MIPT Startup Studio Invests 10 Million Rubles in HiveTrace to Secure AI Systems

Competitive Advantage

The startup studio of Moscow Institute of Physics and Technology (MIPT) has provided HiveTrace with a convertible loan worth 10 million rubles (approximately $115,000) to further develop its AI protection product. The studio invests in DeepTech, BioTech, AI, and other high-potential projects with MIPT founders. Its goal is to expand the number and quality of socially significant projects while equipping students with entrepreneurial skills.

HiveTrace’s solution delivers real-time monitoring and defense for AI systems, with a focus on preventing prompt injection attacks, data leaks, and unethical AI behavior. The technology helps mitigate reputational, legal, and financial risks in deploying large language models (LLMs). Currently, the product is being piloted in Russia’s financial, industrial, and telecom sectors, as well as among GenAI startups.

This step strengthens resilience against AI-related security threats, creates competitive advantages, and represents progress toward national standards. In finance and communications especially, securing AI systems builds trust, ensures data integrity, and reduces operational risks. While the project is local, the trend toward AI security is global—and HiveTrace could serve as a model for similar initiatives worldwide.

Russian companies are only beginning to integrate AI into business processes, but the need for protection will grow rapidly. Large tech firms have expertise in AI security, but their tools are often designed for internal use and are unavailable on the market. Other companies adopting GenAI need accessible products to secure their systems

Evgeny Kokuikin

CEO of HiveTrace and Head of the AI Security Lab at ITMO University

Global Demand for AI Security

Although initially aimed at the Russian market, HiveTrace’s concept is adaptable for international use, especially in collaboration with GenAI startups. Real-time AI security checks are in demand globally, particularly in heavily regulated sectors like finance and telecom.

For Russia, the project is an opportunity to build national expertise in machine learning security and managed AI infrastructure. It also aligns with government initiatives to create domestic AI safety standards.

Systemic Controls and Global Context

Security concerns are among the main barriers to enterprise AI adoption worldwide. A CB Insights survey of 50 corporate leaders found that 46 percent see AI security risks as the top obstacle to deploying generative AI. Experts recommend partnerships to create systemic controls for secure data access and use, with clear governance frameworks.

Global efforts to improve ML security are growing. For example, U.S.-based Protect AI has launched NB Defense, a tool for safeguarding Jupyter notebooks from vulnerabilities. It integrates with JupyterLab and includes a CLI tool to detect sensitive credentials like API keys, private keys, and tokens, as well as known vulnerabilities (CVEs) in ML frameworks and libraries.

MIPT’s startup studio is working on other projects too, including SalesAI, a neural network that evaluates sales managers and updates CRMs; an autonomous pallet-moving system dubbed 'Driverless Loader'; and Karma, an AI model that boosts offline business efficiency through audio and video analytics.

From Pilot to Full Product

MIPT’s startup studio is focused on innovation across DeepTech, BioTech, AI, and related fields. Building effective AI defense tools is essential for responsibly scaling GenAI and LLM applications.

HiveTrace’s pilot may evolve into a full-fledged commercial product. Future development could include integration with domestic RegTech and GovTech solutions, as well as expanded collaboration with international partners to establish joint ML security standards.