[Digital Today reporter Chi-gyu Hwang] As the security market is reorganising around AI, mergers and acquisitions among related companies are active as they seek to expand their stakes.

Large and small security firms are moving aggressively on M&A to strengthen AI-based security capabilities and respond to risks stemming from the spread of AI. According to Security Week, there were 21 M&A deals worldwide in July alone, and the trend appears to be continuing into August.

Network security company Fortinet is acquiring AI security company Virtue AI and will strengthen AI security capabilities spanning models, applications and agentic systems. Virtue AI provides an enterprise security and governance platform targeting AI models, conversational applications and autonomous agents. It supports automated testing, real-time protection and compliance management.

Brinqa, a security company specialising in exposure management, has acquired FlexTrack. The acquisition will allow Brinqa to cover the entire continuous threat exposure management process on its platform, from vulnerability discovery to verification of remediation. Cribl acquired Radiant Security's AI-based security operations technology assets. The deal is Cribl's second acquisition in the security field in 2026. Cribl also acquired Israeli detection engineering startup CardinalOps in July.

• Fortinet acquires AI security company Virtue AI, securing real-time runtime guardrail technology • Brinqa acquires FlexTrack, adding penetration test validation capabilities to exposure management • Cribl acquires Radiant Security technology assets, strengthening AI SOC platform

Concerns about AI-driven hacking are also growing. Some analysis says the recent string of AI-origin hacking incidents is not an accident that should not have happened, but the flip side of AI evolving rapidly. According to CrowdStrike's '2026 Threat Hunting Report', AI is now embedded across modern cyber attacks. The report concludes that AI has become both an attacker operational capability and a target. It also pointed to companies' biggest cybersecurity weakness as failing to properly understand internal connection structures that hackers can move through, rather than new hacking techniques.

• Misconceptions and truths about a string of AI hacks: "Not a mistake, it happened because it could" • "AI embedded across the latest cyber attacks, AI-targeted attacks also spreading" • Corporate security slower than hackers, the problem is invisible connections

Other corporate moves and issues surrounding security at home and abroad were also compiled.

AI digital forensics specialist U-RAC will participate in the 'Global Corporate Collaboration Program N-UP' promoted by the Ministry of SMEs and Startups and supported by NVIDIA. Cybersecurity specialist C-Cubista will work with security consulting specialist Novainsecu to target the security market for national core industries, including the public, finance and aerospace sectors.

• U-RAC joins NVIDIA's 'N-UP', advancing on-device AI forensics technology • C-Cubista cooperates with Novainsecu to jointly develop integrated security services

OpenAI released 10 guidelines for responding to AI autonomous attacks after an incident involving an attack on its model being tested on Hugging Face.

• OpenAI tightens security after Hugging Face attack, blocks attacks with AI • AI attack era, 10 security rules highlighted by OpenAI chairman

Chinese AI companies are accelerating efforts to expand their influence into the cybersecurity market by putting forward open-weight models. Chinese AI startup Z.ai unveiled GLM-5.3, an AI model specialised for cybersecurity.

• China's open-weight models also target security, Z.ai challenges Mythos with GLM 5.3

Apple sent threat notifications to some users in 110 countries that they may have been targeted by mercenary spyware attacks.

• Apple warns users in 110 countries about mercenary spyware, also issues new support document

Swimlane added an intelligent routing feature that automatically classifies alerts coming into security operations centres (SOC) and splits investigation paths. The feature is designed to send alerts through an appropriate path among deterministic automation, AI-based investigation and agentic automation. Palo Alto Networks is introducing a new collaboration programme, the 'Frontier AI Critical Defense Program', to respond to vulnerabilities in the frontier AI era. CrowdStrike global chief technology officer Elia Zaitsev (엘리아 자이체프) is leaving the company to establish a new venture fund called Cognition.

• Swimlane strengthens security operations centres with intelligent routing • Palo Alto Networks expands its AI security threat response collaboration programme, forward-deploying virtual patches • CrowdStrike CTO sets up cyber venture fund 'Cognition', betting on an AI-driven reshaping of the security landscape