A new report from Black Kite warns that over 48,000 vulnerabilities were published in 2025, with hackers exploiting them before patches even become available—a negative seven-day window according to Mandiant's data. While this creates an impossible patching challenge, the report found that only 58 of these vulnerabilities pose genuine, discoverable threats to enterprise supply chains, highlighting that organizations need better visibility to focus on truly critical risks. The problem is expected to worsen as AI discovers more vulnerabilities and introduces new weaknesses through rapid coding and autonomous agents that operate hidden within corporate infrastructure.

Google has seen a dramatic surge in Chrome vulnerability discoveries, jumping from handfuls of bugs in early April to over 100 in early May, with strong indications that AI tools are behind the spike. While Google hasn't explicitly confirmed AI's role, the timing aligns with their recent statements about AI helping security teams remediate risks at unprecedented rates, and they've been developing AI-powered tools like CodeMender and Big Sleep for vulnerability detection. This mirrors similar breakthroughs at Mozilla, Microsoft, and Palo Alto Networks, which have also reported major increases in bugs discovered using AI-assisted scanning.

Microsoft has patched two critical zero-day vulnerabilities in Windows Defender that were actively exploited in the wild. The flaws, known as UnDefend and RedSun, allow attackers to escalate privileges to system-level access and launch denial-of-service attacks, and they're linked to the publicly released BlueHammer exploit toolkit. CISA has added both vulnerabilities to its Known Exploited Vulnerabilities list, giving federal agencies until June 3rd to apply the patches, while urging all organizations to update immediately.

Socket, a California-based supply chain protection provider, has raised 60 million dollars in Series C funding at a 1 billion dollar valuation, bringing its total funding to 125 million dollars. The company, founded in 2020, uses AI-assisted analysis combined with human verification to scan open source dependencies for malicious behavior before they affect enterprise products. Socket plans to use the new funding to enhance its Socket Firewall product, expand its certified patches for exploited vulnerabilities, and extend protection to browser extensions, code editor tools, and AI applications as software development becomes increasingly automated.

Drupal has patched a highly critical vulnerability that could allow hackers to compromise websites using the popular open source content management system through SQL injection attacks. The flaw, which affects sites using PostgreSQL databases, can be exploited without authentication to steal information and potentially achieve remote code execution. This is Drupal's first highly critical security issue in years, and patches are now available for versions 11 and 10 of the CMS that powers hundreds of thousands of websites worldwide.

Apple blocked over 2 million apps from entering the App Store in 2025 and prevented more than 2.2 billion dollars in potentially fraudulent transactions using a combination of artificial intelligence and human review. The company also deactivated more than 40 million accounts for fraud and abuse, terminated 193,000 developer accounts suspected of fraudulent activities, and blocked nearly 195 million fake ratings and reviews. Apple's AI-powered security measures have proven increasingly effective at detecting complex malicious patterns including apps engaged in bait-and-switch tactics, hidden features, and clones, while also preventing nearly 3 million attempts to install apps distributed illicitly outside official stores.

Ocean, a New York and Tel Aviv-based startup, has emerged from stealth with 28 million dollars in funding for an AI-powered email security platform. The company deploys specialized AI agents to inspect every incoming message, evaluating sender intent and analyzing conversation context to detect sophisticated threats like business email compromise and AI-generated phishing attacks that are designed to look completely legitimate. Ocean also automates security tasks for IT teams including email triage and incident handling, while providing real-time guidance to employees when they encounter suspicious messages.

Cisco has patched a critical vulnerability in its Secure Workload software that earned a perfect 10 out of 10 severity score. The flaw could allow attackers to send crafted API requests to gain Site Admin privileges, letting them read sensitive information and modify configurations across tenant boundaries in both cloud and on-premises deployments. Cisco says the vulnerability hasn't been exploited in the wild yet, but is urging all users to update immediately to avoid potential attacks.

A critical nine-year-old vulnerability in the Linux kernel has been discovered that allows attackers to execute commands with root privileges on major Linux distributions. The flaw, which has existed undetected since its introduction nearly a decade ago, poses a significant security risk as it affects widely-used operating systems across the Linux ecosystem. Security teams are now working to patch the vulnerability across affected distributions.

Identity-based attacks have become a primary pathway for cybersecurity breaches, as attackers increasingly exploit credentials and access rights rather than traditional network vulnerabilities. Modern threats leverage compromised identities to move rapidly through systems, particularly through remote access tools like VPNs that were designed for security but now provide attackers with fast lateral movement capabilities. Organizations need to fundamentally rethink their security strategies to protect identity infrastructure, as artificial intelligence is accelerating these attacks and shrinking the window for human response.