Introduction

Twiceayear,BlackRock’sseniorportfoliomanagersandinvestmentexecutivesgatherfortwodaystodebatetheoutlookfortheglobaleconomyandmarkets–anditsimplicationsforportfolios.TheGlobalOutlookistheculminationofthatdebateanddiscussionatour

latestForumandinformsourmacroframing.Whathasbecomeclearthisyear:ourmegaforcesframework,launchedafewyearsago,isacriticalwayofunderstandingthesedriversofeconomictransformationandtheinvestmentopportunitiestheycreate.ThisOutlookiscenteredinreconcilingtheclashingordersofmagnitudeoninvestmentandpotentialrevenuesintheAIbuildout–andhowitis

intersectingwithothermegaforces,suchastheenergytransitionandfutureoffinance.

Authors

JeanBoivin

Head—BlackRock

InvestmentInstitute

VivekPaul

GlobalHeadof

PortfolioResearch–

BlackRockInvestmentInstitute

EdFishwick

HeadofRiskand

QuantitativeAnalysis–BlackRock

GlennPurves

GlobalHeadofMacro—BlackRockInvestmentInstitute

WeiLi

GlobalChief

InvestmentStrategist—BlackRockInvestmentInstitute

RickRieder

Headof

FundamentalFixedIncome–

BlackRock

RaffaeleSavi

GlobalHeadofSystematic–

BlackRock

Executivesponsors

RobKapito

President–BlackRock

RichKushel

Head–BlackRock

Portfolio

ManagementGroup

RajRao

FoundingPartner,

PresidentandChiefOperatingOfficer,

GlobalInfrastructure

ScotFrench

Founding

PartnerandCo-

President,HPS,a

partofBlackRock2

Partners,apartof

Contents

Theglobaleconomyandfinancialmarketsarebeingtransformedbymega

forces,especiallyAI.Technologyisbecomingcapital-intensive,andtheAI

buildoutcouldbeunprecedentedinbothspeedandscale.Withafewmega

forcesdrivingmarkets,itishardtoavoidmakingabigcallontheirdirection-

Introduction

2

andassuch,thereisnoneutralstance,notevenexposuretobroadindexes.We

remainpro-riskandseetheAIthemestillthemaindriverofU.S.equities.Yetthis

Summary

3

environmentisripeforactiveinvesting-pickingwinnersandlosersfromamongthebuildersnowandlaterasAIgainsstarttospread,inourview.

Investmentenvironment

ClashingordersofmagnitudeGrowthbreakout:now

conceivable

4-5

4

5

We’velongarguedwe’reinaworldofstructural

transformationshapedbyafewmegaforces,

includinggeopoliticalfragmentation,thefuture

offinanceandtheenergytransition.ButthemostobviousnowisAI,withabuildoutofapotentially

ThecapitalspendingambitionstiedtotheAI

buildoutaresolargethatthemicroismacro–ourfirsttheme.Theoverallrevenuescouldjustifythespend–yetit’sunclearhowmuchwillaccrueto

thetechcompaniesbuildingAI.That’swhythey

Themes

Microismacro

Leveragingup

Diversificationmirage

6-8

6

7

8

unprecedentedspeedandscale.Thisshiftto

capital-intensivegrowthfromcapital-lightis

profoundlychangingtheinvestmentenvironmentandpushinglimitsonmultiplefronts–physical,financialandsocio-political.

arenotonautopilotandcouldadjustplanswith

greatervisibilityonrevenuesandasstarkenergyconstraintsbite.Thelattercouldslowthebuildoutbutcreateotheropportunities.Westaypro-risk

andoverweightU.S.stocksontheAItheme.This

Megaforces

AIbuildoutfacesconstraints

Fragmentation:AIanddefenseFutureoffinanceevolving

quickly

PrivatecreditentersnewphaseInfrastructure:atimelymomentRoomtorunforEM

Bigcalls

Granularviews

9-14

9

Afewbigmacrodriversimpliesafewbigmarketdrivers.Today’smarketconcentrationreflects

isagreattimeforactiveinvestingforthosewithinsightsonwhowillcapturetherevenues.

10

11

12

13

14

15

16

underlyingeconomicconcentration.Thatmeansinvestorscan’tavoidmakingbigcalls,inourview.

AIisthedominantmegaforcerightnow,helpingpropelU.S.stockstoall-timehighsthisyear.In

recentmonths,investorshavestartedtofret

aboutequityvaluationsandwhetheranAIbubbleisforming.TheShillerprice-to-earningsratio

showsU.S.stockvaluationsarethemost

expensivesincethedot-comand1929bubbles.Marketbubbleshaveariseninallmajorhistoricaltransformations–andthatcouldhappenagain.Butthosebubblesalsogrewforsometimeand

TheAIbuildersareleveragingup,oursecond

theme.Wethinkthatisneeded:Thespendingis

front-loaded–andunavoidabletorealizethe

benefits–whilerevenuesareback-loaded.Alongwithhighlyindebtedgovernments,thiscreatesamoreleveredfinancialsystemvulnerableto

shocks–includingbondyieldspikestiedtopolicytensionsbetweeninflationanddebt

sustainability.Weseeprivatecreditand

infrastructuresupportingthisfinancing.Wegounderweightlong-termU.S.Treasuriestactically.

Thisenvironmentalsocreatesadiversification

onlybecameobviousaftertheyburst.That’swhyweseektoreconcilethepotentialordersof

magnitudeofinvestmentandrevenues.That’s

ourbasisfortrackingtheAItransformationandiswhatwedointhisOutlook.

mirage,ourthirdtheme.Allocationsmadeunder

theguiseofdiversificationmaynowinfactbebig

activebets.Wethinkportfoliosinsteadrequirea

preferidiosyncraticexposuresinprivatemarket3s.

clearplanBandareadinesstopivotquickly.We

Clashingordersofmagnitude

Arapidtransformation

Lengthandcapitaldeepeningofnotableinnovations,1760-2040

Capitaldeepeningcontributionto

GDPgrowth(percentagepoints)

25

20

15

10

5

0

InformationandCommunication

Technology(1974-2012)

Steam(1760-1910)

AI(2022-2040)

Electricity(1899-1941)

TheAIbuildoutcouldbefasterandgreaterthanallpasttechnologicalrevolutions.Seethechart.ExternalestimatesofAIcorporatecapital

spendingambitionsrangebetweenU.S.$5-8trilliongloballythrough

2030,mostofthatintheU.S.Thechallengeforinvestors:reconcilingthehugecapitalspendingplans

withthepotentialAIrevenues.Willtheirordersofmagnitudematch?

Basedonthehighestendofthe

spendingrange,theanswer

dependsonwhetherU.S.growth

canbreakoutfromitslong-term

2%trend.Thisisatallorder,butAImakesitconceivableforthefirst

time.Seenextpage.Evenif

spendingandrevenuesreconcileatthemacrolevel,thereisstilla

timingmismatch.TheAIbuildoutrequiresfront-loadedinvestment:compute,datacentersandenergyinfrastructure.Theeventual

revenuefromthatinvestment

comeslater.Thebuildoutisjuststartingandconcentratedinthetechsector,fornow,butweseeproductivityandrevenuegainsspreadingacrosstheeconomy.

Thegapintimebetweencapex

needsandeventualrevenuesmeansAIbuildershavestartedusingdebttogetoverafinancing“hump.”Thisfrontloadingofspendingis

necessarytorealizeeventualgains.Privatesectorleveragewilladdtoaheavilyindebtedpublicsector,

creatingfinancialsystem

vulnerabilities.Bondyieldspikescouldposearisktothisfinancing.

Allthiscreatesaverydifferent

investmentenvironmentwithsomecorefeatures.First,higherleveragewillresultingreatercreditissuanceacrosspublicandprivatemarkets.

Second,abroadlyhighercostof

capitalasthisbigborrowingputs

upwardpressureoninterestrates.

Third,beforenewpoolsofAI

revenuesspreadacrossthe

economy,furtherconcentrated

marketgainswithintechwillrequirebigcalls.Andfourth,moreroomforalphaandactiveinvestingasthoserevenuesspreadbeyondtech.All

thismeanstraditionalapproachestoportfolioconstructionneeda

rethink.

020406080100120140160

Lengthoftransformation(years)

Charttakeaway:TheAItransformationisunfoldingfastandisdrivingrecordinvestment—wecouldreachalevelclosetothelargestbuildoutinhalfthetime.

Forward-lookingestimatesmaynotcometopass.Source:BlackRockInvestmentInstitutewithdatafrom

Crafts(2021),December2025.Notes:ThechartshowstheaverageannualcontributionofcapitalspendingtoGDPgrowthforpreviousU.S.technologies(except“steam,”fortheUK)againstthelengthoftimethe

capitalwasspent.Estimatesforsteam,electricityandICTaretakenfromCrafts(2021).Thespendneededforartificialintelligence(AI)iscalculatedusingrealizedcapitalspendingbetween2022-2024,theupperendoftheU.S.$5-8trillionrangeoftotalcapitalspendingintentionsspanningtheperiod2025-2030,andan

assumptionthatcapexintensitycontinuesatasimilarratethrough2030-2040.

Thisinvestmentenvironmenthassomecorefeatures:higherleverage,ahighercostofcapital,concentratedgains

requiringtheneedtomakebigcalls,andmoreopportunitiesforactiveviews.

4

Growthbreakout:nowconceivable

Neverbrokenout

U.S.GDPpercapitaandlong-termtrend,1870-2024

Logscale

10.5

10.0

9.5

9.0

8.5

8.0

7.5

Historic

1.9%growthtrend

WeseeAIcapitalspendingstill

supportinggrowthin2026,with

thecontributiontoU.S.growth

frominvestmenttotalingthree

timesitshistoricalaveragethis

year.Thiscapital-intensiveboostislikelytopersistintonextyear,

allowinggrowthtoholdupevenasthelabormarketkeepscooling.

This“nohiring,nofiring”stasis

givestheFederalReserveroomtokeeptrimmingpolicyratesin2026.Yetinflationcouldprovesticky

abovetheFed’s2%target.WeseeafavorablepolicyandregulatorybackdropheadingintotheU.S.

midtermelections.Anyreboundinbusinessconfidenceandhiring

couldreigniteinflation,bringingbackpolicytensionswithdebt

sustainability.

Zoomingout,AIwilllikelykeep

trumpingtariffsandtraditional

macrodriversasithasin2025.InaworldwhereU.S.trendgrowth

staysunder2%,absorbingthat

capexwillrequirebigmacro

adjustments–crowdingoutothernon-AIspendingintheeconomy,andpotentiallyresultinginmuchhigherinflation.

Butthebigquestioniswhetherit’s

possibleforU.S.growthtoeventuallybreakoutofitslong-termtrend.TheU.S.sitsattheglobaleconomic

frontier.Butallmajorinnovationsofthelast150years–includingsteam,electricityandthedigitalrevolution

–werenotenoughforittobreakoutofits2%growthtrend.Doingsoisatallorder.

Yetitisconceivableforthefirsttime.Why?AIisnotonlyaninnovation

itselfbuthasthepotentialto

innovatetheprocessofinnovation.AIcouldbegintogenerate,testandimprovenewconceptsonitsown.Ifthathappens,therateofdiscoverycouldaccelerate,drivingscientificbreakthroughssuchasinmaterials,drugsandtechnology.Thisself-

reinforcingloopofacceleratinginnovationiskeytoachievingthebreakout.

Whileconceivable,wedon’tknowifitwillhappen.Wetrackevidencethatcanserveasaleadingindicatorof

fasterinnovation–forexample,a

surgeinthenumberofnewpatentseachyear–beforeanybenefitsstarttoshowupinbroaderactivitydata.

18701890191019301950197019902010

Charttakeaway:Overthepast150years,U.S.growthhasstayedremarkablyclosetoa2%trendevenwithpasttechnological

revolutions.Wethinkit’sconceivablethatacceleratingAI-driveninnovationcouldboostgrowthabovethattrend.

Source:BlackRockInvestmentInstitute,BureauofEconomicAnalysisandMacrohistoryDatabase,withdatafromHaverAnalytics,December2025.Note:HistoricaldatacompiledbyòscarJordà,MoritzSchularick,andAlanM.Taylor.2017.MacrofinancialHistoryandtheNewBusinessCycleFacts.inNBERMacroeconomicsAnnual2016,volume31,editedbyMartinEichenbaumandJonathanA.Parker.Chicago:Universityof

ChicagoPress.

FortheU.S.tobreakoutofa2%growthtrend,itwillrequireinnovationinhow

weinnovate-andthatfirstrequiresmassiveupfrontinvestment.

5

Microismacro

Doingthemath

EstimatesofannualU.S.corporaterevenuegrowththrough2030

TheAIbuildoutisdominatedbya

handfulofcompanieswhose

spendingissolargethatithasa

macroimpact.Takingaviewon

thesecompaniesrequiresassessingwhetherthemacromathaddsup.

TaketheupperendoftherangeofAIinvestmentambitionsthrough2030of$8trillion.Seepage4.Broker

analystsseetotalannual

hyperscalerrevenueincreasingby

$1.6trillion.Thatwouldfallshortoftherevenuesneededtodelivera

reasonablerateofreturnof9-12%onthatupper-endofAIinvestment.Seethechart.Butthat’snotthefullpicture:thetruepotentialofAIgoesbeyondexistingbusinesslines.AsAIbecomesembeddedintheeconomy,weexpectittocreateentirelynew

poolsofrevenueinthetechsectorandbeyond.Whereexactlyishighlyuncertain.Butwecanstilltryto

gaugethemacrosize.

Todothat,supposewebreakoutofa2%growthworldandAIdeliversa1.5%boosttogrowththrough

productivitygains.Weestimatethatthiswouldexpandeconomy-wide

revenuesby$1.1trillionperyear–morethanenoughtojustifythe

upper-endofspending.

Butevenifthenumbersmightadd

upatthemacrolevel,thereisno

guaranteethoserevenueswillflow

totheAIbuilders.Ifthetechsector

growsitsshareofrevenuefrom25%now,LSEGdatashow,to35%,that’sa$400billionboost.Seetheright

bar.Combiningthesenewrevenueswithcurrentanalystestimates

suggeststhiscoulddelivera

reasonablereturnoninvestment.

Weseeroomfortechrevenuesto

moveinthatdirectionascurrentAIbuilderseatintopartsofthetech

ecosystem,likesoftware.AI’sabilitytoautomatetasks,suchascoding,meansitcouldreplaceothertech

services.Yetrevenuegainswill

increasinglyspreadacrosssectors

andtheeconomy.EntirelynewAI-

createdrevenuestreamsarelikelytodevelop.Howthoserevenuesare

sharedislikelytoevolve–andwe

don’tyetknowhow.Findingwinnerswillbeanactiveinvestmentstory.

WestayoverweightU.S.stocksandtheAItheme,supportedbyrobustearningsexpectations.Thecapexmaypayoffoverallevenifnotfor

individualcompanies.Thenext

phasemaybemoreaboutenergyandresolvingbottlenecks.

Revenuegrowththrough2030

$2.5T

$2.0T

$1.5T

$1.0T

$0.5T

$0.0T

New

$1.7-2.5T

AI-driven

revenue

IncrementalrevenuesneededBottom-uprevenuegrowth

forareasonablereturnforecast

Cloud

Ads

Retail

Platforms

Services

Software

Hardware

Charttakeaway:WefindthatlargetechfirmswillneedtogrowrevenuefasterforAI-driveninvestmenttopayoff.

Forward-lookingestimatesmaynotcometopass.Source:BlackRockInvestmentInstitute,Bloombergand

others,December2025.Notes:Theleftbarshowsthenecessaryincrementalrevenuestobeachieved2025-2030toreacha9-12%lifetimeinternalrateofreturn.Thisisbasedondatacenterdeploymentestimates

(gigawattterms)andcapexpergigawattassumptionsreflectingcostinflation,GPUlife,phasingand

hyperscalermarketshare.Wecross-referencedwithothersourcesincludingMcKinsey&Company,BostonConsultingGroup,LawrenceBerkeleyNationalLaboratory,InternationalEnergyAgency,U.S.DepartmentofEnergy,BloombergNEF,SchneiderElectric,ElectricPowerResearchInstituteandothers.Thesecalculationsincluderetailrevenuegrowth(Amazon)andothernon-datacenterrevenuesaccordingtoBloomberg

consensusestimates.TherightbarshowstheBloombergconsensusfortheincreaseinannualrevenuesover2025-2030forU.S.hyperscalersincludingMicrosoft,MetaandGoogle.

Investmentimplications

?Westayrisk-onandoverweightU.S.stocksontheAItheme.

?WebelievetheAIthemewillbecomeanactiveinvestmentstoryofidentifyingnewwinnersasAIrevenuesspreadacrosstheeconomy.6

Leveragingup

Debtdivergence

U.S.governmentandcorporatedebt,1950-2025

ShareofGDP

120%

100%

80%

60%

40%

20%

0%

U.S.governmentdebt

Corporatedebtex.financials(right)

24x

20x

16x

Debt-to

GettingovertheAIbuiIdout’s

financinghumpbetweenfront-

loadedinvestmentandback-loadedrevenuesneedslong-termfinancing.Thatmeansgreaterleverageis

inevitable.Thishasalreadystartedasseenfromrecentbondsalesfromlargetechfirms.

Thegoodnews:thestartingpointforprivatesectorleverageishealthy,

particularlylistedtech.Thelargestcloudspenderssportanaverage

debt-to-equityratioof0.54timesasofNov.26,Bloombergdatashow,

pointingtoroomtoleverup.New

issuancefrommajortechfirmshassofarbeenwellreceivedbymarkets.

ThefinancingneedstiedtoAIcapexfarexceedwhateventhelargest

firmscanmeetinternally.Butthisisoccurringagainstabackdropof

alreadyhighlyleveragedpublic

sectorbalancesheets.Seethechart.Elevateddebtandcompetingfiscalpriorities-fromGermany’sdefensepush,stretchedU.S.andUKbudgetdeficitsandJapan’snewspending

pledges-IimitthepubIicsector’sroleinfinancingtheAIbuildout.

That’swhyweexpectcompaniestokeeptappingpublicandprivate

creditmarkets,creating

opportunitiesforinvestors.

Higherborrowingacrosspublicandprivatesectorsislikelytokeep

upwardpressureoninterestrates.

Elevateddebtservicingcostsare

onereasonweseetermpremium-orcompensationinvestorsdemandtoholdlong-termbonds-risingandpushingupyields.Weturn

underweightlong-termTreasuries.AIcouldeventuallydeliverthe

productivitygainsneededtoeasepublicdebtburdens,yetthatwilltaketime.

Onerisk:astructurallyhighercostofcapitalraisesthecostofAI-related

investmentandaffectsthebroadereconomy.Amoreleveragedsystemalsocreatesvulnerabilitiestoshockssuchasbondyieldspikestiedto

fiscalconcernsorpolicytensions

betweenmanaginginflationand

debtservicingcosts.Indebted

governmentshavelesscapacitytocushionsuchshocks.ThisiswhereAIfinancingneedsandgovernmentdebtconstraintsintersect.

-profitratio

12x

8x

4x

0x

19501960197019801990200020102020

Charttakeaway:Publicandprivatebalancesheetshave

divergedsharply:governmentdebthassurgedtopostwarhighswhilecorporateleveragehaseased.Thisiswhythecorporate

sectorhasroomtoleverageuptofundtheAIbuildout.

Source:BlackRockInvestmentInstitute,CongressionalBudgetOffice,BureauofEconomicAnalysis,FederalReserveBoard,withdatafromHaverAnalyticsandLSEGDatastream,December2025.Notes:Thechart

showsU.S.federaldebtasashareofnominalGDP,anddebtofU.S.nonfinancialcorporationsasashareofpre-taxprofits.

Investmentimplications

?WeseerisingAIexposureinpublicandprivatecreditmarkets.

?Wegotacticallyunderweightlong-term

Treasuriesasweseeinvestorsdemandingmoretermpremium.

7

Diversificationmirage

Apowerfulcommondriver

VarianceinS&P500returnsexplainedbyadominantunderlyingfactor

Varianceexplained

13%

12%

11%

10%

9%

8%

7%

6%

5%

Portfoliodecisionstakenundertheguiseof“diversifying”awayfromthehandfulofforcesdrivingmarketsisnowabiggeractivecall.Ouranalysisshowsthatafteraccountingfor

commondriversofequityreturns,

likevalueandmomentum,anever

largershareofU.S.equityreturns

reflectsasingle,commondriver.Seethechart.Marketsaremore

concentrated,withlessbreadth.

AttemptstodiversifyawayfromtheU.S.orAI-towardotherregionsor

equal-weightedindexes,forexample

-amounttolargeractivecallsthanbefore.Theyhaveleftportfolioslessexposedtothisyear’sdominant

driverofreturns-adriverwehave

convictionin.Theequal-weighted

S&P500isupjust3%thisyear

versus11%forthemarketcap-

weightedS&P,LSEGdatashow.Wedon’tthinkthiscanbeviewedas

indiscriminatelyspreadingoutbetsfordiversification.Theyshouldbe

recognizedasbigactivecallsthat

needtobemadewithconviction.

Moreover,iftheAIthemestumbles,theimpactwilllikelydwarfany

seemingdiversificationawayfromit.

Traditionaldiversifierslikelong-termTreasuriesdonotoffertheportfoliobaIIasttheyoncedid.That’swhy

investorshavebeenhuntingfor

otherportfoliohedges-reviving

interestingoldthisyear.Weseeit

moreasatacticalplaywithdistinctreturndriversandareskepticalofitsroleasalong-termportfoliohedge.

Wethinkdiversificationmeans

shiftingawayfrombroadasset

classesorregionalviewstomore

granular,nimblerpositioningandthemesthatworkacrossscenarios.PortfoliosneedaclearplanBandareadinesstopivotquickly.

Wedon’tseeeasypassive

diversificationoptionsinthis

environment.Wethinkinvestors

shouldfocuslessonspreadingriskindiscriminatelyandmoreon

owningitmoredeliberately-in

short,amoreactiveapproach.We

thinkhigh-convictionstrategieslikeprivatemarketsandhedgefunds

pairdiversificationwithstrong

returnpotential,inourview.And

muchoftheirperformanceis

idiosyncratic,allowingmanagerskilltoshowthrough.

20152017201920212023

Charttakeaway:Aftercontrollingforequitystylefactors,theshareofdailyS&P500returnsexplainedbyasinglefactorhasjumpedsincelastyear-asignofaconcentratedmarket.

Thefigureshownrelatestopastperformance.Pastperformanceisnotareliableindicatorofcurrentorfutureresults.Indexperformancereturnsdonotreflectanymanagementfees,transactioncostsorexpenses.Indicesareunmanagedandonecannotinvestdirectlyinanindex.Source:BlackRockInvestmentInstitute,withdatafromBloombergandKennethR.French,December2025.Note:ThelineshowsthevarianceofdailyS&P500stockreturnsexplainedbyacommondriverafteraccountingforfactorslikestockvalue,sizeandmomentum.Thiswascalculatedusingfirstprincipalcomponent(PC1)ofaprincipalcomponentanalysis(PCA)overa

rolling252-daywindow.PC1attemptstodeterminethecommondriverinasetofnumbersthatchangeoften

-likestockreturnsforexample.

Investmentimplications

?Traditionaldiversifierslikelong-datedbondsofferlesspotentialportfolioballast.

?Thisenvironmentcallsforseekingtruly

idiosyncraticreturnsources,suchasprivate

marketsandhedgefunds,andstayingtactical8

AIbuildoutfacesconstraints

Facingconstraints

U.S.datacenterpowerdemandasashareoftotal,2024-30

Shareofcurrentdemand

25%

20%

15%

10%

5%

0%

Demand-based

Supply-basedandother

TheAIbuildoutintheU.S.faces

manyconstraints:political,

financialaswehavenoted,but

especiallyphysicalconstraints

involvingcomputeandenergy–

withenergythemostbindingnow.

AIdatacenterscoulduse15-20%ofcurrentU.S.electricitydemandby2030–ascalesuretotestthe

limitsofpowergrid,fossiland

materialsindustries.Seethechart.Someestimatesevensuggestdatacenterscoulduseaquarterof

currentelectricitydemand.Risingpowerdemandcollideswitha

backlogofprojectswaitingto

connectwiththeelectricgridandgenerallyslowpermittinginthe

West.Capexplanscouldbewalkedbackifsuchlimitsbite,crimping

ambitionaroundtheAIbuildout.

Chinastandsincontrast,buildingpowergenerationandtransmissionatpace.Thatincludeson-timeandon-budgetnuclearreactorsplus

coal,hydropowerandrenewables.China’ssolarandbattery

manufacturingisdrivingdomesticsolarandbatterycostsdown.

WebelievethatgivesChinaa

numberofadvantagesindeployingAIinfrastructure:power-intensive

datacentersneedcheap,stableandcleanenergy.Thefocuson

DeepSeekthisyear,aseemingly

moreenergy-efficientChineseAI

model,showedthatgreatlyreducingtheenergyneededforcompute

couldpartlyeaseenergyconstraints.

Weseeopportunitiesemerging

whereconstraintswilllikelybitethemost:powersystems,grids,criticalmineralsandpotentialbeneficiariesfrompermittingreform.Private

capitalwillbecentraltobridgingthegapbetweenthecurrentsupplyof

energyandfuturedemandas

elevateddebtlimitsgovernmentspending,inourview.

Companieshaven’tstruggledtogetchips—therealconstraintislandandenergy.”

2024202520262027202820292030

Charttakeaway:TheelectricityneedsofdatacenterscouldmakeupaquarterofcurrentU.S.electricitydemandby2030—withdemand-basedestimatesexceedingthosefactoringin

supplyconstraints.Thatunderscoresthescaleofthechallengeandtheuncertaintyaroundhowthedemandwillbemet.

Forwardlookingestimatesmaynotcometopass.Source:BlackRockInvestmentInstitute,basedonsources

reviewedbyBloombergNEF,December2025.Note:Greybarsshowthetotalrange,whilelinesdenotethemedianofdatacenterpowerdemandforecastsbymethodasashareof2024U.S.totalpowerdemand,

includingtraditionaldatacentersandartificialintelligence(AI)datacentersandexcludingcryptocurrencyanddatatransmissionnetworksdemand.

Investmentimplications

?Majorchokepoints–likepowersystemsandthepermittingofland–couldconstraintheAI

buildout.

?Weseeprivatecapitalplayingacriticalroleinfinancingfutureenergyprojects.

9

AlastairBishop

Portfoliomanager,

BlackRock

FundamentalEquities

Fragmentation:AIanddefense

Europestepsupspending

Defensespending,1960-2027

ShareofGDP

5%

4%

3%

2%

1%

0%

Estimates

Previous2%target

GermanyNATO-EU

WehaveenteredthethirddistinctworldordersinceWorldWarTwoastheU.S.fundamentallyresetsits

economicandgeopolitical

relationshipswiththeworld.

Transactionaltrade,industrial

policyandaresettingofalliancestoadvanceaseriesofU.S.interestsmarkadecisivebreakfromthe

post-ColdWarorder.

TheU.S.-Chinarivalryisthe

definingfeatureofgeopolitics.

Competitionspanstrade,

technology,energyanddefense.AIisatthecenter.BothseeAIasthekeytechnologydeterminingfutureeconomicandmilitaryadvantagethiscentury.Andbothare

undertakingmajoreffortstocurbstrategicdependenciesthatcanbeweaponized.

Elsewhere,anongoingwarof

attritioninUkraine,anuncertainpeaceinGazaamidstructural

regionalchangeandthelargestU.S.build-upofforcesinthe

Caribbeanindecadesshowhowregionalconflictsandcrisescanlayervolatilityatopastructurallyelevatedriskenvironment.

InEurope,risingU.S.protectionism,weakcompetitivenessandRussia’sinvasionofUkrainehaverestored

urgencyinaddressingEurope’s

challenges,withtheEUembracingpartsoftheDraghireporttoguidepolicy.NATOalliesagreedto