Huawei has unveiled a new vision for the future of artificial intelligence and digital transformation, identifying 10 key directions that it says will shape the development of the Intelligent World 2035.

The company presented its vision during an online launch event held on September 26, 2026, under the theme “Act Today. Shape Tomorrow.” The event marked the release of two major research reports: Intelligent World 2035: Turning Vision into Action and the Global Digitalization and Intelligence Index 2026.

Building on the trend landscape outlined in 2024 and the 10 megatrends identified in 2025, the latest Intelligent World 2035 report introduces, for the first time, 10 key directions for building an intelligent world. The report examines the major challenges emerging in the era of agentic artificial intelligence and presents an actionable roadmap for the technology industry.

AI Moves From the Digital World to the Physical World

The report identifies a fundamental shift in the role of artificial intelligence as AI expands beyond the digital environment and increasingly enters the physical world.

Rather than remaining a standalone technological tool, AI is becoming increasingly embedded in core business operations and social processes. According to Huawei, this transformation is changing the underlying logic of industries while accelerating digital and intelligent transformation across a broad range of sectors.

The growing role of agentic AI is central to this transition. Unlike conventional applications that primarily respond to predefined requests, AI agents can continuously perceive their environments, reason, make decisions, dynamically invoke tools and interact with people in real time.

This evolution is expected to significantly increase demand for computing, networking and cloud infrastructure, particularly as the volume and frequency of data generated by AI agents continue to rise.

David Wang Highlights the Role of Agentic AI

David Wang, Deputy Chairman of the Board and Rotating Chairman of Huawei, described agentic AI as a key variable in the ongoing transformation.

Wang noted that for AI to reshape operational paradigms, the foundations of competition and industrial dynamics, several major challenges must first be addressed. While the direction of technological development is becoming increasingly clear, he emphasized that turning the vision into reality requires concrete action.

That principle forms the central focus of this year’s report, Intelligent World 2035: Turning Vision into Action.

According to Wang, the report begins by examining the pivotal changes generated by the current AI revolution, analyzes the challenges associated with those changes and provides practical recommendations for an agentic world. At the center of these recommendations is the development of intelligent infrastructure capable of reshaping user experiences and enterprise operating models.

Agentic AI Drives a New Digital Model

One of the report’s central conclusions is that agentic AI is driving a transition from an application-centric digital environment toward an agent-centric digital world.

In this emerging model, AI agents continuously interact with their surroundings, independently reason and make decisions, dynamically access tools and communicate with humans in real time.

Every stage of this process requires substantial token generation and high-frequency token flows. Huawei’s report predicts that global annual token consumption could increase 100,000-fold by 2035, while traffic generated by AI agents could account for more than 90% of overall traffic.

Such a dramatic change would create significant requirements for computing, networking and cloud infrastructure, far exceeding the demands generated by traditional Internet applications.

Against this backdrop, Huawei has identified 10 key directions for the development of the Intelligent World 2035.

1. Building AGI With Autonomous and General Capabilities

The first direction focuses on building artificial general intelligence, or AGI, with autonomous cognition and broad capabilities that can address a wide range of tasks.

Huawei identifies the expansion of AI into the physical world as an essential path toward AGI. Achieving this goal, according to the report, will require parallel advances in language intelligence, embodied intelligence and scientific intelligence.

The combination of these capabilities is expected to enable AI systems to interact more effectively with physical environments while expanding their ability to understand, reason and perform increasingly complex tasks.

2. Scaling Computing and Reducing Agent Costs

The second direction calls for computing clusters to scale by 100 times, while the cost of agent tasks is reduced by 1,000 times.

Huawei emphasizes that progress cannot depend solely on increasing computing capabilities at the chip level. The industry will also need to transition toward advanced SuperPoD computing systems capable of improving token production efficiency.

This approach is intended to address the rapidly increasing computational requirements associated with agentic AI.

3. New Storage and Memory Systems

The third direction focuses on the development of new storage and memory systems characterized by causal accuracy and trustworthy traceability.

Huawei sees these technologies as essential for supporting the evolution from simple data management toward more advanced forms of knowledge and memory.

Such systems would provide AI with more effective mechanisms for retaining, tracing and using information as intelligent applications become increasingly autonomous.

4. Experience-Centric Intelligent Connectivity

The fourth direction concerns the development of intelligent connectivity infrastructure centered on user experience.

This requires integrating AI directly into communications systems, creating infrastructure capable of supporting the increasingly complex and continuous exchange of information generated by intelligent applications and agents.

As AI agents become more deeply integrated into everyday services, connectivity will become an increasingly important component of the intelligent infrastructure ecosystem.

5. The Tau Scaling Law

The fifth direction introduces the Tau Scaling Law, which explores new methodologies and tools for chip design.

The objective is to jointly optimize latency and energy efficiency by moving beyond conventional geometric scaling and introducing time scaling as a new dimension for semiconductor development.

Huawei believes this approach could contribute to the continued evolution of semiconductor technologies as AI workloads place increasingly demanding requirements on performance and energy efficiency.

6. Establishing an Agent OS

The sixth direction focuses on the development of an Agent OS, representing a shift from simply harnessing the environment toward autonomous evolution.

Huawei describes this as an agent-native technical foundation designed specifically around the needs of intelligent agents.

The report summarizes the capabilities of such an environment through the formula:

(Coordination × Execution × Memory × Connectivity) ^ Evolution

The approach places coordination, execution, memory and connectivity at the center of agent capabilities while emphasizing continuous evolution.

7. Adaptive Intelligence Across Devices

The seventh direction calls for a new paradigm of seamlessly adaptive and ubiquitous device intelligence.

The objective is to enable intelligent services across a wide range of devices and scenarios while maintaining flexibility and continuity.

As intelligent capabilities become embedded in more products and environments, adaptive device intelligence could play an increasingly important role in delivering consistent services regardless of the device or operating scenario.

8. Self-Learning and Safety in Intelligent Driving

The eighth direction focuses on self-learning and advanced safety in intelligent driving.

Huawei envisions vehicles evolving into mobile embodied agents capable of combining movement with increasingly sophisticated reasoning and decision-making.

The report emphasizes both autonomous learning and safety as key components of this development, reflecting the growing role of AI in intelligent transportation and mobility.

9. Breaking Power and Thermal Limits

The ninth direction addresses one of the major infrastructure challenges facing the expansion of AI: energy consumption and thermal management.

Huawei proposes efforts to overcome power and thermal constraints in order to increase computing capacity within AI data centers while improving efficiency.

As AI workloads continue to expand, data center infrastructure will require solutions capable of supporting higher computing performance without proportionally increasing energy and thermal pressures.

10. Protecting Security and Privacy

The tenth direction focuses on safeguarding the intelligent world through stronger security and privacy protection for autonomous agents.

As AI agents gain greater independence and interact with data, applications, devices and enterprise systems, security and privacy become increasingly important elements of intelligent infrastructure.

Huawei’s approach therefore places protection mechanisms alongside computing, connectivity and cloud infrastructure as fundamental components of the future intelligent ecosystem.

Token-Centric Infrastructure for the Agentic AI Era

To address the challenges emerging in the agentic AI era, Huawei’s report calls for the development of end-to-end intelligent infrastructure centered on tokens.

This token-centric infrastructure would cover the complete lifecycle of tokens, including generation, delivery, consumption and security.

The report proposes building SuperPoDs and open ecosystems to improve token production efficiency, developing intelligent connectivity infrastructure to ensure seamless token flows, and deploying agentic-native cloud platforms to accelerate AI adoption across industries.

The ultimate objective is to transform tokens into perceivable intelligent services that can reshape both personal user experiences and enterprise operating paradigms.

Five Strategic Initiatives From Huawei

The report concludes by proposing five strategic initiatives aimed at translating technological innovation into tangible industrial progress.

These include taking a long-term approach, remaining people-centric, upholding the principle of “tech for good,” maintaining a commitment to openness and collaboration, and adopting a systems approach.

Together, the initiatives reflect Huawei’s emphasis on combining technological development with human-centered innovation, collaboration and broader ecosystem development.

Global Digitalization and Intelligence Index 2026

The online event also marked the release of the Global Digitalization and Intelligence Index 2026, jointly developed by Huawei and the Institute of Economics at Tsinghua University.

The report brings together insights from experts in economics, sociology and information and communications technology, with the objective of providing countries with pathways toward the development of an intelligent economy.

According to the report, AI is projected to generate more than US$27 trillion in cumulative economic value over the next five years.

At the same time, global investment in digital and intelligent infrastructure is projected to grow at a compound annual growth rate of 19.14%, exceeding US$4 trillion by 2030.

These investments are expected to provide an important foundation for the continued expansion of the intelligent economy.

Data, Talent and Technology as Core Production Factors

The report also redefines the core factors of production in the digital and intelligent world.

These factors are identified as data, ICT talent, and digital and intelligent technologies.

Huawei further emphasizes that the integration of networking, computing and storage provides the foundation needed for industrial upgrading. The convergence of these capabilities is increasingly important as industries move toward more intelligent operating models.

90 Countries Assessed for Digital and Intelligent Development

The Global Digitalization and Intelligence Index 2026 assesses the digital and intelligent development of 90 surveyed countries.

The report categorizes countries into three stages of development: Builder, Adopter and Frontrunner.

Rather than proposing a single universal model for the intelligent era, the index seeks to provide policymakers with clearer pathways based on the circumstances of individual countries.

The quantitative assessment allows each country to identify areas requiring greater focus and develop a customized pathway that reflects its own conditions, capabilities and development priorities.

Turning Today’s Actions Into Progress Toward 2035

In his foreword to Intelligent World 2035: Turning Vision into Action, David Wang summarized Huawei’s approach by emphasizing the importance of translating visions into practical action.

The report ultimately presents the Intelligent World 2035 as a long-term transformation involving far more than artificial intelligence alone. Computing, connectivity, storage, cloud infrastructure, semiconductors, energy efficiency, intelligent devices, mobility, security and privacy are all presented as interconnected components of the emerging intelligent ecosystem.

Huawei’s latest research therefore places agentic AI at the center of a broader technological transformation, while emphasizing that infrastructure development and collaboration will be essential to turning today’s technological advances into practical progress toward 2035.