Binarly, a leader in software supply chain security, has been granted U.S. Patent No. US 12,236,262 B1 for its machine learning-driven technique that optimizes large-scale binary executable analysis. The patent, issued on February 25, 2025, introduces an AI-based system that predicts peak memory usage and dynamically allocates computing resources, enhancing efficiency in software security assessments.
By leveraging advanced preprocessing and machine learning models trained on extensive datasets, the technology estimates memory requirements and intelligently queues tasks when resources are limited. This innovation reduces processing delays, lowers operational costs, and improves the accuracy of vulnerability detection in software supply chains.
CEO and Head of Research Alex Matrosov emphasized the significance of this advancement, stating that the patented technology dramatically increases the throughput and precision of security analysis. He confirmed that these techniques have already been integrated into the company’s flagship Binarly Transparency Platform, reinforcing its capability to detect vulnerabilities and mitigate security risks at scale.
The invention, developed by researchers Alex Matrosov, Sam Thomas, Yegor Vasilenko, and Lukas Seidel, is expected to have broad applications in enterprise environments where rapid software analysis is essential. By improving automated vulnerability management and risk assessment, Binarly’s AI-powered system sets a new standard for securing firmware and software supply chains.
