Kaspersky and Sber Tackle AI Vulnerabilities

A New System

Kaspersky and Sber announced plans to co-develop a platform based on Sber’s multi-agent generative AI system to automate vulnerability detection and management. The solution is expected to reach the market in 2026.

Vulnerabilities remain one of the main attack vectors in IT infrastructure. Automating vulnerability management with AI promises faster response times and fewer human errors. The joint initiative is therefore both a practical security measure and a statement of Russia’s push toward technological sovereignty, offering a domestic solution competitive with global standards.

Risks and Growth Opportunities

The product is likely to find demand across multiple industries, from government and banking to corporations and infrastructure providers—anywhere vulnerabilities pose a high risk. Beyond detection, the system is expected to feature automated remediation tools and integration with SOC, MDR, and other services. This functionality could even attract international customers, leveraging Kaspersky’s global brand.

Software vulnerabilities are among the most common attack vectors into IT infrastructures. Automated AI-based solutions help organizations identify, assess, and remediate weak spots before they can be exploited. This joint project with Sber will establish best practices for applying cutting-edge AI technologies, including multi-agent systems, in vulnerability management. Our collaboration is a prime example of how market leaders unite in long-term partnerships to deliver reliable and effective protective solutions

Anton Ivanov

Chief Technology Officer at Kaspersky Lab

However, external markets may resist due to distrust of Russian technologies. Other challenges include strong competition from Western vendors, potential technical issues such as false positives or lack of AI explainability, and the risk of vulnerabilities within the AI itself. Each of these hurdles will require targeted solutions.

At the same time, the project offers room for growth: autonomous patching, blockchain verification of fixes, and federated learning—enabling AI models to train without exposing sensitive data. Future expansion could cover configuration flaws, cloud environments, and containerized infrastructures.

Toward an Industry Standard

The integration of AI into SOC, XDR, and threat intelligence tools has been a global trend for years. In Russia, reports already suggest updates to Kaspersky’s cloud security services with enhanced automated analysis. Industry news also points to the use of multi-agent AI systems in vulnerability management under this partnership.

Globally, major vendors, including Microsoft, are deploying AI-driven security. It has quickly become an industry standard. The Russian collaboration highlights domestic efforts to close cybersecurity gaps without reliance on foreign solutions.