57% of U.S. enterprises expect providers to pre-build agents, while 54% of Chinese enterprises prefer customizable agent building.According to Caijing Tuya, a corporate intelligence expert, on December 2, Rick Villars, Vice President of IDC Global Research, delivered a keynote speech titled “Toward a New Era of Intelligence: Three Drivers Reshaping the Global IT Industry”at the IDC FutureScape 2026: China ICT Market Prediction Forum.He noted that the global technology industry is entering an era of expansion. By 2027, total spending on servers and storage will exceed 700billion∗∗,andsoftwarespendingwillsurpass∗∗16.7 trillion. However, only 13.6% of North American enterprises​ and 2.4% of Asia-Pacific enterprises​ are able to achieve measurable benefits from the majority of their AI projects, highlighting that value realization remains a key challenge.“To break through the next barrier in AI application, enterprises need to establish enterprise-level AI strategies, build AI-ready workforces, and construct AI-ready technology stacks,” Rick Villars emphasized.Overall, both China and the United States are driving advancements in AI, but with different emphases. The U.S. is a major force behind AI expansion, primarily driven by AI infrastructure and software, while China’s growth is more infrastructure-driven.

Breaking Through the Next Barrier in AI Application: Three Key Directions

When it comes to AI spending, the growth priorities of China and the U.S. in 2026 will vary significantly.Specifically:

  • •In the U.S., efforts will focus on:
    • •Building AI agents​ to automate business processes (expected growth: 48%),
    • •Enhancing network recovery and resilience​ (33%),
    • •And modernizing enterprise data center infrastructure​ (31%).
  • •In China, priorities include:
    • •Modernizing core enterprise applications​ (39%),
    • •Migrating applications from public infrastructure to on-premises infrastructure​ (36%),
    • •And moving applications from on-premises to the cloud​ (34%).

Measurable AI Benefits Remain Limited

Globally, only 13.6% of North American enterprises​ are able to derive measurable benefits from more than 75% of their AI projects. Over the past two years, the average proportion of AI projects that have produced measurable outcomes stands at 47%.Key challenges hindering enterprises from fully realizing the value of their AI investments include:

  • •Resource competition between AI and other IT/digital initiatives​ (36%),
  • •Resistance to the process changes required for AI integration​ (33%),
  • •Regulatory uncertainty affecting AI investment decisions​ (29%),
  • •And difficulty in quantifying and demonstrating ROI to stakeholders​ (28%).

In the Asia-Pacific region, the percentage of enterprises gaining measurable benefits from over 75% of AI projects is even lower, at just 2.4%, with the average proportion of projects delivering measurable results at 38%​ over the past two years.Challenges specific to the region include:

  • •Lack of clarity around ownership and accountability for AI outcomes​ (31%),
  • •Difficulty in quantifying and demonstrating AI ROI to stakeholders​ (30%),
  • •Resource competition with other IT/digital initiatives​ (30%),
  • •And resistance to process changes needed for AI integration​ (30%).

Three Key Directions to Break Through the Next AI Barrier

Rick outlined three strategic directions​ to overcome the next barrier in AI application:

  1. 1.Developing an enterprise-level AI strategy​ to identify core business areas and prioritize transformation.
  2. 2.Building an AI-ready workforce​ by planning and promoting necessary organizational changes.
  3. 3.Constructing an AI-ready technology stack​ by optimizing the technological architecture to support AI and agent workflows.

“Going forward, CIOs will bear the highest responsibility (46%) for driving their companies’ AI transformations, while Chief AI Officers and CEOs account for only about 18% and 16%, respectively,” he pointed out.He added that in 2026, corporate priorities will include:

  • •Identifying key business areas requiring transformation,
  • •Aligning investments and roadmaps with strategic goals,
  • •And establishing core teams to coordinate cross-company initiatives.

Over 1 Billion Active Agents by 2029

“Agent workflows are reshaping the employee lifecycle, and companies need to rethink future work models,” Rick said.By 2026, 40% of jobs​ will involve collaboration with AI agents, redefining traditional roles at junior, mid-level, and senior levels.By 2027, the usage of agents among the Global 2000 companies​ is expected to grow 10x, with invocation loads increasing 1,000x. Agent selection, orchestration, and optimization will become core responsibilities. Without a high-quality, AI-ready data foundation, companies risk a 15% productivity loss​ due to suboptimal performance of generative AI and agentic systems.In terms of agent sources:

  • •57% of U.S. enterprises​ expect application providers to deliver pre-built agents,
  • •While 54% of Chinese enterprises​ prefer the ability to customize agents themselves.

“By 2025, there will be approximately 28.8 million agents​ globally. By 2029, the number of active agents will exceed 1 billion, a more than 40x increase​ from 2025—of which 39% will be unique, low-code/no-code custom agents.”Rick further noted:

  • •In 2025, agents will perform 120 million actions per day; by 2029, this will rise to nearly 217 billion actions per day, a 1,798x increase, with 38% completed by custom agents.
  • •In terms of Tokens/Calls, daily token delivery in 2029 will surpass 3.7 trillion, a 2,626,000x increase​ from 2025, with 40% generated by custom agents.
  • •By 2029, the average token/call delivery cost per agent action​ will be 87% lower​ than in 2025.

“In the new era of intelligence, the focus will be on designing for orchestration, delivering at scale, and governing for trust,” Rick emphasized.

Recommendations for Enterprises

To thrive in this new paradigm, Rick advised enterprises to focus on the following:

  1. 1.Ensure data integrity​ and invest in AI governance, observability, and interoperability.
  2. 2.Embrace modularity and interoperability, collaborating across ecosystems to build open agent frameworks.
  3. 3.Design for scale and sustainability, creating architectures capable of handling the exponential growth in the number of agents, interactions, and token/call demands.
  4. 4.Reassess pricing and delivery models, moving beyond traditional per-seat or per-license models toward outcome-based and usage-driven approaches​ that reflect ongoing autonomous operations.
  5. 5.Adopt a responsibility-oriented approach, establishing safeguards and compliance mechanisms to build trust with both enterprises and the public.

In summary, as AI agents become ubiquitous, organizations worldwide must prepare for a future where intelligent automation reshapes workflows, data strategies, and business models—requiring not only technological readiness but also governance, scalability, and trust at scale.