人工智能创新是不间断的。你的云基础也应该如此。
The no-regret movesto get your cloudAI-ready now There is more cloud ahead than behind The need for a modern, adaptable,cloud-powered digital core 11The cost of standing still 14Three strategic pathways to AI-readiness Move fast to close the cloud gaps 29 Authors Andy Tay Jason Dess Lan Guan Group Chief Executive,Consulting Chief AI & Data Officer Global Lead, Cloud First Andy Taydrives Accenture’scloud strategy, shaping businessinnovation and extendingpartnerships across public andhybrid cloud, data & AI, andsoftware engineering. Jason Desshelps global executivesharness technology to driveinnovation and growth, drawingon over 25 years of experiencein enterprise transformation andperformance management. Dr. Lan Guanleads a globalteam of 80,000 experts helpingorganizations embed AI to drivereinvention, holds over 30 patentand serves on the advisory boardof AI4All. Shalabh Kumar Singh Jefferson Wang Principal Director- Accenture Research Senior Managing Director -Chief Strategy Officer, Cloud First Jefferson Wanghas more thantwo decades of experience atthe intersection of business andtechnology. He aligns businessproblems to use cases andhelps companies make the righttechnology decisions. Shalabh Kumar Singhleadstechnology thought leadershipat Accenture, focusing on topicsrelated to cloud, infrastructureand software engineering, andsustainable technology. There is more cloudahead than behind Many companies treat their cloud journeys as complete once scalabilityand uptime targets are met and modernization checklists are signedoff. But the reality is, there is more cloud ahead than behind. Today’scloud is no longer just “public cloud,” but a flexible, governed fabricthat aligns technology choices to business and industry needs.Meanwhile, AI is accelerating—from classical and machine learningto generative, agentic, ambient and physical. This has redefinedwhat cloud must do to make AI a driver of productivity, growth andcompetitive advantage across an organization. When modern cloudbecomes the foundation of an adaptable digital core, AI can delivermeasurable impact by operating as an integrated system rather thana collection of disconnected initiatives. Every other dimension of theenterprise—strategy and business model, work and workforce—restson this foundation. The expanding definition of cloud To scale AI, you need a modern, resilient digital core that is designedfor continuous change. For most organizations, that foundation iscloud-based. Today’s cloud is not a single destination. It’s a journeythat spans public, private, hybrid-, multi-, sovereign cloud and edge,where workload placement is driven by factors like latency, governmentregulations, risk and economics. It means running the right workloads inthe right places, with governance, security and observability built in, andembracing cloud-native tools and practices. The shift from cloud as a destination to the foundation for AI is alreadyunderway. BBVA, one of Europe’s largest banks, for example, neededa modern infrastructure that could scale globally while meeting localregulatory requirements and unlock agility to respond to clients’changing needs. To achieve this, it launched Analytics + Data + AI(ADA), a cloud-native data and AI platform built from the ground up toserve as a single operating foundation across all its markets. ADA gave BBVA the scale, speed and security to embed AI deeplyinto the enterprise, supporting everything from advanced analytics togenerative AI. By consolidating data across geographies, the platformeliminated legacy complexity and enabled real-time insight. Users canaccess governed, scalable AI and analytics tools, acceleratingdecision-making, improving efficiency and driving the development ofmore personalized client services. The platform also laid the foundationfor decentralized data ownership and enterprise-wide experimentation,giving BBVA a lasting strategic advantage.1 In our analysis of 216 clients’ cloud environments, drawing on a surveyof our senior practitioners working directly with those organizations’cloud and digital cores, only 8% operate at a level similar to BBVA’stoday (See ‘About the research’ for further details.). Meanwhile, inCisco’s recent AI Readiness Index, more than half of respondentsreport that their cloud infrastructure still cannot support AI at scale.This leads to shortcuts that create “AI infrastructure debt”—hiddenpoint integrations, one-off data pipelines, shadow GPU spend,duplicated vector stores, unchecked model drift and manual approvalworkarounds—which compounds every quarter.2 8%of organizationshave clouddedicated toexperimentingwith advancedtechnologies. In our research, cloud estates fell into three categories based onmaturity across cloud migration, observability, data andAI-readiness and automation: Stabilizers are constrained by legacysystems or regulatory requirements that limit cloud migration and slow AI adoption. Optimizers have modernized selectively
