2025年数据、人工智能与分析趋势

infrastructure-related tasks.spglobal.com/451research.Chris Marsh,Research Director, Workforce Productivity &Collaboration and Data, AI & AnalyticsPaige Bartley,Senior Research Analyst, Data, AI & AnalyticsJames Curtis,Senior Research Analyst, Data, AI & AnalyticsMelissa Incera,Research Analyst, Data, AI & AnalyticsAlexander Johnston,Senior Research Analyst,Data, AI & AnalyticsKrishna Roy,Senior Research Analyst, Data, AI & Analytics2025 Trends inData, AI & Analyticsspglobal.com/marketintelligence Table of contentsExecutive summaryIntroductionAbout this reportThe TakeTrends we anticipate in 2025Trend 1: ROI pressure will increase for GenAI as costs mountTrend 2: Rich media generation will find its nicheFigure 1: Purchase or development of GenAI capabilities to create outputsTrend 3: AI governance will be business-criticalTrend 4: Agentic AI will rise in importanceTrend 5: Governance will be required for expanding agentic AI architecturesFigure 2: Slim majority of organizations using AI are engaging in AI governancepractices/initiativesTrend 6: Inferencing efficiency will become an industrywide pursuitFigure 3: Inferencing is becoming more demanding for organizationsTrend 7: Models become more specialized and domain-specificTrend 8: Synthetic data practices mature in the era of GenAITrend 9: Unstructured content will experience a reckoning for privacy and securityFigure 4: Reliance on archive storage for sourcing data to build modelsTrend 10: GenAIOps will take flightTrend 11: Vector support will become ubiquitousFigure 5: Vector database market revenue forecast ($M)Trend 12: Optionality will be coming soon to RAG environmentsFigure 6: Use of RAG techniquesMethodologyFurther readingAbout the authors 3334445566778899101011121213131414152025 Trends in Data, AI & Analytics| 2spglobal.com/marketintelligence Executive summaryIntroductionAs we approach 2025, the data, AI and analytics landscape is poised for significanttransformation. Organizations increasingly recognize the complexity involved indesigning and managing the data infrastructure necessary to support machine learning(ML) at scale. Many are recalibrating their expectations and placing greater emphasison evaluating their initiatives’ performance and return on investment (ROI). This shiftacknowledges the rising costs associated with AI projects, particularly as organizationsmove toward more sophisticated applications. Further, achieving enterprise-gradeimplementation at scale requires not only innovative technology but considerableexpertise and governance.The trends we highlight provide a nuanced understanding of the challengesorganizations face as they prepare for increased sophistication in their data andapplication strategies. Our research insights serve as a guide for organizations strivingto leverage these technologies effectively and for technology enablers seeking tounderstand these evolving market dynamics.About this reportReports such as this showcase insights derived from a variety of market-level researchinputs, including financial data, M&A information and other market data sources bothproprietary to S&P Global and publicly available. This input is combined with ongoingobservation of markets and regular interaction with vendors and other key marketplayers.This report specifically includes data from the following sources:–Voice of the Enterprise: AI & Machine Learning, Use Cases 2024— This web-basedsurvey was fielded November 14 – December 14, 2024, among approximately 1,001 ITend-user decision-makers worldwide.–Voice of the Enterprise: AI & Machine Learning, Infrastructure 2024— This web-based survey was fielded May 10 to June 4, 2024, among approximately 712 IT end-user decision-makers worldwide.–Voice of the Enterprise: Data & Analytics, Data Architecture for AI 2024— Thisweb-based survey was fielded March 5 to April 24, 2024, among approximately 1,112 ITend-user decision-makers worldwide.–Data Platforms Market Monitor & Forecast 2024— This report leverages 451Research’s deep knowledge of and relationships within the data platforms market,resulting in a proprietary forecast based on a hybrid bottom-up analysis of 174vendors’ current revenue and growth expectations through 2028.–Generative AI Market Monitor & Forecast— This report leverages 451 Research’sdeep knowledge of and relationships within the generative AI market, resulting in aproprietary forecast based on a hybrid bottom-up analysis of 398 vendors’ currentrevenue and growth expectations through 2028. Trends we anticipate in 2025Trend 1: ROI pressure will increase for GenAI as costs mountWhile experimentation with GenAI spiked in 2024, 2025 will be the year organizations startto meaningfully assess the performance of their initial investments. With demand drivenfrom the top, organizations have demonstrated significant appetite for GenAI, but manycompanies have been challenged by a lack of expertise and the technical complexity ofdeploying enterprise-grade projects at scale, whic