Notifire's running roll-up of the year's highest-impact vulnerabilities — what they were, who was affected, what defenders did.
Each year a small handful of CVEs reshape security practice. Heartbleed (2014), Shellshock (2014), Spectre/Meltdown (2018), Log4Shell (2021), XZ (2024). This page is Notifire's running index of the 2026 entrants: which were exploited in the wild, which forced industry-wide patching, and which changed default trust assumptions for years to come.
Notifire's editorial team curates this list against three criteria: severity (CVSS ≥ 9.0 or active exploitation), reach (millions of affected systems or supply-chain blast radius), and persistence (the disclosure changed defensive practice). News briefings on the individual CVEs are linked below as they appear.
Anthropic is expanding its AI vulnerability detection program, Project Glasswing, to 150 critical infrastructure companies. The project uses AI to find security flaws in sectors like power and telecom, but experts warn it could create a massive patching bottleneck for vendors.
xAI has released Grok Imagine Video 1.5, now available on Vercel's AI Gateway. The model generates video with synchronized audio from a single image. This version features improved audio quality, better prompt adherence, enhanced photorealism, and more consistent characters in longer sequences.
Vercel has introduced signed URLs for its Blob storage service. This feature allows developers to generate temporary, secure links for specific files. Each URL is scoped to a single action like uploading or downloading and has a configurable expiry of up to seven days, enhancing security.
Following its acquisition, Broadcom has overhauled VMware's pricing and licensing, leading to significant price increases for many customers. These changes are disrupting long-standing IT strategies, forcing organizations to urgently reconsider on-premise infrastructure and accelerate decisions to migrate to the cloud.
Alibaba's new multimodal AI model, Qwen 3.7 Plus, is now available on the Vercel AI Gateway. The model combines vision and language capabilities, allowing developers to build advanced agentic applications for tasks like coding, visual reasoning, and operating graphical user interfaces directly through the platform.
A vulnerability in Meta's AI support chatbot allowed hackers to take over Instagram accounts. Attackers tricked the chatbot into changing an account's email address, enabling a password reset. Meta has confirmed the security flaw is now patched.
Vercel has introduced Elastic Build Machines, a new feature that automatically monitors memory usage during deployments. It dynamically adjusts compute resources to prevent common out-of-memory (OOM) failures, improving deployment reliability and developer productivity by reducing the need for manual intervention.
A new analysis highlights how advanced AI models can now autonomously discover software vulnerabilities at an unprecedented scale. This fundamentally changes the balance between offense and defense in cybersecurity, making traditional disclosure timelines potentially obsolete and requiring an urgent rethinking of security strategies.
Vercel has rolled out an update for microfrontend deployments. Alias and branch-assigned domains now fully inherit the routing configuration from the source deployment. This change simplifies the process of creating preview environments, ensuring aliased URLs behave exactly like the original deployment for developers.
A developer has released an open-source AI racing harness for AI Grand Prix contestants. Built with Rust and the Bevy engine, the tool provides a real-time flight software simulation that matches competition constraints, allowing teams to test their code while waiting for the official qualifier.
Salesforce has launched Headless 360, a new offering marking a significant strategic shift. The move is a response to the rise of AI agents and automated workflows, which reduce the need for users to stay within a single application. This change will impact IT architecture and budgeting.
Salesforce has launched Headless 360, a major architectural shift away from traditional application interfaces. The move allows AI agents, bots, and external tools to access Salesforce data directly via APIs, paving the way for new automated workflows and potential usage-based pricing models.
European tech companies are launching Euro-Office, a new open-source productivity suite, on June 9. It is designed as a sovereign alternative to Microsoft Office and Google Docs, offering governments and businesses more control over their data and software infrastructure.
Vercel has integrated the MiniMax M3 model into its AI Gateway. This is MiniMax's first model with a 1-million-token context window and native multimodal capabilities, designed for complex tasks like software engineering, agentic web browsing, and multi-turn collaboration for developers using the platform.
Palo Alto Networks warns a medium-severity security flaw in its PAN-OS and Prisma Access products is under active attack. The vulnerability, CVE-2026-0257, allows attackers to bypass authentication and potentially establish unauthorized VPN connections, posing a significant risk to affected networks.
An attacker exploited a vulnerability in a Marimo notebook (CVE-2026-39987) to gain access to a system. They then used a large language model (LLM) agent to perform post-compromise actions, including stealing cloud credentials. This marks a new evolution in automated attack techniques.
Hugging Face has released an open-source project for 3D-printable humanoid robot legs. The hardware and software stack costs around $2,500 to build, significantly lowering the barrier for researchers and developers to experiment with physical AI and robotics software in real-world environments.
California lawmakers are proposing an amendment to an upcoming age-verification law. The change would exempt most open-source operating systems, including major Linux distributions, following criticism from privacy advocates who warned the original rules would force decentralized projects to collect user data.
Arm has open-sourced Metis, an AI-powered security framework designed to find complex software vulnerabilities. Unlike traditional tools that rely on pattern matching, Metis uses agentic AI and semantic reasoning to analyze dependencies across components, providing clear explanations for its findings.
The Kubernetes project is correcting historical CVE records for some older, unfixed vulnerabilities. This update will cause vulnerability scanners to flag new issues that were previously miscategorized. Teams should prepare to reassess their security posture based on the updated data before 2026.
Microsoft has launched MDASH, a new AI-powered security platform for large-scale vulnerability discovery. The system uses more than 100 specialized AI agents that work together to automatically scan, validate, and prove security flaws across complex codebases like Windows and other Microsoft software.
Anthropic's Project Glasswing, an AI-powered security initiative, has discovered over 10,000 high- or critical-severity vulnerabilities in widely used software. The project, which launched last month with around 50 partners, demonstrates AI's growing capability in automated vulnerability detection for critical systems.
Critical vulnerabilities have been found in the SEPPMail Secure E-Mail Gateway, an enterprise email security solution. Attackers could exploit these flaws to execute code remotely, read all email traffic passing through the appliance, and potentially gain access to the company's internal network, posing a significant security risk.
Multiple security vulnerabilities have been discovered in the Linux kernel, affecting its cryptographic API and packet sockets. A significant flaw, dubbed "Copy Fail," could allow a local attacker to escalate privileges or even escape a container, potentially leading to a full system compromise. An update has been released.
Ubuntu has issued a security notice for multiple vulnerabilities discovered in the Linux kernel. The flaws affect key subsystems, including network drivers and the BTRFS file system. An attacker could potentially exploit these issues to compromise a system, making immediate updates essential for all affected users.
Cloudflare tested Anthropic's security-focused AI model, Mythos Preview, on its own infrastructure. As part of Project Glasswing, they used the model to scan over fifty internal code repositories to identify potential vulnerabilities and understand how attackers might leverage similar advanced AI tools in the future.
Ubuntu has released a security update for the Linux kernel used in NVIDIA Tegra systems. The update addresses several vulnerabilities in network drivers, NVME drivers, and Netfilter that could allow an attacker to compromise an affected system. Applying the patch is recommended for all users.
The US cybersecurity agency has updated its Known Exploited Vulnerabilities (KEV) catalog with seven new entries, including flaws in Microsoft Windows, Defender, and Adobe Acrobat. The additions signal that these vulnerabilities are being actively used by attackers, requiring urgent attention from IT and security teams.
A new Linux kernel vulnerability named "Dirty Frag" has been found in versions 4.10 and later. It allows for local privilege escalation by combining two previously known flaws. The issue is in the IP packet fragmentation system and was publicly disclosed on May 7, 2026.
Microsoft's May security update for Windows 11 is failing for some users, leaving their systems unprotected. The installation fails and rolls back when the device's boot partition has insufficient free space. The issue occurs during the reboot phase, leaving dozens of security patches uninstalled.
Neeraj Dhiman ·
Frequently asked questions
What makes a CVE “critical”?
A CVSS base score ≥ 9.0, plus at least one of: active in-the-wild exploitation, very large affected population, or remote unauthenticated RCE. Notifire applies the same bar to its critical list.
Where can I subscribe to CVE alerts?
The NVD RSS feed, CISA's Known Exploited Vulnerabilities (KEV) catalog, and vendor PSIRT mailing lists. Notifire's /security RSS feed (notifire.in/rss.xml filtered to security) covers the highest-impact disclosures.
How fast should we patch a critical CVE?
Active in-the-wild exploitation: same day. Critical remote unauthenticated RCE without active exploitation: 72 hours. Critical authenticated or local: 14 days, prioritised against business risk. CISA mandates 14 days for KEV-listed CVEs on US federal systems.
What is SBOM-driven CVE response?
Maintain a signed Software Bill of Materials for every artifact in production; when a CVE drops, query the SBOM database to find every running workload that contains the vulnerable component. Reduces MTTD on a new CVE from days to minutes.