Something unprecedented happened in June 2026: Google patched more security vulnerabilities in two Chrome releases than it had in the previous two years combined. That's not a typo. It's the clearest sign yet that AI-powered security tooling has crossed from promising experiment to operational game-changer — and it has profound implications for every security team managing enterprise browser risk today.

The Numbers That Stopped the Security World in Its Tracks

Google reported that Chrome 149 and Chrome 150 fixed 1,072 security bugs, surpassing the total number fixed across the previous 23 Chrome milestones combined.

To put that in historical context,

the figure exceeds the 1,036 fixes recorded across the previous 23 Chrome versions released over the last two years.

The divergence is significant: it suggests the exponential increase in found-and-fixed bugs is not an artifact of software becoming less secure, but a consequence of companies choosing to deploy AI at scale in their security pipelines.

And Google isn't alone.

Microsoft's July 2026 Patch Tuesday fixed a record 570 vulnerabilities

, with

Microsoft linking the increase in patched vulnerabilities to an AI-powered vulnerability discovery system that has identified more security flaws across the Windows codebase.

The age of AI-accelerated patching is here — whether security teams are ready for it or not.

Meet the AI Behind the Milestone: Big Sleep and Gemini

Google introduced an AI-driven agent named Big Sleep, developed in collaboration with Google DeepMind and Project Zero, as well as a dedicated Gemini-powered agent to scan the codebase.

The evolution of this toolchain didn't happen overnight.

In 2023, Google developed ways to use LLMs to increase security fuzzing coverage and performance. In 2024, it worked with Project Zero on Naptime, giving LLMs specialized tools for vulnerability research. In 2025, it collaborated with DeepMind and Project Zero on Big Sleep. And in early 2026, it built an agent harness that used Gemini to find vulnerabilities across the broader Chrome codebase with higher efficiency and lower false positives.

Unlike traditional approaches that rely primarily on manual code review and static analysis, Big Sleep leverages machine learning to identify patterns and anomalies that may indicate previously unknown vulnerabilities.

This makes it capable of surfacing subtle, logic-level bugs that conventional fuzzing tools would likely miss entirely.

The 13-Year-Old Bug That Proved AI's Worth

Perhaps the most striking demonstration of this technology's power came in the discovery of a vulnerability that had been hiding in plain sight for over a decade.

One of the bugs found was a sandbox escape that would allow a compromised renderer to trick the browser into reading local files — a bug that had quietly survived in the codebase for more than 13 years. For many at Google, this moment cemented the potential of AI-powered vulnerability detection.

The vulnerability discovered by the system was a Chrome sandbox escape that had remained in the codebase for more than 13 years. If exploited, the flaw would have allowed a compromised renderer to escape the sandbox and trick the browser into reading local files.

This example underscores a critical truth: long-lived codebases accumulate "legacy debt" that human reviewers simply cannot audit comprehensively at scale. AI doesn't get tired, doesn't skip edge cases, and doesn't carry assumptions about which parts of a codebase are "already safe."

How the AI Pipeline Works End-to-End

What makes Google's approach especially powerful is that AI isn't just being used at the discovery stage — it's embedded across the entire vulnerability management lifecycle.

The company now uses large language models throughout the vulnerability management process, including discovering flaws, reproducing reports, determining severity, assigning bugs to developers, generating candidate patches, and creating tests.

Once a bug is flagged,

"Fixing agents" draft potential candidate code, while separate "critic agents" evaluate the patches against Chromium standards, and "test-writing agents" generate cross-platform tests.

This is done "before a developer reviews the fix," Google confirmed, "saving up to weeks of developer time."

On the triage side,

AI now automates parts of vulnerability triage, including filtering duplicate or invalid reports, reproducing bugs, assigning severity ratings, and routing issues to engineering teams.

The company estimated this saves developers hundreds of hours every month.

Additionally,

AI tools integrated into its continuous integration system blocked more than 20 vulnerabilities from reaching production in May, including one critical issue.

The Patch Gap Problem — And How Google Is Solving It

Finding and fixing vulnerabilities faster creates a new, urgent challenge: the patch gap.

Once a patch lands in Chrome's public open-source codebase, attackers race to reverse-engineer it to exploit unupdated browsers — a window known as the "patch gap."

While it takes Google 1–2 days to triage, fix, test, and release a patch, the "time spent waiting for the user to restart Chrome can be a significant contributor to N-day exploitation risk."

Google is attacking this problem on multiple fronts:

The company is moving to a two-week Chrome milestone release cycle with weekly security updates and is piloting two security releases each week.

Google is developing "dynamic patching," which would allow Chrome to apply updates without restarting the browser.

Google is "working on automating the generation of release notes and CVE descriptions from security bug fixes to eliminate manual bottlenecks and shorten the window between vulnerability discovery and public disclosure."

What This Means for Your Security Team

This shift isn't just Google's internal success story — it's a wake-up call for every enterprise security and IT team managing Chrome deployments at scale.

Voluminous patch updates could soon be the norm, as artificial intelligence accelerates the speed and scale of vulnerability discovery.

The implications are significant:

The patching burden is permanently higher. Microsoft has updated its guidance to deploy quality updates within three days, not the 30-day window most enterprise IT teams use.

Cybersecurity experts have long warned that advances in large language models could allow AI systems to discover security flaws at a much larger scale, requiring defenders to adopt similar technologies to stay ahead of malicious actors.

Practical Tips Security Teams Can Act On Right Now

Here's how to adapt your security operations to this new AI-driven patching reality:

  1. Enable automatic Chrome updates across your fleet. With Google piloting twice-weekly security releases, manual update processes create unacceptable lag. Use MDM or enterprise policies to enforce silent, automatic browser updates.

  2. Shrink your patch SLA to 72 hours for critical browser vulnerabilities.

Fixes committed to the main "tree" typically take weeks to reach the Chrome Stable channel

— once they arrive, deploy them immediately.

  1. Monitor Chrome version telemetry continuously. Ensure your asset inventory tracks Chrome versions in real time so you can immediately identify unpatched endpoints when a critical CVE drops.

  2. Review your VRP and bug bounty posture.

Google saw a sharp increase in reports submitted through the Chrome Vulnerability Reward Program, and by March 2026, it had received more security bug reports than during all of 2025.

External researchers are more active than ever; make sure your own program reflects this new pace.

  1. Prioritise memory-safe coding in your own development pipelines.

Google's long-term architectural solution is a migration toward memory-safe languages like Rust

— consider similar investments in your own codebases.

  1. Educate end users on browser restart hygiene. Until dynamic patching is fully rolled out, the most critical gap is often a browser sitting open for days without a restart. A simple user awareness campaign here can meaningfully reduce your attack surface.

  2. Evaluate AI-assisted vulnerability scanning tools for your own environment.

The technique Google's AI agent used — cross-referencing current code against historical CVE patterns and Git history — is applicable to virtually any large open-source codebase, including the Linux kernel, major JavaScript runtimes, and cryptographic libraries.

Conclusion: The New Baseline for Browser Security

What Google has achieved with Chrome 149 and 150 isn't just an impressive headline — it's a preview of the new normal. AI is compressing the vulnerability discovery cycle from months and years to days and hours, and the entire industry is following suit.

There is no doubt that AI has been a critical cog in the Chrome security machine that helps Google get security updates out faster and faster, and will likely continue to accelerate this process.

For security teams, the message is clear: passive patch management is no longer a viable strategy. The organisations that will stay ahead are the ones that match their patch velocity to the AI-driven pace being set by vendors — and begin exploring AI-assisted tooling within their own security operations.

Is your security team's patch management process built for the AI era? Now is the time to find out. Review your browser update policies, compress your patch windows, and start evaluating AI-driven vulnerability scanning tools before attackers use the same technology against you.