Google says it fixed more Chrome bugs in June than over the past two years, thanks to AI

By Aoife Gallagher · Reporting from Dublin · July 30, 2026

The sheer volume of it is staggering: 1,072 security bugs fixed in the last two Chrome milestones alone.

The Codebase as Infrastructure Crisis

The sheer volume of it is staggering: 1,072 security bugs fixed in the last two Chrome milestones alone. That figure—more than the combined total from the previous twenty-three versions over two years—is not merely a technical boast; it is an announcement of infrastructural transformation. Google claims that by deploying Gemini AI and specialized agent harnesses, they have made fixing vulnerabilities into an "automated, industrial-scale operation," as Doug Turner stated. The reporting across outlets like techcrunch.com and blog.google confirms this leap: the team has moved from manual patching to a system where AI agents triage issues, draft fixes, and write cross-platform tests. While Microsoft’s record 570 patches or Apple's independent count of 482 bugs are impressive numbers in their own right, they fail to capture the mechanism at play here. What Google has built is not just a faster patch cycle; it is a new kind of gatekeeping power.

The Exponential Curve and the Industrial Revolution

This isn't simply better engineering; this is an exponential leap comparable only to the most profound shifts in human history. I am speaking, naturally, of the Industrial Revolution. That era was defined by a fundamental technological shift that automated and scaled production, leading to output volumes previously unattainable by human labour alone. The mechanism here is identical: AI has taken the painstaking, highly skilled work of security auditing—the manual process of finding flaws in complex codebases like V8 or BoringSSL—and scaled it into an autonomous pipeline. We are witnessing a digital Industrial Revolution for software maintenance.

A New Dependency on Silicon Valley’s Plumbing

The implications are immense and deeply troubling. When the ability to maintain basic, functional security is concentrated within proprietary AI models and centralized corporate pipelines, the periphery suffers. The small nation survives by maintaining its own infrastructure; it cannot simply wait for a benevolent superpower to update its plumbing. Google's accelerated cadence—piloting two security releases per week, as noted by pcmag.com—is presented as a necessary response to "fast-moving, AI-powered attacks." But what we are really seeing is the establishment of an unchallengeable dependency. The critical flaw that survived for thirteen years, which androidauthority.com reported was squashed using these new tools, reminds us that even decades of human effort can leave deep structural vulnerabilities.

The promise of "dynamic patching"—hot-swapping background processes without a full browser restart—sounds miraculous, but it is merely the latest layer of proprietary control over our digital lives. The sheer scale means that any failure in this automated system, any flaw in the AI's training data or its judgment, will not just affect one user; it will threaten global connectivity.

The era of the independent developer patching their own niche service, or the small nation maintaining its bespoke digital systems without constant reliance on Silicon Valley’s benevolence, is drawing to a close. This isn't progress; it is consolidation. The ability to secure our shared digital infrastructure has become an oligopoly function, and until there are robust, open-source, and decentralized mechanisms for verifying and contributing to this foundational code—mechanisms that do not require the continuous patronage of one corporation’s AI breakthroughs—we will remain subjects of its update cycle.

Google says it fixed more Chrome bugs in June than over the past two years, thanks to AI

The Codebase as Infrastructure Crisis

The Exponential Curve and the Industrial Revolution

A New Dependency on Silicon Valley’s Plumbing

Sources