2026全球人工智能效能研究中国报告

PwC Global AI Performance Study China Report Foreword As the integration of Artificial Intelligence (AI) with the real economy deepens, AI has emerged asa pivotal engine driving the transformation of the digital economy and a cornerstone for building According to PwC’s 2026 AI Performance Study, intense AI adoption has not yet translated intomeasurable returns for many enterprises. At this stage, value remains highly concentrated What sets these AI leaders apart from other businesses, and where do Chinese enterprises standregarding their AI maturity? We conducted a comprehensive evaluationutilisingthe AI FitnessIndex, to measure the performance of each enterprise. Our research reveals that successful AIleaders are able to precisely pinpoint pain points, construct tailor-made infrastructure, and deeply By benchmarking the AI fitness indices of AI leaders, Chinese enterprises, and others, this reportexamines how leading companies translate AI development into performance growth. It alsoanalyses the competitive strengths and structural bottlenecks shaping the AI development of clearly shows that systematic AI deployment candeliver significant value for enterprises. China’s AIapplication ecosystem is vibrant and dynamic,demonstrating remarkable progress in cross-industryintegration, with implementation practices that rankamongst the most advanced globally. As China’s AImarket continues to develop, it is giving rise to a wide Hemione HudsonChair and CEO, PwC China KeyfindingsforChineseenterprises 1.Chinese enterprises1lead the world in AI fitness, surpassing the median values of global AIleaders across the majority of metrics. 2.Strong advantage in industry convergence: Chinese enterprises exhibit significantly higherpenetration of AI applications related to sector convergence—including cross-enterprisecollaboration, responsiveness to shifts in customer needs, and unlocking new value fromcross-sector ecosystems—than global AI leaders. In particular, the penetration rate of cross- 3.Superior execution-layer implementation efficiency: Execution-oriented AI applications, suchas process automation and replacement of standardised operations, slightly outperform those 4.Gaps remain in closing the value loop: Although most enterprises have begun to focus on theactual business value of AI and have established strategic plans that combine short-and long-term horizons, their capabilities in systematically tracking AI’s business impact and executing 5.Lower efficiency in converting innovation into value: Investment in innovation experimentation infrastructure exceeds that of global AI leaders, yet there is a lack of mechanisms to expandfrom individual pilots to large-scale rollout. As a result, innovation outcomes tend to remainconfined to the pilot stage, limiting the long-term compounding returns on investment. The 6.Trust and governance require further strengthening: The foundation of enterprise trust in AIremains relatively weak. Employees generally remain cautious about acting on AI-generatedinsights and have not yet integrated them into day-to-day decision-making to the same extent Contents 01AIperformancestudy forglobalenterprises 05 I.Themost AI-fitcompanies aregetting a 7.2times AI-drivenperformanceboostovertheirpeers II.What is AI Fitness Index? III.Characteristics ofglobal AIleaders 02Analysis of Chineseenterprises’AI Fitness Index 13 I.Chineseenterprisesrank in theglobaltoptier for AIfitness II. AI use: outstanding performance, with growth-oriented applications andscaling exceeding global benchmarks;autonomous and self-optimising III. AI foundations: reaching global upper-tier levels through high investmentand strong infrastructure; innovation and execution-to-scale conversion 03The potential and outlook for AI performance 34 I.Comparison withglobal AIleadersII.Characteristics of AIperformancedevelopmentamongenterprises in AI performance study forglobal enterprises As enterprises accelerate their AI investments, business leaders often find thatthese investments do not always translate into superior financial performance.This global AI performance study shows that the value of AI is currently Among the 1,217 enterprises surveyed globally across 25 industries, the top20% of enterprises account for 74% of all AI-driven performance. I. The most AI-fit companies are getting a 7.2times AI-driven performance boost over their To understand why some companies are seeing real returns while most are not, we benchmarkedthe surveyed enterprises on their AI-driven financial performance, defined as the revenue andefficiency gains derived from AI and adjusted so each company could be compared against its We also asked senior executives at these companies about their engagement in 60 areas of AImanagement and investment practice to test those areas’ effects on AI-driven financialperformance. We grouped these practices into nine factors related to the ways in which Note:Unless otherwise stated