2026技术趋势：私营公司的预测智能

Predictive Intelligenceon Private Companies Access 10 days offreepredictive intelligencefor private marketstoday. ✓Run scouting reports with full SWOT analyses on more than10M companies ✓AskChatCBIwhat your next strategic move should be✓Get personalized insights on how to react to market shifts ▮▮▮IIIIIIStart free trial Table of Contents Enterprise&work Industrial&national infrastructure AI agents make ROI a moving targetThe back office runs itself915 Data centers become grid assetsSovereign AI creates regional AI ecosystemsThe commercial space race enters a new eraDefense AI startups mobilize for mass production60667381 Private becomes the new publicStablecoinsgo corporateThe storefront disappears212835 PhysicalAI&robotics AI learns how the physical world worksRobots learn to work together8895 Health&bio Healthcare puts voice AI to the testPharma’s path to self-driving labsThe unbundling of employer health plans414753 Did we hit the mark? Rating our 2025 tech trends (report published 11/19/2024) Did we hit the mark? Rating our 2025 tech trends (report published 11/19/2024) CB Insights Predictive Intelligence Each trend in this report combines CB Insights’ proprietary signals across company fundamentals and marketbehavior to show where innovation, capital, and customer momentum are moving next. Enterprise& work Enterprise&work AI agents make ROIa moving target New tools help companies keep up with howagent impact actually shows up ENTERPRISE&WORK Cost savings and time tracking dominate early AI agentmeasurement efforts Organizations track efficiency metrics while revenue impact remains difficult to quantify Leading KPIs used totrack AI agent success What KPIs do you track forAI agent success? Select allthat apply. ENTERPRISE&WORK Vendors have an opening to define agent ROI frameworks With deployments in the early stages, organizations still rely on informal, anecdotal assessment approaches “At present [ROI] is being monitored atboard level, where they are tracking spendon AI agents.How the effectiveness isbeing measured then filtersup, butis nota precise science and fully trackedasyet.” “We are still struggling with this–[measuring AI agent ROI] seems to be amajor issue for all of these tools.” Director of Information Security,$1B revenue media & entertainment company VP of ApplicationDevelopment,Fortune 500finservcompany ENTERPRISE&WORK Software engineering analytics platforms establishmeasurement approaches in coding Companies like Span move to provide clear ways to measure agentic AI impact across software engineering workflows,a leading area of AI agent usage and spend Span targets AI code measurement as engineeringteams navigate development transformation Softwareengineeringanalyticsplatforms Span launched span-detect-1 in September 2025, aproprietary model thatdetects AI-generated code at thechunk level and tracks it through the developmentlifecycle, differentiating from competitors that rely onvendor data or surveys. $25M in Series A funding will deepen capabilities formeasuring ROI of AI-augmented developmentand buildtoward answering complex engineering questions inseconds rather than days. ENTERPRISE&WORK While process intelligence platforms evolve to tacklemeasurement across enterprise workflows New breed of solutions help enterprises identify agentic automation opportunities, track investments, andquantify impact beyond engineering teams WorkhelixScouting Report Key takeaways •Elite AI investor backing:The company secured$15.3M Series A from top-tier investors including ReidHoffman, YannLeCun, Jeff Dean, and MiraMurati,signaling validation of its approach to enterprise AImeasurement and optimization. •Strategic Accenture integration:Integration withAccentureLearnVantageand participation in ProjectSpotlightprovides direct access to enterprise clientsstruggling with AI ROI measurement. ENTERPRISE&WORK Billing platforms emerge to help companies monetize AI agentactivity Vendors experiment with pricing models (usage-based, outcome-based, and hybrid) as agent monetizationapproaches remain unsettled Scouting Report Paid buildsbilling infrastructure for AI agents, enablingoutcome-based pricing modelsas companies move away fromtraditional seat-based SaaS pricing. Enterprise&work The back officeruns itself Enterprises move from tool-assisted workto autonomous execution in core operationalworkflows ENTERPRISE&WORK Agent platforms have crossed from pilots to production As AI agent companies reach deployment maturity, enterprises begin automating routine operational work ratherthan experimenting with copilots ENTERPRISE&WORK FinancialServices Financial servicespulls ahead in AIagent deployment 118 relationships21% share The industry’s document-heavy, rules-based workflows make it an ideal provingground for autonomous systems Industrials10619% ▮▮▮▮▮▮▮▮▮▮▮▮▮▮IIITrack1,900+ AI agent companies Legal8715% ENTERPRISE&WORK Withinfinserv, agents