Nvidia pushes the boundaries of artificial intelligence: from agent systems to quantum physics

At Davos Forum 2026, Nvidia CEO Jensen Huang highlighted three key directions in AI technology development that have radically transformed the industry over the past year. He paid special attention to bringing artificial intelligence closer to practical real-world applications, including revolutionary advances in quantum physics and physical understanding of the world. These achievements signal a shift from theoretical models to tools that truly transform science and industry.

Agent AI: From Theory to Solving Real-World Problems

The first major breakthrough concerns the development of autonomous AI systems. Previously, AI models had numerous limitations and a tendency to hallucinate—generating false information. However, in 2025, the company observed a qualitative shift: models began demonstrating the ability to reason logically, plan sequences of actions, and answer complex questions without prior specific training on similar tasks. These capabilities marked the beginning of the era of agent AI, where systems can autonomously solve complex problems in research fields and real-world production scenarios.

Open Model Ecosystem: Democratization of Innovation

The second major achievement was the large-scale deployment and dissemination of open inference models. Jensen Huang particularly emphasized the role of DeepSeek — the first open inference model that became a catalyst for an entire industry. Since its launch, the ecosystem has experienced dynamic growth: research institutions, corporations, and educational organizations gained access to powerful tools for developing their own AI applications. This democratization of technology accelerated innovation and made cutting-edge AI technologies accessible not only to large corporations but also to startups and academic institutions.

Physical AI and Quantum Physics: A New Frontier

The third breakthrough is defined by the development of physical AI — systems that go beyond text processing and acquire an understanding of the physical world. In this area, AI has demonstrated notable successes in several key fields. Systems can now analyze complex biological structures, including protein mechanisms and biochemical processes. Additionally, quantum physics has become one of the most promising areas of physical AI application, where models successfully operate with concepts of quantum systems, quantum mechanics, and interactions of subatomic particles.

Achievements in quantum physics also include understanding fundamental physical laws — models can now predict fluid dynamics, analyze particle behavior under extreme conditions, and approximate solutions to complex quantum-mechanical equations. These advances open new opportunities for materials science, pharmaceuticals, and energy sectors, where understanding quantum physics is critical for developing new materials and drugs.

Thus, Nvidia’s strategy focuses on three interconnected directions: developing intelligent agents, expanding open software, and advancing physical AI. Quantum physics is no longer just a separate field but a fundamental competence necessary to solve the most complex scientific and technological challenges of the future.