Russian industrial companies are increasingly turning to artificial intelligence, but several serious obstacles stand in the way of its mass adoption. Experts from the STAQ platform identified four key barriers, "Pervy Tekhnichesky" was informed by "System IX" company.
The first is infrastructure. Deploying a full-fledged data processing center for large language models requires significant investment. The market is overheated, components are in short supply, and delivery times are extended. However, this is a solvable engineering problem.
The second barrier is data. Until equipment information is normalized, structured, and brought to a single standard, talking about industrial AI is meaningless. In real projects, the first months are spent not on implementing algorithms, but on collecting data into a single point and organizing it.
The third barrier is datasets. Obtaining a high-quality industrial dataset from external sources is practically impossible due to process specifics, commercial secrets, and security requirements. Large enterprises have enough of their own data and are already training models on their own information. It is more difficult for medium-sized businesses – they will have to look for partnership models or industry consortia.
The fourth barrier is security. AI access to industrial data is practically regulated through proper architecture. The neural network is deployed within the corporate perimeter, it does not directly affect the technological process and operates in a recommendatory mode: the system sees a deviation, suggests a solution, and the final decision is made by a human.
MES – Manufacturing Execution System – plays a key role in AI implementation. This is where data from equipment, operational logic of the enterprise, and the decision-making human converge. AI begins to yield measurable results not in a pilot, but in real work, when telemetry, incidents, service requests, and production tasks are linked within a single perimeter.
According to STAQ, half of Russian industrial companies already have experience using domestic automation systems. Russian software is used in more than ten million workplaces in aircraft and shipbuilding, railway engineering, and other fields. At CIPR-2026, experts noted a shift in the agenda: industrial automation moved from the phase of "catching up with the West" to the mode of "building our own."
Комментарии