CrowdStrike: AI & Cloud Fueling Cyberattack Surge

Cybercriminals are increasingly using trusted identities, cloud services, and AI tools to launch attacks. This is according to CrowdStrike’s 2026 Threat Hunting Report. Intrusion activity rose by about 4% over the past year. This shows a growing focus on more targeted campaigns. Attackers are spending more time exploiting trusted access paths and legitimate infrastructure. Here's the thing: AI is also speeding up these attacks. Threat actors are using large language models to generate malware and phishing emails. They are also using AI to identify vulnerabilities and create exploits, shortening the time between a vulnerability being found and it being used in an attack. One incident involved attackers gaining administrative access to a cloud account. They then sent nearly 200,000 requests to large language models within two minutes, consuming the victim's AI resources. Attackers are also targeting developer ecosystems. They hide malware in software packages uploaded to public repositories. Once installed, these packages can distribute malicious code to thousands of users. The npm package registry accounted for 87% of malicious software packages found during the reporting period. What's more, vishing, or voice phishing, is becoming one of the fastest-growing ways attackers gain initial access. They often impersonate IT support staff. The bottom line is that organizations must secure AI as aggressively as they adopt it to defend against these evolving threats.