Google fixed more Chrome bugs in June 2026 than in the previous two years combined, thanks to AI

In June 2026, Google reported an unprecedented jump in Chrome security work: in a single month, the company found and fixed more vulnerabilities than in the previous two years combined. The cause was the large-scale use of large language models (LLMs) and specialized AI tools to hunt for bugs in code automatically.

How AI changed the speed of bug fixing

Google began applying LLMs systematically to the Chrome codebase, and the result was unprecedented. June 2026 set a record: the number of bugs found and fixed exceeded the combined total for the entire preceding two-year period.

Chrome is one of the largest browsers in the world, with a codebase of tens of millions of lines in C++ and other languages. At that scale, a traditional security audit required hundreds of engineer-hours and covered only a limited share of the code. Language models change that equation radically: LLMs can scan huge volumes of code for known classes of vulnerabilities, such as buffer overflows, use-after-free bugs, unsafe memory operations and injection vectors, far faster and more systematically than any manual review.

LLMs can also detect context-dependent bugs, which surface only under a particular sequence of calls or specific conditions. Traditional static analyzers often produce false positives on such tasks, or miss the threat entirely.

• June 2026 was a record month for the number of Chrome vulnerabilities fixed
• One month's result exceeded the total for the previous two years
• Tools: language models (LLMs) and AI systems for static code analysis
• Microsoft has recorded a similar effect for its own products

Why experts are not surprised

Cybersecurity specialists have spent the last two years warning that this shift was coming: LLMs can fundamentally change the balance of power between attack and defense in software.

"Companies like

Microsoft and now Google are finding and fixing exponentially more bugs in their products thanks to the use of LLMs and AI tools," according to TechCrunch.

Microsoft, a tech giant with a comparable codebase, also publicly confirms a sharp rise in the number of vulnerabilities found and fixed after adopting AI tools. According to TechCrunch, the trend appeared at the same time at the two largest tech vendors. That points to a systemic shift in the industry, not the isolated success of a single team.

Importantly, the acceleration cuts both ways: the same LLMs that help Google find and fix bugs could, in theory, let attackers discover vulnerabilities faster. The balance between the growing number of fixes and a possible rise in attacker activity remains the key question for the industry in the coming years.

What this means

Using LLMs to find vulnerabilities is moving beyond experiments and becoming standard engineering practice at the largest tech companies. If Google gets more fixes in a single month than in the previous two years, that signals a structural break for the entire cybersecurity industry: the security of mass-market software will improve much faster than before. The question is not whether AI will change security work. It already is.