Technology leaders and analysts at the IDC Directions 2026 conference in Beijing recently highlighted the rapid evolution of artificial intelligence, signalling a decisive shift from infrastructure development to large-scale enterprise adoption.

The event, which brought together more than 400 decision-makers, investors and experts, underscored how AI is entering a new phase of global expansion, with markets accelerating deployment across industries and use cases.

A symbolic opening moment featured an AI-powered robot sharing the stage with IDC chief executive officer Lorenzo Larini, reflecting the growing real-world integration of advanced technologies.

“IDC has been on the ground here since 1986, and we’ve never seen the pace of change move faster than it is right now,” he said.

“What has become clear is that this is not a market you can afford to observe from a distance because it is a technology force actively shaping how the world moves,” he added.

The discussions centred on what experts described as the AI supercycle, with global enterprise spending expected to reach 940 billion dollars in 2026 and expand to 2.1 trillion dollars by 2029.

According to Kitty Fok, the industry has already transitioned from its first phase, focused on computing power and foundational models, into a second phase driven by applications and intelligent services.

“The global AI industry has entered a super cycle, and the market is now moving from infrastructure build-out to enterprise application explosion,” she said.

This transition is being reflected in the rapid growth of robotics and automation, with forecasts indicating that the global robotics market will expand significantly in the coming years, supported by advances in embodied intelligence and industrial AI.

Spending on embodied intelligence alone is projected to rise from 1.4 billion dollars today to 77 billion dollars within five years, marking an annual growth rate of 94 per cent.

At the same time, the cost structure of enterprise AI is undergoing a transformation, with the concept of the “token” emerging as a key metric for both cost and value.

Zhenshan Zhong explained that enterprise AI has shifted from simple content generation to execution-based systems, where automated agents drive operational outcomes.

“Tokens are the core of cost, and agents are the core of value,” he said.

The Model-as-a-Service market is also expanding rapidly, with forecasts suggesting it will reach 40,000 trillion token calls in 2026, generating approximately 18.6 billion renminbi in revenue.

More than 60 per cent of leading enterprises have already integrated generative AI into core business processes, signalling a broad shift towards AI-native operations.

Efficiency is becoming a critical differentiator, with experts pointing to “tokens per watt” as a more meaningful performance metric than traditional computing benchmarks.

According to Thomas Zhou, inference is expected to account for more than 70 per cent of intelligent computing demand by 2027, while edge infrastructure continues to grow faster than centralised data centres.

“The competitive advantage in AI has shifted because it is no longer about who has the most computing power but about who converts AI into sustainable business capability at the lowest token cost,” he said.

The broader strategic landscape is also evolving, with digital economy policies placing emphasis on innovation, ecosystem development and long-term capability building.

Lianfeng Wu said companies are increasingly focusing on platform-based growth and developer ecosystems to maintain competitiveness in the next phase of expansion.

Industrial AI is playing a central role in this transition, moving beyond pilot projects into full-scale deployment across production, supply chains and decision-making systems.

Kai Cui explained that modern industrial software now integrates perception, prediction and collaborative execution, enabling end-to-end transformation.

“Traditional systems focused on recording and control, while the new generation adds intelligence across the value chain,” he said.

At the consumer level, the evolution of smart devices is also reshaping market dynamics, with buyers placing increasing emphasis on intelligent experiences rather than hardware specifications.

Shipments are expected to reach around 900 million units in 2026, although cost pressures linked to component supply constraints remain a key challenge.

Antonio Wang noted that AI-native devices represent a shift in how value is distributed across ecosystems, rather than a simple upgrade cycle.

The discussions at the forum pointed to a structural transformation of the global technology landscape, where adaptability, efficiency and ecosystem integration will determine long-term leadership.