JULY2025

MAPPINGTHEAIECONOMY:

WHICHREGIONSAREREADYFOR

THENEXTTECHNOLOGICALLEAP?

MarkMuroandShriyaMethkupally

Tableofcontents

Executivesummary 2

Introduction 4

Background 5

Approach 6

Data 7

WhohasAItalent? 7

Whoisinnovating? 8

WhoisactuallyadoptingAI? 8

Groupingmethod 10

Findings 11

Finding#1:Inaggregate,thenation’sAIenterpriseisgrowingrapidly,thoughitremains

modestinsize 11

Finding#2:Thenation’smainAIenterpriseisconcentratedinalimitednumberofmetro

areas,butnumerousotherregionsarehometomeaningfulAIactivity 20

Finding#3:CurrentandemergingregionalAIperformanceisinformedbyvariedlocalmixes

ofsuccessfactors 26

Discussion 35

ThenationshouldbuildastrongAI-supportplatform 36

Regionsshouldengageinstrategicclusterdevelopment 40

Conclusion 53

AppendixA:Technicalbackground 54

References 60

Acknowledgements 63

Endnotes 64

MappingtheAIeconomy:Whichregionsarereadyforthenexttechnologicalleap?

MarkMuroandShriyaMethkupally

Executivesummary

ArtificialintelligenceisrapidlytransformingtheU.S.economy.However,regional

capacitiesforAItalentdevelopment,research,andenterpriseadoptiondifferdramaticallyacrossthecountry,andpolicymakershavelittleinsightonwherethesegapsare

emerging.

Thenation’sreadinesstobenefitfromAIiscriticalbecausethetechnologyisgoingto

playasignificantroleineconomicdevelopmentgivenitspotentialtodriveinnovationandproductivityineveryindustry,bothingeneralandwithinregions.

Toaddresstheseissues,thisreportmapstheunfoldingAIrevolutionacross387U.S.

metropolitanareasandassesseshowthenationandlocaleconomiesarepositionedto

create,apply,andharnessAI.ThereportfirstunderscoresAI’simplicationsfornational

andregionaleconomicdevelopment.Thenitturnstoabenchmarkingexercisethat

groups14indicatorsintothreedimensionsofregionalAIreadiness:talent,innovation,andadoption.The“talent”pillarmeasurestheflowofAI-capableworkers;“innovation”

capturesresearchandinnovationstrengths;and“adoption”chartsindustryuptake.

Employingthisframework,thereportgroupsU.S.metroareasintosixtiersofAI

economicreadinessand,insodoing,illustratesthegeographyofAIprogressasitvariesacrossthenation.Continuingapatternfroma2021Brookingsreport,1themajorAI

communitytypesrevealedbythenewclusteranalysisare:

?Superstars:TheSanFranciscoandSanJose,Calif.,metropolitanareasexhibitunmatchedstrengthacrossallthreeAIpillars(talent,innovation,andadoption).

?StarHubs:Thisgroupof28metroareasformsasecondechelonofuniformly

strongAIecosystems,balancingtop?tiertalent,research,andenterpriseuptake.

?EmergingCenters:Thisgroupof14metroareascombinestopperformanceintwopillarswithonedevelopingarea.

?FocusedMovers:These29metroareasexcelinonepillarwhilemaintainingfoundationsintheothertwo.

?NascentAdopters:Thisgroupof79metroareasshowsmoderateperformanceacrossallthreepillars.

?Others:Agroupof43metroareasthatcurrentlylagsonmultiplepillars.

Bytrackingtheclusters’progress,thereportdevelopsseveralmainfindingsthatshow:

2

3

?ArapidlygrowingbutnascentAImarket.Between2010and2025,AI-relatedjobpostingsgrewatanannualrateof28.5%,yetstillcompriseonlyabout2.5%ofallU.S.jobopenings.

?Extremeconcentration.TheBayAreaaccountsfor13%ofnationaljobpostingsfeaturingAIskills,andtogetherwiththeStarHubs,thisgroupof30top-

performingmetroareascaptures67%oftotalAIjobpostings.

?Signsofgeographicdiffusion.Severalnoncoastalmetroareas—including

Pittsburgh,Detroit,Madison,Wis.,andHuntsville,Ala.—nowrankinthetop

quartileonatleasttwopillars,indicatingemerginghubsbeyondthetraditionaltechcoasts.

?Persistentopportunitygaps.OverhalfofU.S.metroareasremaininthebottomtworeadinesstiers,revealingsignificantshortfallsintalentpipelines,research

infrastructure,andenterpriseadoption.

Thereportconcludeswithacallforatwo-prongedpolicyagenda.First,thisagendalaysoutarobustnationalAIplatformthatboostsnondefenseR&Dfunding;expandssharedresearchandcomputationalresources;acceleratesAIclusterscale-up;streamlines

innovationinfrastructureinvestments;andfundsAIcurriculumdevelopmentandresearchathighereducationinstitutions.Second,thereportarticulatesaregion-by-region

strategythatbeginswithregions’individualstartingpointsandusesthemtoshapelocalAIresearchagendas,fosterregionalclusterdevelopment,andbuildlocaltalentinwaysthatareorientedtolocalneeds.

Ultimately,theregion-tailoredprioritiesvarydependingonaregion’sdegreeofAIreadiness,asfollows:

?Superstars:Supportemergingtechcompanies,maintainappealtoimmigranttalent,investheavilyinlocaltecheducation,andconsideroptionsforworker-transitionsupport.

?StarHubsandEmergingCenters:Investindevelopingregionalclusters,increaseaccesstohigh-speedandaffordablecomputingresources,andintensifyeffortsintecheducation.

?FocusedMovers:Leanintosignaturestrengths,investinthecomputing

infrastructurenecessarytotrainandretaintoptalent,prioritizetechtransferand

commercialization,andleveragelocalbusinessenvironmentstopromoteadoption.

?NascentAdoptersandOthers:PromotebroadAIliteracy,demonstratepracticalAIapplicationsinroutinetasks,andthinkaboutAIcareerpathways.

Insum,theemergenceofAIasageneralpurposetechnologypresentsaninflectionpointforregionaleconomicdevelopmentintheUnitedStates.Leadersshouldmoveurgentlytopromotelocaldevelopmentthatwillcontributetomoreevenlydistributedand

transformativeAIgrowthfromcoasttocoast.

4

Introduction

Dramaticadvancesingenerativeartificialintelligencehavesparkedintensediscussions

aboutgeopoliticalcompetition.WilltheUnitedStates“win”theglobalracetodevelopanddeployAI,orwillChina?Wherewilladoptionoccurfirst?

Techvisionaries,securityexperts,andpoliticiansallinsistthattheU.S.mustout-prepareandout-innovateChinatoensurethatthenationmaintainsitssecurity,maximizesits

productivity,andunleashesthepotentialofitstalent.

YetthereisanotherquestionabouttheAIreadinessquest:HowisAIcreationandadoptionunfoldinghereathome?HowisitproceedingwithintheU.S.,andnamely,amongandbetweenthenation’shundredsofregionaleconomies?

Theanswerstothesequestionsmatterimmensely.Aswithpreviouswavesof

technologicaltransformation,howmuch,howeffectively,andinwhatspecificways

regionsharnessAI—especiallyintheircoreindustries—willbehugelyimportantin

shapingthegeographyoftheU.S.economyanddeterminingwhichregionsprosperandwhichdonot.

Inthisregard,AIhasemergedasa“generalpurposetechnology”—onewithfar-reachingconsequencesforindustries,places,andpeople.2AIsystemspromisetodrive

productivitybyautomatingroutinetasksandatothertimesaugmentingwork,allowinghumanstofocusonhigher-valueactivities.AIisacceleratingthepaceofdiscoveryandinnovationbyanalyzingvastdatasetsandidentifyingpatternshumansmightmiss.Andforthatmatter,AIenablesmoreefficientresourceallocationthroughintelligent

forecastingandoptimization.Assuch,AIcouldheavilyinfluencethenation’sabilitytoachieveitslargergoals,whetheritbethroughfasterdrugdevelopment,personalizedlearning,or“virtualemployees”optimizingsupplychaincomplexities.

YetmanyoftheseachievementscouldbejeopardizediftoolittleAIdevelopmentoccurswithtoolittlethoughtintoofewindustries.

AllofwhichsuggeststhatitmattersalotwhetherandwhichU.S.citiesandregionsarereadytofacilitateAIdevelopmentinhigh-qualityways,andarethusdemonstratinga

readinesstotrulybenefitfromfutureAIbuild-out.

Whichiswherethisreportcomesin.AsAI’spotentialtospurgrowthandtransformeconomiesbecomesincreasinglyclear,localleaderswanttounderstandwheretheirregionsstand.Specifically,theywanttogetahandleonthetypesofcapacitiesthatmatterinAIpreparednessandtheirlocalstatusonthosesuccessfactors.

Andso,thedataandassessmentsprovidedhereseektofillavoid.FollowinguponanearlierBrookingsreportonthegeographyofAI,thediscussionhereprovidesauniquebenchmarkingofU.S.metropolitanareasontheirstandingacross14metricsreflectingthreepillarsofcommunities’readinesstointegrateAIintoeconomicdevelopment.3

5

Assuch,thefollowinganalysisprovidesleadersbasicinformationonwhatmattersinestablishingAIreadinessasalocaleconomicdevelopmentpriority,andhowindividualregionspresentlyfareonthosepriorities.Atthesametime,theanalysisgroupsU.S.

metroareasintosixtiersofAIreadinessand,indoingso,providesanewviewofthegeographyofAIprogressasitvariesacrossthenation.

Whatdoesthatnewviewshow?Wefindbothcontinuednarrownessandadegreeofrecentdiffusion.Tobesure,thegeographyofAIadoptionprogressremainshighly

concentratedinashortlistof“Superstar”metroareas(SanFranciscoandSanJose,

Calif.)and“StarHubs”(suchasSeattle;Boston;Austin,Texas;andWashington,D.C.).

Therapiddiffusionofchatbotsinrecentyearshasnotchangedthat.Yetatthesame

time,thechatbotboomandnowthespreadofagenticAIisclearlybeginningtowidenthegeographyofAIactivityintheU.S.Inthatvein,thepresentanalysishighlightsseveral

additionalregions(includingPittsburgh;Rochester,N.Y.;Detroit;Tampa,Fla.;Madison,Wis.;andHuntsville,Ala.)thatarenowstrongontwoofthethreepillarsofadoption

readiness.

Wenotonlyreportontheprogressofindividualmetropolitanareas,butalsoonthe

integrationofregionaleconomiesintothenationalenterprise.Tostart,thereportreviewsbroadAIadoptiontrendsacrossthenation,introducesitsbenchmarkingapproach,and

outlinesaseriesoffindingsaboutthevariationandgeographicaldistributionofAI

activitiesinU.S.places.Afterthat,thereportreflectsonthetrendsandreviewsthetypesofstrategiesthenation,states,andregionscouldpursuetoaccelerateAIadoption

consideringtheircommunities’currentAIassetsandcapabilities.Ultimately,thegoalistohelpleaderspromotelocaldevelopmentthatwillcontributetomoreevenlydistributedAIgrowthfromcoasttocoast.

Background

Thenation’sreadinesstobenefitfromAIiscriticalbecausethetechnologyisgoingtoplayasignificantroleineconomicdevelopmentgivenitspotentialtodriveefficiency,innovation,andproductivityineveryindustry,bothingeneralandwithinregions.4

Overall,AIreadiness—asanemergent,innovation-driventechnology—willdependonthenation’sabilitytodeliveronthreesuccessfactors:

?TheavailabilityofabundantAItalent,sincetalentclustersarecriticalin

generatingself-reinforcingeconomicgrowthforpeople,firms,andplaces.5

?TheaccessibilityofAIinnovationandinnovationinfrastructure,sincetechnicalprogressplaysadisproportionateroleineconomicgrowthandbuildsonitself.6

?ActualadoptionofAIbyorganizations,becausebroadtechnologyadoption

remainsanimportantdriverofgrowthinproductivityandlivingstandards.7

6

Thisbenchmarkingtakesthesekeydimensionsoftechno-industrialcompetitivenessascore“pillars”ofAIreadiness,thoughitshouldbenotedthatsafety,equity,and

accountabilityareequallyimportant.MakingsurethatAIisdevelopedandadoptedresponsiblyiscriticalforlong-termsuccess.8

Atthesametime,AIadoptionattheregionallevelmattersequallyforeconomic

development,prosperity,andtheflourishingofcommunities.Assuch,individualplacesmustpayattentiontothelocalpresenceofthethreeAIreadinesspillarstoensuretheirsuccess.

Thenation,states,and“bigtech”actorsalsoneedtoconsiderthebroadergeographyofAIdiffusionacrossthenation.AI,afterall,verymuchreflectsthetendencyofemergingdigitalindustriestoclusterintenselyinashortlistoflarge,tech-orientedhubs,as

Brookingshasdescribedinearlierreports.9Inthatfashion,strongclusteringhaslong

beenviewedasapowerful“X-factor”forhigh-techgrowthintheU.S.,asscholarssuchasMichaelPorter,EdwardGlaeser,andEnricoMorettihaveargued.10

Recentanalysishasincreasinglysurfacedconcernsaboutthenation’sconcentratedhigh-techgeography—concernsthatcouldwellpertaintothegeographyofAIadoption.Inthisfashion,AIepitomizesthekindofunevenbuild-outeconomistNicholasBloomhas

trackedamongemergingtechnologies.In2021,forexample,Bloomexaminedthe

geographicalspreadof29disruptivetechnologiesandmappedagradualdiffusionof

lower-skilljobsawayfromtheiroriginal“frontier”hubs,yetalsoapersistentcentralizationofthehighest-valuejobsinthosehubs.11

Thesefamiliar“winner-take-most"patternsofhigh-techindustriesraiseimportant

questionsaboutthenatureanddistributionofAIactivitiesacrosstheU.S.TrackingbothlocalandbroadergeographicalpatternsofAIreadinessisnecessarytoassessthehealthoftheindustry.

Approach

TomapthestrengthsandgeographyoftheU.S.AIenterprise,thisreportundertakesto

benchmarkthevariationandintensityofkeysignalsofAIeconomicdevelopment

readinessinregionsacrossthenation’smajormetropolitanareas.Usingcross-sectional

datacollectedonvarioustypesofAIassetsandcapabilitiesatthemetropolitanlevel,theassessmentemploysasimplegroupingmethodtoidentifyindividualmetroareasand

groups(alsoreferredtoas“clusters”inthisreport)ofmetroareas’readinesstoengageintheAIenterprise.

Inthisvein,thebenchmarkingaimstoprovidelocalleaders—includinggovernment

officials,businesses,nonprofitintermediaries,andcivicorganizations—withimportantinsightsforharnessingAItodriveregionalgrowth.Ametro-area-levelassessmentisparticularlyvaluableasitprovidesintelligencethatcaninformtargetedstrategiesand

7

initiativesthataddresstalentandworkforcedynamics,innovationneeds,andbusinessdevelopmentatthelocallevel,wheretheywilllikelybemosteffective.

Data

Inordertocarryoutthisanalysis,metroareadatawerecollectedforallU.S.metroareasacrossthreekeypillarsofregionalAIreadiness:1)localAItalent;2)localAIknowledge

creationandinfrastructure;and3)actuallocalbusinessadoptionofAIapplications.

Together,thesecategoriesofdatacollectionreflecttheessentialelementsofarobustAIecosystem.TheycapturetheavailabilityoftheskilledprofessionalsneededtodriveAI

techinnovationanddeployment;theresearcheffortsandhardwarethatadvancethe

field;andbusinesses’currentengagementandreadinessforadoption.Analysisoftheseinterconnectedfactorsyieldsvaluableinsightsintoeachregion'sAIreadinessandis

usefulforidentifyingstrategicopportunitiesforinvestment.

Asagroup,thesemetricsrepresentareasonedselectionfromamongthelimitedarrayofavailablelocalstatisticsusefulforapproximatingplaces’standingonparticular

capabilities.Inmostcases,nosinglemeasure—suchasthenumberofresidentswith

computersciencedegrees—capturesthefullspectrumofrelevantcapacity,whichinthisexamplemightalsoberepresentedbylessavailablecredentialsorskillsgainedthroughexperience.

Alongtheselines,thesectionsthatfollowask:

WhohasAItalent?

AnabundantsupplyofAI-capableprofessionals(andothertypesofskilledworkers,notallwithprofessionaldegrees)isessentialfordrivinglocalinnovationanddeployment,asskilledtalentaffectsaregion’scapacitytodevelopandsupporttechnologydevelopmentandtheadoptionofAItechnologies.Regionswithqualityeducationprogramsthat

producenumerouscomputersciencegraduatesarealsomorelikelytoattractinvestmentandsupportemergingAIstartups.Additionally,awell-trainedworkforcehasthepotentialtofostercollaborationbetweenacademiaandindustry.

SelectedindicatorsofAItalentinclude:

?Computersciencebachelor,sdegreeholders.Censusdataonthenumberof

computerscience,engineering,andmathematicsgraduatesineachregionreflectthesupplyofprofessionalswhocansupportAIinnovationandadoption.

?ComputersciencePh.D.holders.Censusdataonthenumberofcomputer

science,engineering,andmathematicsPh.D.graduatesineachregionhighlighttheavailabilityofadvancedAItalenttohelpdevelopnewtechnologies.

8

?AIjobprofiles.StatisticsfromLightcastthattrackthevolumeofU.S.worker

onlineprofilesmentioningAIskillsindicatethesupplyofAItalentacrossdifferentregions.

Whoisinnovating?

ThecreationofAIknowledgethroughresearchanddevelopment(R&D)activityplaysanimportantroleinadvancingAIlocallyandnationally.Suchactivitynotonlyengages

criticalAItalent,butalsoenhancesaregion’sabilitytogenerateknowledgeandconvertitintopracticaltechnology.Innovationcapacitymayalsoincludeinnovationinfrastructure

suchascomputinghardwareandprograms.

SelectedindicatorsoflocalAIknowledgecreationinclude:

?FederalAIR&Dcontractspending.FundingdatafromUSAcapturesfederalinvestmentinAIR&D,highlightingfinancialsupportforinnovation.

?AcademicpapersattopAIconferences.DatafromAIRankingsonAIresearch

paperspublishedattopAIconferenceshighlightlocalcontributionstoadvancingthescienceofAI.

?AIpatents.DatafromtheUnitedStatesPatentandTrademarkOffice’s(USPTO)AIpatentdatasetsuggesthoweffectivelyresearchisbeingtranslatedintonew

technologies.

?High-performancecomputing(HPC)usagefromacademicusers.Metricson

HPCusageinacademicandresearchsettingsthroughtheNationalScience

Foundation’s(NSF)ACCESSprogramreflecttheextenttowhichlocalresearchersareaccessingadvancedcomputingresourcestoaccelerateAIresearch.

WhoisactuallyadoptingAI?

Understandinglocalfirms’currentadoptionofAIintheiroperationsisessentialfor

assessingaregion’soveralltechnologyengagement,creativity,competitiveness,andpotentialforproductivitygainsandgrowth.

SelectedindicatorsofAIadoptioninclude:

?AIstartups.DatafromCrunchbaseonyoungentrepreneurialcompanies

specializingincreatingAI-drivenproductsandservicesdepictlocalbusinessdynamismandtheconditionsforlaunchingAIinitiativesintheregion.

?FundingforAIstartups.DatafromPitchBookonventurecapital(VC)fundingraisedbyAIstartupsmeasurethecapitalavailableforentrepreneurialadoptionefforts.

?Enterprisetechadoption.DatafromtheNSFAnnualBusinessSurveymeasuringtheadoptionofAItechnologiesshowhoweasilycompaniesareintegratingAI.

9

?Enterprisedatareadiness.DatafromtheNSFAnnualBusinessSurveymeasurethedigitizationofcoreactivitiesnecessaryforAIintegration.

?Enterprisecloudreadiness.DatafromtheNSFAnnualBusinessSurveymeasurethecloudreadinessnecessaryforAIintegration.

?AIjobpostings.DatafromLightcastmeasuringjobpostingsrequiringAIskillshighlightthedemandfortalentandindicatearegion’sreadinesstoembraceAItechnologies.

?OccupationalexposuretoAI.DatafromOpenAIontheshareofanoccupation’stasksthatcouldsoonbeexposedtoChatGPT-4speaktothegeneralinvolvementofaregioninAIadoption.

Table1.Metro-area-levelindicatorsandtheirsources

Category

Data

Source

Description

WhohasAItalent?

Numberof

computersciencebachelor’s

degrees

U.S.Census

Bureau:AmericanCommunitySurvey(ACS),2023

Bachelor’sdegreein

scienceand

engineering,computers,mathematicsand

statistics,orrelatedfields

Numberof

computersciencePh.D.degrees

NSF:Higher

Education

Researchand

Development

(HERD)Survey,2023

Part-timeandfull-time

post-doctorate

researchersincomputerandinformationscience,mathematicsand

statistics,and

engineeringfields

JobprofileswithAIskills

Lightcast,2024

Onlineprofilesof

workerswithAIskills

thatstartedjobsin2024

WhoisdrivingAIinnovation?

FederalAIR&D

contractspending

USA,2024

FederalAIR&Dcontractspending

Publications

presentedattopAIconferences

AIRankings,2024

AIacademic

publications

AIpatents

USPTO,2023

AIpatents

High-performancecomputing(HPC)

NSFACCESS,2024

HPCusage

10

usagefrom

academicusers

WhoisreadytoadoptAI?

AIstartups

Crunchbase,2014to2024

AIstartups

VCfundingforAIstartups

PitchBook,2023to2024

FundingdealsdonebyAIstartups

Enterprisedatareadiness

U.S.Census

BureauandNSFAnnualBusinessSurvey,2021

Firmswithbusiness

activityprimarilyin

digitalformatper100firms

Enterprisecloudreadiness

U.S.Census

BureauandNSFAnnualBusinessSurvey,2021

Firmswithbusiness

activityprimarilyincloudformatper100firms

EnterpriseAIadoption

U.S.Census

BureauandNSFAnnualBusinessSurvey,2022

FirmsusingAIper100

firms

AIjobpostings

Lightcast,2024

OnlinejobpostingswithAIskills

LocaloccupationalexposuretoAI

OpenAI,2022

ShareofjobsexposedtogenerativeAI

Groupingmethod

ToevaluateandgroupU.S.metropolitanareasbasedontheirrelativestrengthin

AI?relatedcapabilities,thisreportdevelopsasimpleclassificationapproach.For387

metroareas,wemeasureAIcapacityacrossthethreehighlightedpillars(localtalent,

innovation,andbusinessadoption),andscalethemfrom0to1usingmax-min

normalization.Wethenrank195metroareaswithpopulationsgreaterthan250,000in

eachpillar.Thoseinthetop25%onapillarearna“T”(for“top”)forthatpillar.Thoseinthemiddle50%garneran“M”(for“middle”performance).Andthoseinthelowest25%earna“B”(for“bottom”).Inthisway,wecreategroupprofiles(e.g.,“TMB”)thatreflectaplace’sstrengthsandweaknesses.

Thethree-lettercombinationsarethenusedtoidentifysixgroupsofmetroareas.Thistaxonomyallowsforthecreationofmetroareaandgroupprofilesthatcanhelp

11

policymakersandinvestorspinpointwheretodirecttraining,funding,orsupporttoboostAIreadiness.

Table2.‘TMB’taxonomyofmetropolitanareas

Group

Definitions

Superstars*

SanFranciscoandSanJose(allthreeTs)

StarHubs

Strengthacrossallpillars(allthreeTs)

EmergingCenters

Strengthacrossatleasttwopillars(atleasttwoTs)

FocusedMovers

Focusedstrengthinonepillar(atleastoneT)

NascentAdopters

Mediumstrengthinatleasttwopillars(atleasttwoMsandnoT)

Others

Lowstrengthinatleasttwopillars(atleasttwoBsandnoT)

*SanFranciscoandSanJoseareplacedinaseparategrouptoreflecttheirhighperformanceacrossallthreeofourclassifyingpillars.ThisdecisionalsoensurescomparabilitywiththeBrookingsInstitution's“GeographyofAI”report,whichalsotreatedtheBayAreaasanoutlierduetoitsoutsizedlead.

DataandadetailedmethodologyareavailableinAppendixA.

Findings

Analysisofthe14measuresofAIeconomicactivityacrosstheU.S.yieldsseveralbroadfindingsabouttheindustry,itscharacteristics,anditsgeography.Thefollowingsectionsdetailthreecorefindings.

Finding#1:Inaggregate,thenation’sAIenterpriseisgrowingrapidly,thoughitremainsmodestinsize

Aggregatestatisticsthatbroadlyexaminethenation’sAIenterpriseprovideagaugeofthesector’sgrowthandcurrentsize.

OneoftheclearestindicatorsofAI’sexpandingfootprintintheU.S.economyistherisingshareofbusinessesreportingcurrentoranticipateduseofAItechnologies.AccordingtotheU.S.CensusBureau’sBusinessTrendsandOutlookSurvey(BTOS),theshareof

businessescurrentlyusingAIintheproductionofgoodsorservicesrosefromaround4%inearly2023to8.7%bymid-2025.Expectationsfornear-termAIusearehigher:Asof

thelatestsurvey,11%offirmsreportedthattheyplantoadoptAIwithinthenextsixmonths(seeFigure1).Thesetrendspointtosteadygrowthinenterprise-levelAI

integration.

12

Figure1.

JobpostingsthatnameAIskillsserveasanothersignalofAIadoption,eveniftheydonotdirectlyequatetoemploymentlevels.Inanyevent,LightcastdatareportthatAI-specificjobpostingshavebeengrowingrapidlyinabsolutenumbersandasashareofallpostingsacrosstheeconomy.Inabsoluteterms,AI-specificjobpostingsincreasedfrom3,780

uniqueinstancesinApril2010toaround82,980in2025,despitedeclinesin2023afterseveralroundsoflayoffsfromtechgiants(seeFigure2).Thatincreaseamountedto

28.5%averageannualizedgrowthforAIjobpostingsoverthe2010-to-2025period,whichoutstrippedthe11.1%averageannualizedgrowthofpostingsinthegeneral

economyandthe13.3%annualizedgrowthofthoseintheITsector.

13

Figure2.

AlongwithgreaterAI-specificjobactivityhascomegreaterdiffusionofthatactivity

acrossthecountry.Asrepresentedbyjobpostings,AIemploymentactivityhasdiffusedquiteextensivelyacrossmostofthenation,albeitofteninrelativelysparselocal

numbers.Figure3showsthatthisprocesshasnowbroughtAIactivitytomostareasofthecountryinthelastdecade.

14

Figure3.

Atthesametime,AIstartups—anothermeasureoftheindustry’sgrowth—havealsobeenproliferating.AccordingtodatafromCrunchbase,aplatformthattrackstechnology-

basedstartups,theannualnumberofAIstartupsfoundedintheU.S.increasedfrom311in2010to1,348in2024,surgingtheirshareofalltechstartupsfromaround1%to28%(seeFigure4).

15

Figure4.

Inshort,AIjobpostingsandstartupactivityeachunderscorethestronggrowthofAIactivityinboththelastdecadeandrecentyears.

YetdespitetheAIindustry’sgrowth,itsoverallsizeintheU.S.remainshardtoquantify

andstillmodest.ThisislargelyduetoAI'sbroaddiffusionintovarioussectors,which

makesitchallengingtoisolateitseconomicimpact.Nevertheless,multipleindicatorsfromouranalysispointtoarelativelymodestsizeforthenation’sAIindustry,despitethe

excitementthatenvelopedgenerativeAIinlate2022.

Inthisregard,