderer access established, the attacker presents a crafted HTML page that triggers concurrent GPU operations, exploiting the TOCTOU race window in the GPU component to corrupt shared state. Winning the race allows the attacker to escape the Chrome sandbox and execute arbitrary code at the privilege level of the Chrome process on the underlying host OS.