Software Vulnerability Discoveries Nearly Double in 2026, Driven by AI Tools

The discovery of software security flaws is surging in 2026, with the number of reported vulnerabilities on track to nearly double the total from 2025, a trend attributed to the growing use of powerful artificial intelligence tools by security researchers.

A Surge in Disclosures

According to a Bloomberg News report on Monday, the U.S. National Vulnerabilities Database has recorded 45,207 new flaws between January and July 27. This figure is already approaching the total number of vulnerabilities discovered throughout all of 2025. The database serves as a critical repository, tracking digital security holes that malicious actors could exploit for criminal or espionage purposes.

The increasing volume of disclosures is putting pressure on corporate cybersecurity teams to patch systems at an unprecedented rate. The trend highlights a double-edged sword: while more flaws are being found and fixed, it also reveals a vastly expanded potential attack surface for hackers.

Tech Giants Report Record Patches

Several major technology companies have reported a significant spike in the number of bugs they are addressing. The latest software updates show a dramatic year-over-year increase in patched vulnerabilities:

Ad

• Oracle Corp. (ORCL) fixed a company-record 1,449 security issues in its July software update, a sharp increase from the 309 fixes in the same update last year.
• Microsoft Corp. (MSFT) disclosed 642 security bugs in July, a figure nearly five times higher than the count from July 2025.
• Alphabet Inc.’s (GOOGL) Google found and patched 433 bugs in a recent Chrome browser update, compared to just 11 in an equivalent update one year prior.

AI's Role in Vulnerability Discovery

The proliferation of AI-powered tools is seen as the primary driver behind the accelerated pace of discovery. These systems can analyze code and identify potential weaknesses far more efficiently than human researchers alone.

"These tools are increasing people’s ability to find vulnerabilities in software," said Gabriel Shapiro, distinguished AI research scientist at cybersecurity firm SentinelOne Inc. (S), in a comment to Bloomberg. The trend suggests that both ethical hackers and malicious actors are leveraging AI, intensifying the race between cyber defense and offense.