重新思考人工智能主權(quán):通過戰(zhàn)略投資實(shí)現(xiàn)競爭力的途徑_第1頁
重新思考人工智能主權(quán):通過戰(zhàn)略投資實(shí)現(xiàn)競爭力的途徑_第2頁
重新思考人工智能主權(quán):通過戰(zhàn)略投資實(shí)現(xiàn)競爭力的途徑_第3頁
重新思考人工智能主權(quán):通過戰(zhàn)略投資實(shí)現(xiàn)競爭力的途徑_第4頁
重新思考人工智能主權(quán):通過戰(zhàn)略投資實(shí)現(xiàn)競爭力的途徑_第5頁
已閱讀5頁,還剩62頁未讀 繼續(xù)免費(fèi)閱讀

下載本文檔

版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進(jìn)行舉報(bào)或認(rèn)領(lǐng)

文檔簡介

AIGlobalAlliance

IncollaborationwithBain&Company

RethinkingAISovereignty:PathwaystoCompetitivenessthroughStrategicInvestments

WHITEPAPERJANUARY2026

Images:AdobeStock

Contents

Foreword3

Executivesummary4

1

InvestmentpatternsacrosstheAIvaluechain5

1.1HistoricalinvestmentsintheAIvaluechain6

1.2InvestmenttrendsinkeyAIvaluechainelements9

2

AIinfrastructureasthebackbonefordrivingAIcompetitiveness11

2.1GlobalAIinfrastructureinvestmenttrends12

2.2AIinfrastructurechallengesandstrategicapproaches14

3

DifferentpathstowardsAIcompetitiveness16

Pathway1:

Fromselectiveplayerstoecosystembuilders21

Pathway2:

Fromadoptionacceleratorstoecosystembuilders22

Pathway3:

Fromemergingcollaboratorstoecosystembuilders

viaadoptionaccelerators22

Pathway4:

Fromemergingcollaboratorstoecosystembuilders

viaselectiveplayers24

Pathway5:

Fromemergingcollaboratorstoecosystembuilders24

4

Keyconsiderationsforpolicy-makers25

Conclusion27

Contributors28

Endnotes30

Disclaimer

Thisdocumentispublishedbythe

WorldEconomicForumasacontributiontoaproject,insightareaorinteraction.

Thefindings,interpretationsand

conclusionsexpressedhereinarearesultofacollaborativeprocessfacilitatedand

endorsedbytheWorldEconomicForumbutwhoseresultsdonotnecessarily

representtheviewsoftheWorldEconomicForum,northeentiretyofitsMembers,

Partnersorotherstakeholders.

?2026WorldEconomicForum.Allrightsreserved.Nopartofthispublicationmaybereproducedortransmittedinanyformorbyanymeans,includingphotocopyingandrecording,orbyanyinformation

storageandretrievalsystem.

RethinkingAISovereignty2

RethinkingAISovereignty3

January2026

RethinkingAISovereignty:

PathwaystoCompetitivenessthroughStrategicInvestments

Foreword

CathyLi

Head,CentreforAI

Excellence;Memberof

theExecutiveCommittee,WorldEconomicForum

Artificialintelligence(AI)israpidlybecomingthe

definingcapabilityofthe21stcentury,reshaping

economies,redrawingindustrialboundariesand

redefiningthenatureofcompetitivenessand

sovereigntyintheworld.Itisnolongerjusta

technologybutadriverofproductivity,innovationandgeopoliticalinfluence.Economiesthatbuild

strongfoundationsandmakestrategicinvestmentstodaywilldefinethetrajectoryoftheintelligent

economyfordecadestocome.

Whynow?

Whatbeganasaraceforinnovationhasevolved

intoaraceforAIinfrastructureeconomies

arecompetingtostrengthencontrol,secureAI

competitivenessanddeterminewhosetstherules,capturesvalueandsustainslong-termadvantage.WhiledatacentrescontinuetoattractasignificantshareofAIinvestments,manyeconomiesare

grapplingwithamorefundamentalquestion:howtoparticipatemeaningfullyinthisacceleratingrace.Atthesametime,waitingforclarityisnotanoption.FailingtoactriskswideningAIandeconomic

gapsbetweenmarkets.Atthiscriticaljuncture,economiesmustrethinktheirapproachtoAI

sovereigntyanddeterminehowtoinvestwisely.

FlorianMueller

SeniorPartnerandHead,AI,Insights&SolutionsforEurope,MiddleEastandAfrica,Bain&Company

Theapproach

Thiswhitepaper,RethinkingAISovereignty:

PathwaystoCompetitivenessthroughStrategic

Investments,emergesfromtheWorldEconomic

ForumsAIGlobalAlliancesworkonAI

competitiveness,incollaborationwithBain&

Company.ItbuildsontheForumsJanuary2025whitepaper,

BlueprintforIntelligentEconomies:AI

CompetitivenessthroughRegionalCollaboration

,andwillbefollowedbytargetedpublications

detailinghowtobuilddifferentelementsof

sovereignAIecosystemsandAIcompetitiveness.

WeprovideanewapproachtoAIsovereignty:

onethatprioritizesstrategiccontrolandresilienceoverrigidself-sufficiency,andexploreshow

economiescanstrengthentheirAIcompetitivenessthroughdeliberateinvestmentchoices,strategic

AIinfrastructuredevelopment,andAIdeployment,aswellastrustedalliances.

Thispaperisacalltoactiontoshapeafuture

whereAIsovereigntyservesasasharedengineofgrowth,ensuringstrategiccontrolwhileenablingalleconomiestobenefitfromtheadvancesofAI.Weinvitepolicy-makersandbusinessleaderstorethinkAIsovereigntyandjoinusinadvancingglobalAI

competitiveness,empoweringeveryeconomytothriveintheintelligentera.

RethinkingAISovereignty4

Executivesummary

IntheraceforAIcompetitiveness,

economiesmustpursueAIinfrastructure,interoperability,policyandpartnershipstobuildlastingcompetitiveadvantages.

AIsovereigntyisemergingasonekeyelement

oflong-termcompetitiveness.AIsovereigntyreferstotheabilityofeconomiestoshape,deployand

governAIecosystemsinaccordancewiththeirownvalues,whilstensuringstrategicandoperational

control,flexibilityand,ultimately,resiliencethroughacombinationoflocalizedinvestmentandtrusted

internationalcollaboration.Inpractice,AIsovereigntyagendasreflecteconomies’strategicpriorities

tostrengthenAIcapabilities,aimingtoreduce

dependenceonforeignentities,protectnationalinterestsandenhancecompetitiveness.SeveraleconomieshavethereforeattemptedtocompetebyowningtheentireAIvaluechain,fromraw

materialstoAI-basedapplications.

Basedoninvestmentpatterns,however,

“AIsovereignty”hasbeenconflatedwith“AI

infrastructure”anddatacentres.WhileAI

infrastructureisapressingconcernforgovernmentsplacingbigbetsonAI,itisnottheonlydeciding

factor.AIcompetitivenessdependsequallyon

whereeconomiesinvest,howtheybuildanddeployAIcapabilities,andwhotheypartnerwithacross

theAIvaluechain.

ThispaperadvocatesforaredefinedapproachtoAIsovereignty–onethatprioritizesstrategiccontrolandresilienceoverrigidself-sufficiency.SuccessintheAIerawillbedeterminednotbyisolation,but

bystrategicinterdependence–balancingdomesticinvestmentinkeyAIinfrastructurewithinternationalcollaboration.Economiesthatfocusontheir

comparativeadvantages,ensureinteroperabilityacrossAIsystemsandcultivateregionalallianceswillbebestpositionedtocapturelong-termvaluefromtheAIrevolution.

Investwisely–noteverywhere–andwithalong-termvision

Decisionsmadeinthenextfewyearswillshapewhoremainsorbecomescompetitiveinan

AI-driveneconomy.Policy-makerscanenableAIcompetitivenessbypromotingfocusand

collaboration,notfullcontrol.Thismeanspolicy-makersshouldcoordinatewithinvestorsand

otherstakeholdersto:

–IdentifystrengthsandnationaladvantagesthatcantranslateintoAIcapabilities.

–Investstrategically,concentratingonareas

ofcomparativeadvantageratherthanspreadingresourcesacrosstheentireAIvaluechain.

–EnsureinteroperableAIinfrastructuretoguaranteescalability,trustandresilience.

–Partnerstrategically,tappingintoalliancestofillcriticalgapsinsteadofduplicating

costlyefforts.

Differentpathstocompetitiveness

EveryeconomyjoinstheAIracefromadifferent

startingpoint.Whileinfrastructure-basedAI

sovereigntyisoutofreachformosteconomies,

therearedifferentpathstoAIcompetitiveness.Thispaperdescribesseveralpotentialpaths,tailoredtodifferenteconomicstartingpointsontheAIjourney.

Thebottomline

Policy-makersmustsupportthedevelopment

ofintentionalnationalAIstrategiesthatfocus

investmentsonthecomparativeadvantagesoftheireconomy.TheyshouldhelpreframeAIsovereigntyasstrategicinterdependence,wherelocalized

investmentsarecombinedwithtrustedpartnershipsandalliances.Withaclearviewoftheirlocalstrengths,economiescanleapfrogtowardsAIsuccess.

Partnership,notownership,isakeypathforward.

RethinkingAISovereignty5

1

Investmentpatterns

acrosstheAIvaluechain

DrivingAIcompetitivenessrequiresathoroughunderstandingofinvestmentflows,trendsandcapitalallocationacrosstheAIvaluechain.

Understandingthecurrentandfuturedynamics

ofAIinvestmentsisessentialtobuildingrobustAI

ecosystems(seeFigure1)andstrengtheningAI

sovereignty.Aroundtheworld,economiesaremakingmajorfinancialcommitmentstoadvancedomesticcapabilities,oftenannouncingtheseasmovements

towardsAIsovereignty.Forexample,theUKhas

pledgedtobecome“anAImaker,notjustanAI

taker”1whileChinahasprioritizedthedevelopmentandadoptionofdomesticAIchips,2andtheEuropeanUnionhascommittedtomobilize€200billionto

InvestAI,aninitiativetobuild“AIgigafactories”.3

FIGURE1

TheAIecosystem

AIecosystem

AIvaluechain

Foundationalinputs

(e.g.electricity,rawmaterialslikesilicon)

Keyenablers*

Enabling

technology

(e.g.devices,connectivity,

cybersecurity)

AIpolicies

Note:*Enablersareinterdependent;orderdoesnotrepresentanypriority.

Infrastructure(e.g.compute,

datacentres)

FundamentalR&Dandinnovation

Hardwareinputs(e.g.semi-conductors)

Applicationsandservices

Adoptionenablement

Talentandskills

Foundationmodels

AI

strategy

Capitalaccess

Data

strengthdependsnotonlyontherobustnessof

AsoutlinedinBlueprintforIntelligentEconomies:

AICompetitivenessthroughRegionalCollaboration,theAIecosystemencompassesAIvaluechain

elementsandkeyenablerswhichallowAItobe

developed,deployedandscaled.Puttingpeopleatthecentreofthisecosystemunlocksproductivity,jobcreation,innovationandgrowth.However,its

individualenablersbutalsoonhoweffectivelytheyreinforceoneanother.Economiesshouldthereforebemindfuloftheneedforclosecoordination

amongenablersandremainadaptive,asnewoneswillinevitablyemergeandreshapetheAIecosystemgiventherapidtechnologicalprogress.

AverageannualgrowthinAIinvestments(2010–2024):

approximately33%

1.1HistoricalinvestmentsintheAIvaluechain

capacityequippedtohostadvancedAIworkloads–hasbeenamajorfocus,attractingmorethan

$600billionbetween2010and2024.

Todate,morethanhalfofglobalAI-relatedinvestmenthasbeendirectedtowardsAIinfrastructure,aswellasapplicationsandservices(Figure2).Investment

inAI-dedicatedinfrastructure–suchasdatacentre

FIGURE2

AIinfrastructure,aswellasapplicationsandservices,hasattractedmostglobalinvestmentsintheAIvaluechain

HistoricalinvestmentsacrosstheAIvaluechain,*2010–2024,in$billions

900–1,300

Technology

600–900

Remainingindustries

200–300

100–200

Applicationsandservices

200–300

150–250

Foundationmodels

Foundationalinputs

Hardwareinputs

Infrastructure

Data

Note:*InvestmentestimatesarebasedoncapitalandR&Dspendingacrosselectricitycapacity,siliconprocessing,equipmentandchipmanufacturing,datacentres,foundationmodeltraining,andAIapplicationdevelopmentbymajortechnologyfirms,aswellascorporateinvestmentsinAIinitiativesforotherindustriesandthemarketsizeofdata-relatedsolutions.Land-relatedinvestmentsareexcluded.

Source:PublicsourcesfromWEF;IMF;IEA;IRENA;USGS;NRMMRRD;GoldmanSachs;WorldBank;WHO;IATA;Gartner;

S&PGlobal;OpenAI;EpochAI;Cushman&Wakefield;Bain&Company;MarketResearch.

TheUSandChinadominatetheinvestment

landscape,capturingabout65%ofaggregate

globalinvestmentintheAIvaluechain(Figure3).TheiroutsizedpresenceineveryelementoftheAIvaluechainreflectsafull-stackapproachthatfeweconomiescanmatch,giventhescaleof

investmentneeded.

AIinfrastructureandhardwaredemandisexpectedtoriseevenfurtheraseconomiespursuedomesticcontroloverdatacentresandcriticalinputssuch

asprocessors.Theworldcouldhavenearly2,000hyperscaledatacentresby2030–asharpincreasefromthe300thatexistedin20164and1,136thatexisttoday.5

RethinkingAISovereignty6

RethinkingAISovereignty7

3

FIGURE

TheUSandChinaarethelargestinvestorsacrosstheAIvaluechain

HistoricalinvestmentsacrosstheAIvaluechainsplitbyeconomy,2010–2024,in$billions

100–200200–300600–900150–250200–300900–1,300

100%

80%

60%

40%

~90%

driven

by

TSMC*

20%

and

UMC**

0

Infrastructure

Applicationsandservices

Data

Foundationmodels

Foundational

inputs

Hardwareinputs

口TopinvestingeconomiesperAIvaluechainelement

RestofworldUSChina●Europe(excludingUK).IndiaSingapore.Remainingeconomies***

Notes:*TaiwanSemiconductorManufacturingCompany;**UnitedMicroelectronicsCorporation;***IncludingBrazil,Canada,Japan,Korea,theUnitedArabEmirates,theUK,andallothereconomiesexceptwhere“restofworld”isshownseparately.

Source:PublicsourcesfromWEF;IMF;IEA;IRENA;USGS;NRMMRRD;GoldmanSachs;WorldBank;WHO;IATA;Gartner;S&PGlobal;OpenAI;EpochAI;Cushman&Wakefield;Bain&Company;MarketResearch.

Economiescanthereforeaccruemeasured

advantagesbypositioningthemselvesinsectorsalongtheAIvaluechainwheretheycanscale

demand.Toassessthesuccessofinvestments,leadersshouldconsideradoptionandoutcomes,includingaspectsofresilience,notjustmoney

deployedorreturned.

YetthisisnottheonlypathtoAIcompetitiveness.Balancedandtargetedplayshavehelpedsome

economiesturncapitalintodeepcomparative

strengthsandresilience.Forexample,Singaporehastakenabalancedapproach,intentionallyallocatingresourcesinameasuredwayacrosstheAIvalue

chain(casestudy1).Similarly,SouthKoreahas

initiallyconcentratedtheirinvestmentsinhardwareelementslikechips(Figure4)andisexpanding

effortsacrossfoundationmodels6andapplications.7

CumulativeAIinfrastructureinvestmentsince2010:

morethan$600billion

RethinkingAISovereignty8

ShareofaggregateglobalAIinvestmentbytheUSandChina:

approximately65%

4

FIGURE

Economieshavepursueddifferentinvestmentstrategies

BenchmarkofhistoricalinvestmentsacrosstheAIvaluechain:

Accumulatedinvestments,2010–2024,in%of2024grossdomesticproduct(GDP)

Foundationmodels

Applicationsandservices

Hardwareinputs

Infrastructure

Data

Foundationalinputs

US

3.4一5.1%

3.1一4.6%2.2一3.3%1.7一2.6%1.5一2.3%1.5一2.2%1.5一2.2%1.4一2.0%1.2一1.9%1.2一1.8%1.1一1.6%0.7一1.1%

Singapore

Korea

China

UnitedArabEmirates

Japan

Canada

UK

Approximately90%of

investmentsinhardwareinputsdrivenbyTSMC*andUMC**

Restofworld

India

Europe(excludingUK)

Brazil

0.00.51.01.52.02.53.03.54.04.5

口OutperformingeconomiesinselectAIvaluechainelements

Note:*TaiwanSemiconductorManufacturingCompany;**UnitedMicroelectronicsCorporation.

Source:PublicsourcesfromWEF;IMF;IEA;IRENA;USGS;NRMMRRD;GoldmanSachs;WorldBank;WHO;IATA;Gartner;S&PGlobal;OpenAI;EpochAI;Cushman&Wakefield;Bain&Company;MarketResearch.

RethinkingAISovereignty9

1.2

AIinfrastructurehasexpanded

rapidlyas

organizationshavebuiltinitialcapacityandscaledto

supportfoundationmodelproviders

andgenerativeAIworkloads.

ProjectedannualinvestmentinAIapplicationsby2030:

approximately$1.5trillion

InvestmenttrendsinkeyAIvaluechainelements

Foundationalinputs

CrucialfoundationalinputsforAIincludeenergy,

rawmaterials(e.g.siliconandrareearthelements)andland.Since2010,energyandrawmaterials

haveattractedover$100billionininvestment–

withthevastmajoritydirectedtowardsenergy

systemspoweringdatacentres,whoseelectricity

consumptionhasreached1–2%ofglobaldemand.8AlthoughAIcanimproveenergyefficiency,ithas

thusfarcontributedtorisingelectricitydemand(seetheWorldEconomicForum’s2025whitepaper,

ArtificialIntelligence’sEnergyParadox:Balancing

ChallengesandOpportunities

).Thisincreasealsoledtodebatesonrevivingnuclearpowerand

net-zerovs.fullAIpotentialgoals.Aforthcoming

reportonoptimizingAIandhyperscaledatacentresforenergyefficiencywillbelaunchedaheadoftheForum’s2026conveninginSaudiArabia,informingfutureregionalenergypolicy.

InvestmentsinfoundationalinputswillkeepgrowingbutremainlimitedrelativetotherestoftheAIvaluechain,reachingapproximately$50billionperyear

by2030.Policy-makersshouldhelpintegratethesefoundationalinputswithbroaderAIinfrastructure

investmentstosupportsustainableAIsystems’

growth(seetheWorldEconomicForum’srecent

publicationFromParadoxtoProgress:ANet

PositiveAIEnergyFramework).Chapter2examinestheroletheseinputsplayinenablingAIhardwareandinfrastructure.

Hardwareinputs

Since2010,investmentsinhardwarehave

surpassed$200billion,ledbycapitalexpenditure

fromsemiconductorfoundries,alongside

contributionsfromlithographyequipmentmakers

andfablesschipsuppliers.Withapproximately

90%ofglobalfoundryrevenueconcentrated

infourcompanies–TaiwanSemiconductor

ManufacturingCompany(TSMC),United

MicroelectronicsCorporation(UMC),Samsung

FoundryandSemiconductorManufacturing

InternationalCorporation(SMIC)–whicharebasedacrossthreeeconomies,9akeypartofthedebate

onAIsovereigntyisonthehighgeographical

concentrationofhardwareproduction.Hardware

investmentisprojectedtogrowby15–25%annually,reachingalmost$90billionperyearby2030.

Infrastructure

AIinfrastructurehasexpandedrapidlyas

organizationshavebuiltinitialcapacityandscaledtosupportfoundationmodelprovidersand

generativeAIworkloads.10Investmentisforecastedtoincreaseat10–15%annually,reachingover$400billionperyearby2030,withtotalvendorfinancingacrosschipmakers,modeldevelopersanddata

serviceprovidersexpectedtobeevenhigher.

Whilebusinessmodelsareveryintegrated,with

hyperscalerscoveringasignificantpartoftheAI

valuechain,theAIinfrastructuredevelopmentis

largelyfragmentedacrosstheglobe.TheUSholdsover40%ofinstalledglobaldatacentrecapacity,11whilemanyemergingeconomiesarestillestablishingreliableconnectivityanddigitalinfrastructure.

Data

Investors’focusisshiftingtowardsdata,mirroringamarkettrendofmigratingdatatohyperscalers.

Cumulativeinvestmentindata-relatedsolutionsisestimatedatover$150billion,reflectingstronggrowthafter2018.12,13By2030,investmentis

expectedtoexceed$90billionperyear.This

includestrainingdatasetsforfoundationmodelsanddatasolutionsthatsupportapplications

andservices(e.g.integration,governance,migration,marketplaces).

AIperformanceandcompetitivenesshingeonthesize,diversity,uniqueness,recencyandoverall

qualityofcurateddata,aswellasitsintegrityandavailability.Meta’smulti-billion-dollarinvestmentinScaleAIunderscoresthestrategicvalueofdata.

Thistrendextendsbeyondadvancedeconomies,offeringemergingeconomiesopportunitiestobuildcompetitiveadvantages.

Foundationmodels

Foundationmodelinvestmentisprojectedtogrow25–35%annually,reachingatleast$300billionperyearby2030.Thisgrowthisdrivenbytherising

complexityoflargelanguagemodels(LLMs)and

thecontinuinggrowthofclassicalmachinelearning(ML)andsmalllanguagemodels(SLMs).

RethinkingAISovereignty10

US-basedproviderssuchasAnthropicandOpenAIhavecollectivelyraisedmorethan$85billion,14,15

mostofitinthepasttwoyears.OutsidetheUS,firmssuchasDeepSeek,MistralAIandothersacrossAsia,EuropeandtheMiddleEastare

alsodevelopingopen-sourceandproprietary

modelstailoredtolocallanguageandculture–withsubstantialreturnsexpectedastheyscale.

Theirinvestmentswillalsogeneratedemandfor

applicationsandshapetheAIecosystemevolution.

Applicationsandservices

Inpersonalcomputing,valuehasshiftedfrom

hardwaretosoftware.Inthemobileera,value

movedfromdevicestoapps.Ashistoryrepeats,investmentsintheAIvaluechainareexpectedtomovetowardsAI-basedapplicationsandservices.

AnnualinvestmentinAIapplicationscouldreach$1.5trillionperyearby2030,outpacinggrowth

inAIinfrastructureandfoundationmodelsand

deliveringgreatereconomicvaluethroughdomain-specificusecases.Forexample,adoptionin

healthcareapplicationscouldreducespending

by5–10%16withoutsacrificingquality.Similarly,

AIcouldfreeupapproximately8%ofpublicsectorbudgetsby2030.17Economiesthateffectively

channelinvestmentintohigh-impactapplicationsandbuildsupportiveAIecosystemswillsecurethelargestgains.

Thus,globalinvestmenthassurgedacrosstheAIvaluechain,butAIcompetitivenessdependson

howstrategicallyeconomiesallocatecapital,notjusthowmuchtheyspend.Chapter2exploresthepivotalroleofAIinfrastructureasthebackboneoftheAIvaluechaininmoredetail.

2

AIinfrastructureasthebackbonefordrivingAIcompetitiveness

Policy-makersshouldfacilitateinvestmentinresilientAIinfrastructure,considering

trends,barriersandstrategicimplications.

investmentfocus,deliveringscalablecomputefortrainingclustersandinferenceoperations.

Today,AIisdrivingoneofthelargestinfrastructurebuildoutsinmodernhistory–amulti-trillion-dollar

expansionspanningchips,datacentresandenergysystems.Tonavigatethis,economiesmustpair

globalAIinfrastructureinvestmenttrendswiththeirlocalconstraints,barriersandenablers.Thisdefineseacheconomy’s“AIinfrastructureoptionspace”andsupportsresilientinvestmentchoices(seeFigure5).

AIinfrastructureunderpinsfoundationmodelsand

applicationsandtherebydrivesAIcompetitiveness,attractsforeigninvestmentsandenablesthebroadergrowthofintelligentsystems.Itspansinterdependentlayersincludingdatacentres,computeresources,

cloudplatforms,edgesystemsandhigh-capacity

connectivitynetworks.AIecosystemreadiness

dependsonhoweffectivelytheselayersinteroperatetosupportbothtrainingandinference,enabling

large-scaledataprocessinganddeployment.Withinthisecosystem,datacentresremaintheprimary

FIGURE5

AspectsinfluencingAIinfrastructureoptions(notexhaustive)

Understandinganeconomy’sAIinfrastructureoptionstodriveAIcompetitiveness

Globalinvestmenttrends,suchas:

nDSigni?cantincreaseinAIinfrastructureinvestments

New?nancingmodelsforAIinfrastructure(e.g.partnership-drivenmodels)

nDRiseofneocloudproviderinvestments

withtailoredAItraininganddeploymentofferings

De?ningtheAIinfrastructureoptionspaceandderivingstrategicimplications

Land,energyandwaterascriticalconstraints

forscalingAIinfrastructurewithintheeconomy

AIinfrastructurerequiringahighlyskilledworkforce

LocalregulationslowingdownAIinfrastructuredevelopment

Localbarrierstoaddress,suchas:

RethinkingAISovereignty11

RethinkingAISovereignty12

2.1GlobalAIinfrastructureinvestmenttrends

investedbillionsintoOpenAIandprovidedcomputecapacitytoembedfrontierAIresearchinitscloudplatforms.Similarly,Huaweipartneredwiththe

AIinfrastructurestart-upSiliconFlowtomakemodelsfromDeepSeekavailableviathecloud.

Theseinvestmentsledtoanewmixofstrategic

acquisitions,partnershipsandacceleratingfinancingroundsunderpinnedbyacomplexflowoffunds.

OneexampleistheStargateProject,ledbyOpenAIandSoftbankinpartnershipwithNVIDIA,Oracleandotherindustrycollaborators(Figure6).

AsdiscussedinChapter1,AIinfrastructure

investmentcontinuestorise,alongsidenew

financingandownershipmodels,including

expandedpublic-privatepartnerships(PPPs).Earlydealsfocusedonacquiringhardware,likechips,

computinganddatacentreequipment–suchas

AdvancedMicroDevices’(AMD)acquisitionofXilinx,andNVIDIA’sMellanoxdeal.Since2019,investmentprioritieshaveshiftedfromownershiptopartnership-drivenmodels.Hyperscaler-modellabpartnershipsnowdominateinvestmentactivity.Microsofthas

FIGURE6FundandresourceflowsbetweenOpenAI,NVIDIAandOracle

IllustrationoftheecosystemdrivingAIinfrastructure

Foundationmodels

Applicationsandservices

HardwareinputsInfrastructureData

Foundationalinputs

Equity

USDBandGPUleasing

NVIDIA

USDBanddata

OpenAI

AIservices

USDB

Compute

Onsitenaturalgasplantplanned

Endusersandcompanies

Stargate

TSMC

USDB

USDB

USDBandracks

ASML

USDB

USDB

GPUs

Oracle

Compute

MGX

Enablingtechnology (especiallyconnectivity)

Softbank

Talentandskills

Adoptionenablement

FundamentalR&Dandinnovation

AIstrategy

Capitalaccess

Investment?owXXXAssetsandservicesacquired

Source:AdaptedfromCitiResearch2025.

flexibleaccesstoadvancedNVIDIAhardware.Thecompanyisundergoingamassivecapacityexpansion,havingsecuredatotalof$25billionsince2024.AIinfrastructurebuildoutisalso

expandingthroughnationalcloudproviders

suchasHUMAIN(SaudiArabia),G42(United

ArabEmirates),Outscale(France)andSTACKIT(Germany),reflectingaglobaltrendtowards

digitalsovereigntyandreduceddependencyonhyperscalers.

Surgingdemandforcomputeisspawningnew

AIinfrastructuredevelopmentmodels,suchas

neocloudproviders,nationalcloudprovidersand

industry-specificAIclouds.Whilehypersc

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會(huì)有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
  • 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫網(wǎng)僅提供信息存儲(chǔ)空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負(fù)責(zé)。
  • 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

最新文檔

評論

0/150

提交評論