Huawei’s Intelligent World 2035 sets out ten defining technology trends that serve as both the backbone of its vision and a framework for understanding the coming transformation of the global economy, industry, and governance. The analysis is not just predictive; it positions these trends as catalysts for change with strategic significance and practical value.

The report’s foresight, systematic approach, and bold projections are striking. Below is a summary of the ten trends and their implications.

1. Toward the Physical World: The Path to AGI

Artificial General Intelligence (AGI) is seen as the ultimate goal of AI—systems with human-level cognition, autonomy, and creativity. Huawei argues that AGI will become the most transformative force of the next decade, though it still faces formidable challenges.

• Three Decades of Evolution: 1) Statistical learning revived AI in its first decade; 2) deep learning drove breakthroughs in perception; 3) Transformer models ignited cognitive AI; 4) the next decade will go “Beyond Transformers.”
• Three Engines: Data-driven experience, rule-driven reasoning, and practice-driven action will together form “world models” that allow AI to truly engage with the physical world.

2. From Tools to Decision Partners: AI Agents Drive Industry Transformation

AI “agents” will evolve from simple assistants into decision-making collaborators. Their success depends on maturing technology, trust frameworks, and business ROI. By 2035:

• Early stage: productivity assistants in customer service, sales, office tasks.
• Mid-stage: breakthroughs in drug discovery, professional consulting, cost reduction.
• Long term: embodied robots and AI PCs reshape human experience, opening trillion-dollar industries.
  The human–machine relationship will shift from division of labour to true symbiosis.

3. Human–AI Collaborative Programming: Reinventing Software

The complexity of software is rooted in human limitations. Large models have already transformed programming, alleviating bottlenecks and boosting productivity.

• Co-programming will become mainstream, with models and formal methods complementing each other.
• AI agents will generate adaptive, “malleable” applications from natural-language prompts, while professional developers continue to build mission-critical systems.
  Future software will blend deterministic logic, neural networks, and agent orchestration.

4. AI and the Mirror World: Immersive Interaction

Interaction will move from 2D screens to multi-modal, spatial, and sensory-rich experiences.

• Interfaces will progress from keyboards to gesture sensors, voice, and eventually full five-sense integration.
• Devices may include gesture-sensing rings, pulse-tracking necklaces, ultra-light 3D glasses, and brain–computer interfaces.
  The result: a seamless fusion of physical and digital realities.

5. From Apps to Agent Ecosystems

Instead of millions of isolated mobile apps, the future internet will be an ecosystem of interconnected AI agents.

• User commands will trigger cooperative multi-agent workflows across services.
• Success metric: “end-to-end task success rate” will define competitiveness.
  The formula for multi-agent strength: strong foundational models × (coordination frameworks + specialisation + learning + trust).

6. Embodied Intelligence: Crossing into Trillion-Dollar Industries

AI is gaining a physical form through embodied intelligence—merging AI with robotics, sensors, energy systems, and networks.

• Autonomous Driving: By 2027, L4 pilots; by 2030, scaled L4 in selected contexts; by 2035, widespread L4+ and early L5 trials.
• Robotics: Around 2030, robotics may see its “ChatGPT moment” with vision-tactile-language-action models. By 2035, home robots could cost under US$10,000.
• Low-Altitude Economy: Advances in battery energy density (towards 800 Wh/kg) and hybrid power systems could make personal aerial vehicles feasible by 2035.

7. Beyond the Von Neumann Architecture: New Compute Paradigms

By 2035, global compute demand could be 100,000 times higher than in 2025, driven by pre-AGI models. Meeting this requires breakthroughs in:

• Materials: beyond silicon to low-dimensional materials.
• Engineering: 3D heterogeneous integration, wafer-level computing.
• Architecture: from compute–storage separation to in-memory and neuromorphic designs.
• Paradigms: from classical to analog, optical, brain-inspired, and quantum computing.

8. Data as Intelligence: The New Storage Paradigm

Data is the “new fuel.” By 2035, over 70% of data will be “warm” or “hot,” constantly feeding AI models.

• Traditional three-tier storage will evolve into a two-tier “hot–warm/cold–warm” structure.
• New storage must handle contextual, stateful, and evolving data—enabling continuous learning and fast rollbacks while cutting costs.

9. From Mobile Internet to the Agent Internet

Connectivity will expand from 9 billion people to 900 billion devices and intelligent agents.

• PC Internet: on-demand, file-driven, few producers.
• Mobile Internet: always-on, rich media, everyone a producer.
• Agent Internet: real-time, hyper-dimensional, humans and agents as co-producers and co-consumers.
  Networks will evolve to ultra-broadband, ultra-low-latency, secure, integrated sensing–comms–computing.

10. Energy Tokens: AI as the Nervous System of Energy Networks

Energy is the ultimate constraint. By 2035, renewables will surpass fossil fuels, with wind and solar >50% of generation. Hydrogen and advanced nuclear will complement this shift.

• Space-based solar pilots (10MW in orbit by 2035) and progress in fusion could redefine supply.
• Energy “tokens” managed by AI will dynamically allocate every joule, turning the energy grid into an intelligent nervous system.

Conclusion

Huawei frames Intelligent World 2035 not only as a technology roadmap but as a manifesto for an “intelligent civilisation.” It positions AI’s evolution from a tool of efficiency to a partner in human progress.

As the report notes:
“Only when technology is woven into our lives, families, businesses, and environment can it truly advance civilisation.”

The vision is clear: by 2035, AI will permeate industry, infrastructure, and society, enabling a sustainable, intelligent future.