Novosibirsk State University and AIRI Launch Industrial AI Lab
Novosibirsk State University and the AIRI Institute have signed an agreement to establish a joint research and education laboratory focused on industrial artificial intelligence. Based at NSU's Institute of Intelligent Robotics, the lab will bring together eight full-time researchers alongside 20 to 30 students from the university's Applied Artificial Intelligence program.
The new laboratory is more than just another line item in a research grant report. It represents a bid to strengthen technological sovereignty in some of industry's most demanding domains, including metallurgy, oil and gas, energy, and robotics. Artificial intelligence is steadily moving beyond creative applications and into environments where it must operate under real-world industrial constraints.
From Chatbots to Robot Swarms and Physics
The choice of location is deliberate. The laboratory will be based at NSU's Institute of Intelligent Robotics and the A.P. Ershov Institute of Informatics Systems, where AI algorithms will be tested against the realities of physical systems rather than purely digital environments. Research will focus on five priorities: intelligent equipment diagnostics, physics-informed AI models, generative AI for science and medicine, multi-agent control systems for robotic swarms, and industrial data analytics.
The underlying challenge is straightforward. Industrial companies collect enormous volumes of sensor telemetry but often struggle to turn that data into actionable insights. An AI model capable of predicting a turbine failure a week before it occurs, or coordinating a fleet of drones at a hazardous industrial site, could save billions of rubles while helping prevent accidents and protect human lives.
Training Engineers and Expanding Compute Resources
So why was AIRI chosen as the partner? According to laboratory director Igor Solodov, it came down to practical considerations. The institute is focused exclusively on artificial intelligence, actively shapes current research directions, and, critically for education and experimentation, provides access to high-performance cloud computing resources.
The laboratory's permanent staff will consist of eight researchers, including five from AIRI and three from NSU. The project also incorporates undergraduate students from the Applied Artificial Intelligence program launched in 2025 under the federal Top-AI program. As Alexey Okunev, director of NSU's Institute of Intelligent Robotics, explains, students will base their course projects and theses on real industrial challenges. The goal is to train engineers who understand noisy industrial data and the cost of failure in production environments, rather than graduates whose experience is limited to working with APIs.
AIRI's Strategy and Project-Based Funding
AIRI CEO Ivan Oseledets describes the NSU laboratory as a natural next step, noting that the institute's model of establishing joint laboratories with leading universities has already demonstrated its effectiveness. Funding will follow a strictly project-based approach: first an industrial partner is identified, then a team is assembled around a specific technical challenge, and finally the project is developed with deployment as its intended outcome.
Potential partners span a wide range of industries, from oil and gas and metallurgy to manufacturers of batteries, energy storage systems, and autonomous vehicles. Also, the laboratory intends to maintain a strong focus on fundamental research through publications at leading international conferences such as NeurIPS, ICML, and ICLR, while also developing open benchmarks for the broader research community.
Prototypes, Deployment, and the "Valley of Death"
The research team has set an ambitious roadmap. By the end of its first year, it aims to deliver software prototypes, while by the end of the third year it expects to produce operational systems ready for pilot deployments at industrial facilities. Before that comes the familiar "valley of death" that separates laboratory prototypes from production environments. Industrial operators cannot tolerate hallucinations or unreliable AI behavior. The team's major challenges will therefore be ensuring cybersecurity, integrating AI into existing industrial process control systems, and, above all, proving measurable economic value.
The Era of Heavy Technologies
For NSU, the laboratory offers an opportunity to attract non-budget funding while strengthening the capabilities of its research supervisors. For Russia, it represents a chance to build a center of expertise that brings together higher education, scientific research, and industry. Ultimately, the project's success will be measured not by citation counts but by the number of deployed AI systems that prevent industrial accidents around the clock.
