- Статьи
- Internet and technology
- Investing in robots: Izvestia has collected the main trends in the development of AI in 2026
Investing in robots: Izvestia has collected the main trends in the development of AI in 2026
The year 2025 was a turning point for the artificial intelligence industry. Major technology companies, including Google, have openly announced the transition of AI from the experimental zone to the practical application phase. Artificial intelligence has ceased to be a showcase of innovation and has begun to take shape as a basic technological infrastructure comparable in importance to cloud platforms or enterprise resource planning systems. It is this shift that defines the key trends of 2026: autonomous intelligence, physical AI, and advanced multitasking agent systems. For more information, see the Izvestia article.
AI as an industrial technology
By the end of 2025, the AI segment was growing almost twice as fast as the rest of the IT market. The key factor was the beginning of a large-scale industrialization of technologies, said Sergey Golitsyn, head of T1 AI at T1 IT Holding.
—Companies no longer consider AI as a pilot project, but as a full—fledged element of the production environment — a tool that not only optimizes processes, but also generates new sources of revenue, displacing outdated operational approaches," he said.
Generative AI becomes the basis for automation, on the basis of which AI agents and multi-agent systems are formed. They take on tasks previously performed by humans, from analytics and document management to management decision support and demand forecasting. Conversational interfaces are evolving into intelligent assistants, and visual-language models (VLM) are moving beyond multimodal experiments and finding applications in industry and corporate processes.
The era of AI agents and the "command model"
One of the key trends in 2026 is the abandonment of the concept of a single universal AI in favor of distributed systems. Ruslan Dolgopolov, head of the Gazprom ID Operator's product group, notes that the market is moving towards an AI command model.
— Instead of one large and complex "brain", teams of highly specialized AI agents are used, working on the principle of an office: finance, logistics, legal issues and coordination - each agent is responsible for his own area, — the expert noted.
This approach radically changes the way we think about smart search and analytics. The systems are moving from giving short answers to a full-fledged study of the topic: forming hypotheses, comparing data from different sources and evaluating their reliability. At the same time, a new class of services is being formed — "task schedulers", capable of breaking complex goals into stages and autonomously organizing their implementation.
AI agents are beginning to manage business processes as well. Accounting and management systems are complemented by intelligent platforms where purchases, logistics, and reporting are conducted by automated assistants who constantly interact with each other. In the long run, this forms what experts call an "operating system for the mind."
Explainability and the architecture of trust
The growing autonomy of AI is exacerbating the issue of trust. The power of the models increased faster than the willingness of businesses and regulators to rely on their solutions. In 2026, the explainability of AI is becoming a turning point.
According to Ruslan Dolgopolov, the industry is moving away from the "black box" model towards transparent and verifiable systems. Instead of a single logical chain, AI starts using a "tree of thoughts", comparing several possible solutions. This approach is complemented by the principle of double-checking, in which one AI analyzes the conclusions of another for errors and contradictions.
Mikhail Khlebunov, Director of Products at Servicepipe, emphasizes that explainability is becoming a critical requirement for the financial sector, medicine, and government systems. He is convinced that companies and users are increasingly demanding an understanding of why the system made a specific decision. This is the foundation for the large-scale implementation of AI in sensitive industries.
Democratization and localization of AI
AI is increasingly infiltrating the internal processes of companies. According to Sergey Golitsyn, the use of neural networks to increase efficiency in the next two to three years will become part of the corporate culture of most industries, from industry to education. The democratization of technology plays an important role: employees have the opportunity to independently create AI agents for data analysis and report preparation, without involving external IT teams.
A separate trend is the localization of AI. There is a growing number of solutions that can work outside the cloud, inside corporate circuits, or on user devices. This increases the autonomy of systems and the level of data protection, which is especially important in the context of increasing information security requirements.
In 2026, generative AI will become more predictable and manageable for businesses, so it will begin to scale not point-by-point, but as part of the infrastructure. There will be more platform—based approaches that allow us to measure quality, cost and risks, build control and responsibility, and use this basis to launch autonomous scenarios more broadly and boldly, said Denis Nagaev, Technical Director of OSMI IT.
Import-substituted intelligence
At least 80% of Russian organizations in key industries should fully switch to domestic software by 2030, Russian President Vladimir Putin said at the SPIEF plenary session. In 2026, new asynchronous messengers will become an important trend. It is a tool for improving the efficiency of communication within teams, which combines asynchronous video recording, AI and cloud technologies into a single solution for B2C, SMB and Enterprise. The solution is suitable for both companies and private clients — freelancers and the self-employed. With it, you can save up to 35% of your working time by replacing hours of calls and long texts with screencasts with timecodes, the ability to edit videos, leave emoji reactions, receive AI transcription and sammarification of messages.
MTS has launched the first such AI messenger with asynchronous video messages in Russia. It was named MWS Team Stream, and the company is convinced that it will be able to import foreign solutions: OBS, Screen Studio, Bandicam, StreamLabs.
Physical AI and new robotics
Special attention should be paid to the development of physical AI at the junction of neural networks and robotics. Ivan Kalinov, CEO of Yandex Robotics, notes that the new generation of robots is fundamentally different from previous automated systems.
"Physical AI is able to reason, work with uncertainty and adapt to an unfamiliar environment, transferring the logic of language models to the material world,— Kalinov explained in a conversation with Izvestia.
The training of such systems is based on the so—called imitation learning - learning through imitation of a person in a simulation environment. Robots learn basic skills and combine them into long chains of actions. An illustrative example is a robot capable of brewing coffee in an unfamiliar space, focusing on the context and changes in the world around it.
At the first stage, physical AI will be actively implemented in logistics and industry. As equipment becomes cheaper, technology will begin to enter the consumer segment. At the same time, mass distribution does not depend on the shape of robots, but on the economics and reliability of solutions. The speaker believes that the end user will not realize for a long time that his order has been collected or moved by a robot.
Neural cooperation
In 2026, neurostaffing, in Russian — neural collaboration, will become a trend. This is a transformation of work, when the symbiosis of a specialist and an industry-specific AI solution becomes a key asset, said Ruslan Gainanov, founder of Tim Force.
As part of this approach, the employee receives a redistribution of tasks and an increase in competencies using neurotechnology. The transformation of the usual approach to work was confirmed by experts in September 2025 at the International Technological Congress. The congress participants are convinced that AI—powered work platforms are the modern norm.
Risks and new threats
The growing autonomy of AI also carries new risks. Daniil Glushakov, an analyst at the Spikatel Information Security Monitoring center, warned of the risk of a new class of cyber threats in 2026 — compromising corporate AI assistants.
"Having gained control over the AI assistant, an attacker will be able to act on its behalf, disguising attacks as legitimate activity," he said.
In response, companies will strengthen their security policies and control the behavioral patterns of AI agents. At the same time, experts agree that fully autonomous "hacker robots" will remain more of a theoretical scenario in the coming year.
A look beyond the horizon
In 2026, the market is finally shifting from point-to-point AI cases to platform ecosystems and self-service solutions. Artificial intelligence becomes a service and an infrastructure at the same time. According to experts, by 2030, AI in Russia can become the basic technology for managing business processes, and the economic effect of its implementation can reach 7.9–12.8 trillion rubles per year.
At the same time, the industry is taking the first steps towards AGI — strong artificial intelligence. Although it is still a long way off, it is agent systems, physical AI, and trust architectures that form the foundation for the next technological cycle, which will begin beyond 2026. All experts agree that fears that robots will massively displace humans are premature.
The economy is already experiencing a shortage of labor, and at the first stage, smart robots will close this gap. Regulation and legal frameworks will come later, as technology passes the stages of product landing and economic sustainability. The reliability of the systems will grow along with the scale of production, just as it happened in the automotive industry and other high—tech industries.
Переведено сервисом «Яндекс Переводчик»
