Heimdal, a leading European cybersecurity firm, has been awarded U.S. Patent No. 18333620 for a proprietary method that predicts the likelihood of a domain being malicious. The patented technology, titled “Apparatus and Method of Predicting Malicious Domains,” marks a significant step forward in Heimdal’s mission to redefine DNS security.
This invention forms the foundation of Predictive DNS™, Heimdal’s AI-driven system designed to detect and block harmful domains before they are weaponized in phishing, ransomware, or data theft campaigns. The technology uses a combination of machine learning, natural language processing (NLP), and computer vision — all powered by neural networks trained on millions of data points including domain names and DNS behavior.
“It ensures we can continue to protect our customers with industry-leading innovation.”
The patented method allows for earlier and more accurate threat detection, giving security teams a powerful tool to prevent cyberattacks before they occur. Integrated into the company’s broader Heimdal XDR platform, this development supports fully automated, risk-aware threat prevention strategies.
With DNS traffic involved in over 90% of malware and breach-related activity, Heimdal’s technology addresses one of the most exploited yet often overlooked areas of cybersecurity.
Key features of Heimdal’s Predictive DNS™ include:
High-precision threat detection, measured by strong F1 scores.
Real-time adaptability, with AI models that evolve based on new data.
Seamless platform integration, enhancing Heimdal’s extended detection and response (XDR) capabilities.
As global threats grow more sophisticated, Heimdal’s latest achievement positions it as a critical player in next-generation cybersecurity.