德勤：2024年全球半導(dǎo)體產(chǎn)業(yè)展望

Deloitte.

2024globalsemiconductorindustryoutlook

2

2024globalsemiconductorindustryoutlook

What’sinside

Executivesummary3

GenerativeAIandthenextwaveofchipsupply4

Makingsmartmanufacturingsmarter6

Semiconductorassemblyandtestventureintonewgeographies8

Bolsteringcybersecuritytocombatintensifyingcyberthreats10

Geopolitics,exportcontrols,semiconductors,

advancednodes,andAI11

Signpostsforthefuture12

AboutDeloitte’sOutlooks

Deloitte’s2024globalsemiconductorindustryoutlookseekstoidentifythestrategicissuesandopportunitiesforsemiconductorcompaniesandotherpartsofthesemiconductorsupplychaintoconsiderinthecomingyear,includingtheirimpacts,keyactionstotake,andcriticalquestionstoask.Thegoalistohelpequipcompaniesacrossthesemiconductorsupplychainwiththeinformationandforesighttobetterpositionthemselvesforarobustandresilientfuture.

2024globalsemiconductorindustryoutlook

Executivesummary

utilizationwashighduringtherecentshortage(inthemid-90%range)andisexpectedtofallbelow70%inQ42023.9Theindustrylikelyneedsutilizationtobemuchhigherthanthattobeprofitable,whichcouldtakesometime.Meanwhile,capacityisalsogrowingastheUnitedStatesandEuropeincreasedomesticchipmaking.

Thesetrendsandothersplayintoour2024globalsemiconductor

industryoutlook,wherewedrilldownintofivebigtopicsfortheyearahead:generativeAIacceleratorchipsandhowsemiconductor

companiesareusinggenAI;trendsaroundsmartmanufacturing;

theneedformoreassemblyandtestcapacityworldwide;howchip

industryintellectualproperty(IP)isatargetforcyberattacksatawholenewthreatlevel;andafinalgeopoliticssectionthatlooksatexportcontrolsaroundadvancednodemanufacturingequipmentand

technologies,aswellasadvancedgenAIsemiconductors.

TherearesomeinterestinganglestothegenAIchipboomthataffectthechipindustryasawhole.WithsalespredictedtoreachmorethanUS$50billionin2024,10thismarketisatailwindforthesectorand

isexpectedtoaccountforabout8.5%ofsales.Aportionofthat

willcomefromlogicprocessorsmadeonadvancednodes,someofitfromadvancedhighbandwidthmemory(HBM3),somefromadvanced2.5Dpackaging,plussomefromadvancedconnectivitychips.Ineachcategory,thesegenAI-drivenchipsareamongthepriciestoftheir

kind.In2022,morethanatrillionchipsweresoldatanaveragesellingpriceofUS$0.57perchip.11Meanwhile,somegenAIchipsweresellingforUS$40,000eachin2023,or70,000xhigher,andthereforeUS$50billionworthofchipsmightonlybeavolumeof1.25millionchips,orlessthan0.1%oftotalchipvolumesfortheyear.12

Whydoesthatmatter?AlthoughgenAIchipsareexpectedtobeabigpartof2024chiprevenues(andlikelytobeevenbiggerinthe

future),theyarearelativelysmallpartofunitvolumes,andthereforemanufacturingcapacity.Asmentionedabove,withindustryutilizationbelow70%,havinggenAIchipsflyofftheshelvesisgreatnewsforthehandfulofcompaniesthatsellthosechipsorpartsofthosechips,butitmaynotbeashelpfulfortheoverallindustry.

Fortheindustrytoseeoptimalutilizationacrossallprocess

nodes,otherkindsofchipsinadditiontogenAImayneedtoseestrongerdemand.

Chipsaleslooktobouncebackin2024,ledbygenerativeAI,butcouldbecomplicatedbygeopolitics

Theinfamouslycyclicalsemiconductorindustryhadachallenging

yearin2023,theseventhdownturnsince1990,withsalesexpectedtobedown9.4%(toUS$520billion)fortheyear.1Butthat’snotas

badaswasexpectedinthespring;beforerelativelystrongersecond

andthirdquarterswereinthebooks,thepreviousforecasthad

beenforUS$515billion.22024isnowpredictedtoseeglobalsalesofUS$588billion.3Notonlywouldthatbe13%betterthan2023,butit’s2.5%higherthan2022’srecordindustryrevenuesofUS$574billion.

Thestockmarketisoftenaleadingindicatorofindustry

performance:Asofmid-December2023,thecombinedmarket

capitalizationofthetop10globalchipcompanieswasUS$3.4trillion,up74%fromUS$1.9trillioninNovember2022and17%higherthantheUS$2.9trillionwesawinNovember2021.4

Asissooftenthecase,thememorychipmarketwasthebiggest

swingfactor.In2022,memorysaleswerealmostUS$130billion,

orjustunder23%oftheoverallchipmarket,buttheydropped31%(aboutUS$40billion)in2023,comparedtodown1%forlogic.Themarketisexpectedtogetalmostallofthatbackin2024,withsalesexpectedtoreach2022levels.Ifweexcludememory,therestof

theindustrywasdownin2023,butonlybyabout3%.5

Intermsofendmarkets,bothPCandsmartphonesalesare

expectedtogrow4%in2024,after2023declinesof14%and3.5%,respectively.6Returningtogrowthforthesetwoendmarketsis

likelyimportantforthesemiindustry:In2022,communicationandcomputerchipsales(whichincludedatacenterchips)madeup56%ofoverallsemiconductorsalesfortheyear,comparedtoautoandindustrial,whichaccountedforonly14%ofsaleseach,forexample.7

Twootherimportantmeasuresoftheindustry’shealthare

inventoriesandfabutilization.Asoffall2023,inventoriesremainedhighatmorethanUS$60billion,aboutthesamelevelasthe

previousyear.Andtheprocessofdrawingthosedownwillbeasignificantheadwindforsalesinthefirsthalfof2024.8Inaddition,

3

2024globalsemiconductorindustryoutlook

4

GenerativeAIandthe

nextwaveofchipsupply

SellinggenAIchips:

Themarketforchipsthatacceleratethetrainingandinferenceof

generativeAImodelswasthesemiconductorstoryof2023.Atahighlevel,genAIchipsarepackagesofspecialGPUs,specialCPUs,specialHBM3inadvanced2.5Dpackaging,plusotherspecialchipsneededforconnectivityinthedatacenter.13Foranindustryfightingheadwindsfromweakmemoryprices,andweakdemandforsmartphoneandcomputerchips,genAIchipsprovidedagrowtharea,especiallyatleadingmanufacturingnodes.As2024approaches,themarketforthesechipslookstobestrongandispredictedtoreachmorethanUS$50billioninsalesfortheyear,14or8.5%ofthevalueofallchipsexpectedtobesoldfortheyear.15

Inthelongerterm,thereareforecastssuggestingthatAIchipscouldreachUS$400billioninsalesby2027.16Butwhatwillhappenin2024?Ontheonehand,DeloittepredictedinNovember2023thatgenAIchipsaleswillbemorethanUS$50billionin2024.17Ontheother,therearereasonstobelievequarterlygenAIchipsalesgrowthcouldflattenorevendeclineatsomepoint,atleastforawhile.Fall2023wasaperfectstorm:strongdemand,buyerseagertosecuresupply,andrelativelyfewchoices.Butthiscouldchangeatsomepoint

in2024.

1.Newentrants:NewchipsareexpectedfromexistinggenAIchipmakersaswellasemerginggenAIchipmakers.Andnewchipswilllikelycomefromcompaniesthat,upuntilnow,havebeenbetterknownasbuyersofchipsbuthavedecidedtostartmakingtheirown.Nobodyknowshowanyofthesewillfareatthispoint.

2.Newarchitecturesandmodels:Therearemanydifferent

genAImodelsandapproaches,andnosinglechiparchitectureislikelytobeoptimalforallusecases.Therewilllikelybedatacenterchips,edgechips,trainingchips,andinferencingchips,andit’sexpectedthatbillionsofdollarswillbespentdevelopingthese“flavors”ofgenAIchips.

3.Edgechips:It’spossiblethatin2024moreprocessingwillbedoneattheedge,onsmaller,cheaper,ordifferentacceleratorsorgenAImodelsthantheonesthatdominatedin2023.

4.Thebullwhipeffect:Thechipindustrycanbehighlysusceptibletoover-orderingandexcessinventoriesintimesofundersupply,whichthenneedstounwindwhensupplycatchesuptodemand,oftenleadingtofallingunitsalesandfallingpricesperunit.Thisbullwhipeffect18couldbeseenbythesecondhalfof2024.

5.WillpeoplepayforgenAI?Finally,muchofthecurrentdemandforgenAIchipsiscomingfromenterprisesoftwarecompanieseitherdirectlyforthosebuildingon-premiseprocessing,

orindirectlyviacloudcompaniesthatprovidethesoftware

companieswithgenAIprocessingservices.Theymayplanon

includinggenAIfeaturesandservicesinsidetheirexistingor

newsoftwareofferingsandhopetochargeforthosefeatures:DeloittepredictsthiswillprovideaUS$10billionrevenueupliftinenterprisesoftwarerevenuesbytheendof2024.19AndtheymaybebuyingupgenAIchipsnow(eitherdirectlyorviathe

cloud)tohelpmeetanticipatedprocessingdemand.Should

softwarebuyersbeunwillingtopayapremiumforgenAIservices,orevenbeslowaboutdoingso,thesoftwarecompaniescouldabruptlyreducetheirordersforgenAIprocessing.

2024globalsemiconductorindustryoutlook

5

UsinggenAIformakingchips:

ThechipindustryhasbeenusingAItoolstohelpdesignchipsforacoupleofyears,butthatwasjustthebeginning.20GenAIcan

helpimproveoperationsandproliferatebestpracticesthroughoutthesemiconductorindustryvaluechain:AccordingtoDeloitte’s

upcomingGenAIinSemiconductorsStudy(2024),72%ofrespondentsbelievethatgenAI’simpactontheirindustrywillbe“highto

transformative.”21Respondentsnarrowedthatfurther,describingthefollowingareaswheregenAImightbeused,inadditiontochipdesignandcodecreation:

?Generatingmoreaccurateschedulesandsupplychainforecasts

?Enhancingresearchanddevelopmentviaresearchaugmentation

?Improvinganomalyanddefectdetection;existingAIsolutionsarealreadyatuseintheindustry,butgenAIoffersthepotentialto

speedthisupbycreatingsyntheticdataformodeltraining

?OperationscouldbenefitfrommanufacturingprocesssimulationandgenAI-enableddigitaltwingenerationthatcansimulate

sorting,assembly,testing,andothercomplexmanufacturingprocesseswithouttheneedforpetabytesofdata

?EnhancingsalesandmarketingeffortsbyusinggenAIforbettercontentgeneration,tailoringmarketingcontenttoaudienceandpurpose

Strategicquestionstoconsider:

?AsgenAIchipmakingcapacityincreases,willpricesand

volumescontinuetoincreaseorcomedown?AtaforecastedUS$400billionin2027,genAIchipswouldbeclosetohalfthechipindustry;whatwouldthatmean?Whatearlysignals

shouldbepaidattentiontofromacrossthechipsupplychain,suchasmovementsinchannelinventoriesororderbuildupsathyperscalers?

?Manychipsegmentsaredominatedbyasingleplayer,whosepreeminentroleisbothstrongbenchmarkperformanceandasetoftoolsandsupportsystemsthatlockbuyersin.WillgenAIchipsfollowthattrend,orwillweseeamorefragmented

industry?Tobepreparedforthevariousscenarios,what

changeswillneedtobemadeinsupplierandvendorcontracts,aswellasorderpipeline?WilltherebeaggressiveM&Aor

privateequityinterestinsmallerchipcompanies?

?GenAItoolsmaytransformtheindustry,butwhatwilltheeffectsbe?Willtheyjustlowercosts,orwillweactually

seerevenuegrowth?

Inspiteofthesebenefits,genAIstillhasitschallenges.Buildingor

buyingcustommodelsforgeneratingcircuitdesign,testingplans,

andsyntheticdatacanbecostprohibitiveascomparedtomanual

execution.Runningcostsforverylargemodelscanalsooutweigh

thespeedbenefitsofautomation.Asiswellknown,genAIcancreateinaccurateornonsensicaloutputs,sovalidationbyhumans-in-

the-loopisneededtoimproveaccuracy.Finally,whenworkingwithhuman-centeredapplicationssuchasHRorsalesandmarketing,humanreviewcanhelpensurethatalldataissanitized,personallyidentifiableinformation(PII)removed,guardrailsinstalled,and

validationperformed.

2024globalsemiconductorindustryoutlook

6

Makingsmart

manufacturingsmarter

Overtheyears,semiconductorfabfacilitiesandoutsourced

semiconductorassemblyandtestfacilities(OSATs)haveleveragedIoTdevices,roboticstech,andartificialintelligence/machinelearning(AI/ML)andanalyticswiththegoalofachievingsmart,lights-outchipfactoriesthatarefullyautomated.22Waferfabequipmentmakers,integrateddevicemanufacturers(IDMs),foundries,andback-endATfacilitiesallcontinuetoinvestmoreinsmartmanufacturing

practices,digitaltools,andtechnologies,butthebasicsmart

manufacturingobjectivesmoreorlessremainthesame.They

connectthingsonthefactoryfloor,automatematerialmovementanddatacollection,andapplyanalyticstopromptdecisionsandactions.However,fromachipmanufacturingperspective,twothingsareexpectedtobedifferentforsmartmanufacturinginthesemiindustryin2024.

OneistheavailabilityofsophisticatedandhighlyadvancedAI

tools(includinggenAI)toanalyzelargedatasetsandoffersharp

insights.Deloitte’supcomingGenAIinSemiconductorsStudynotedroughlysevenin10semiexecsacknowledgegenAIwillhavea

“hightotransformative”impactontheirbusiness,indicatingtheyviewitasmorethanjustanothernewtool.23Acrossmanufacturing,operations,andmaintenancefunctions,semiexecssurveyedbelievegenAIhasthepotentialtoaddthegreatestvaluethroughanalysesandinsightspertainingtoprocessesandequipment(notedby28%ofrespondents),followedbypredictivemaintenanceandsmart

diagnosticsandtroubleshooting(18%).

Thesecondrelatestoenhancingperformanceandsustainability

offabsandbuildings.Movingfromamaturetechnologynode(e.g.,28nmmanufacturing)toanadvancednode(2nm)needs3.5timesasmuchenergy,consumes2.3timesasmuchwater,andemits

2.5timesasmuchgreenhousegases.24Thesemiindustryshouldconsiderimplementingmanufacturingtransformationonolder

plants(brownfields),inadditiontobuildingbrand-newgreenfieldplants,tohelprealizeevengreatersustainabilitybenefits.25

Moreover,fabscanlookatinvestinginsmartmanufacturingtoolssuchas6DBIM(buildinginformationmodeling)tohelpimprovecostmanagement,simulateandanalyzeenergyconsumption,enhanceefficiency,andstreamlinefacilitymanagement—makingapositiveimpactonthebottomlineandtheplanet.26

Asfabsexploreusingsmartmanufacturingtoolsandvirtualmodelssuchas6DBIMthroughout2024,theyshouldidentifyspecific

environmentalandsocialfactorsandmetricstobemeasuredforamoreholisticsustainabilityassessmentandenvironmental,social,andgovernance(ESG)reporting.Additionally,companiesshouldconsiderexploringwhereandhowtechnologiesandtoolssuchasgenAI,private5Gnetworks,anddigitaltwinscouldbeintegratedinthemanufacturingplants.27

2024globalsemiconductorindustryoutlook

7

SemiconductorexecutivespolledaspartoftheDeloitte2023

SemiconductorTransformationStudyidentifiedthatlackofenterprise-widealignmentondigitalcapabilitiesisakeychallengeaspartof

theirtransformationinitiatives.28ThoughthelevelofbusinessandITalignmentshowedimprovementinthe2023survey(56%)comparedwiththeprior2021survey(37%),itmayneedtoberevisitedin

2024.29Companiesshouldcontinueinvestinginsmartdatabases

andsystemstoacquiredisparatesetsofdataacrossbusinessunitsandanincreasinglycomplexandextensivesupplychain.Theyshouldbolstercapabilitiesindatamodernization,unifieddataplatforms,

advancedanalytics,andnext-generationSaaS-basedapplications(e.g.,procurement,planning,customer,andenterpriseresource

planning)—allofwhichmayrequirebetterintegrationwiththeirmanufacturingexecutionsystem(MES),withemphasisonclean,accurate,andgoodqualitydata.Andtheseaspectsarevitaltohelpenablecompaniestobenefitfullyfromtheirpriorinvestments

madeinAI/MLanddatamanagementsolutions.

EvenasfabsandATfacilitiesincreasethedegreeofautomation

andusesmarttechnologies,theyshouldlooktodeterminewhat

specializedskillstheirworkersandoperatorsshouldpossess—

especiallythosewhooverseethemachinery,equipment,processes,andcontrolsystems.2024couldbetheyearwhenthesemiindustrytakesstockofwhereandhowthelevelofhumaninvolvementcanbelimited—atleastforsomeperiodoftimeinselectaspectsofthefaboperations—totrulyachieve“smart”lights-outfactoriesthat

operateefficientlyinthedark.

Strategicquestionstoconsider:

?Aspartofadigitalvision,whataspectsshouldbeconsideredbyanenterprise—includingdesign,production,sales,

support,andproductend-of-life/recycling—tomakeitaholisticdigitalapproach?

?Tomakethevarietyofmanufacturingequipment,machinery,andtoolsinthefacilitiestalktoeachotherandbenefitfromthebroaderdatamodernizationeffortsacrosstheother

businessunits(e.g.,insupplychain,finance,procurement,sales),whatotherline-of-businessapplicationsshouldtheproductionsystemsandprocessesbeintegratedwith?

Whatcommunicationprotocolsshouldbestandardized?Andwhatinterfaceprogramsshouldbedeveloped?

?Inviewofthegoaltobelights-out,howcanthe

next-generationsmartfactoryimplementationbemorehuman-centric,andwhatspecificaspectsofmanufacturingprocessesshouldhavehumanworkersdoingthejob

alongsidemachines?Additionally,howcancompanies

integrateworkerwell-beingtomakethesemiindustrynotonlysustainablebutmorehuman-centric?

2024globalsemiconductorindustryoutlook

8

Semiconductorassemblyandtestventureintonewgeographies

Morethan75%oftheglobalsemiconductorfabcapacityisinAsia

(thefront-end),30buttheregion’smarketshareisevenhigher

(90%)inchipassemblyandtesting(theback-end).31Exceptfor

largeIDMs,mostchipplayershavebeenoutsourcingATprocessestothird-partyvendors,orOSATs.32ThemajorityofthebigOSATs

arebasedinChinaandTaiwan,commandingroughly80%ofthe

OSATmarketsharein2022.AlthoughtheUnitedStatesisaiming

tobolsterdomesticATcapacity,33almostallactualATworkisdoneinAsia.34Thelinesbetweentraditionalfront-endandback-endareincreasinglyblurring,witheachattemptingtocapturemoreofthevaluechain.Advancedpackagingisalsoincreasinglybecominga

strategicenablertobuildthemostsophisticatedleading-edgechips.35

Goingforward,astheUnitedStatesandEuropelooktoexpand

domesticchipfabricationcapacity,theyshouldlooktobuildup

theirback-endcapacitytoavoidlengtheningandmakingtheirsupplychainsmorecomplex.Tohelpstayontheleadingedgeofproductperformanceandflexibility,IDMsintheUnitedStatesandSouthKoreaareincreasingeffortstobolstertheirpackagingcapabilities,whichareusuallyprovidedandenabledbytheirrespectiveassemblyoperationsandfacilities.Concurrently,leadingfablesscompaniesarepushingfornearshoreAT.Further,complexgenAIchipsare

fuelingdemandforadvancedpackaging,exposinganacutecapacityshortageforthistechnology.36

Deloitte’s2023SemiconductorTransformationStudynotedthatstrongdemandforAIandcompute-andmemory-intensiveapplications

willspurinnovationinadvancedpackagingtechniquessuchas

2D,2.5D,and3D.37Tokeeppacewiththegrowingpackaging

innovationneeds,boththeEUChipsActandtheUSCHIPSand

ScienceActhavefundsallocatedforadvancedpackagingtechnologydevelopment.38However,asnotedinthe2024PredictionGenAIchipdemandfansasemitailwind,fornow,governmentsmayneedtoaddtoexistingincentiveprogramstoexpanddomesticornearshore

ATmanufacturingcapacity.39Packagingisbecomingmorecomplex,

oftenrequiringnewtechniquessuchasdatamodelingandsimulationstoproactivelyhelpdiscoverandpredicterrorsandotherpackaging

issuesinadvance.

2024globalsemiconductorindustryoutlook

9

In2024,theback-endATmarketcouldexperiencesignificant

transformation,asprominentIDMsandfoundriesmoveevenfurtherintoadvancedpackaging,whiletraditionalOSATsalsocontinuetoenhancetheirpackagingcapabilities.Simultaneously,US-andEU-basedsemiconductorcompaniesareexpandingtheirfront-endwaferfabfacilitiesontheirhometurfs.40Alongsidethisexpansion,steps

arebeingtakentoshifttheirback-endATservicestonewcountries.Forinstance,newATcapacityisbeingbuiltinVietnam,Malaysia,

India,andPoland,41reflectinghowIDMsandOSATsarediversifyingandde-riskingtheirsupplychain;thistrendisinlinewithDeloitte’sperspectiveinthe2023globalsemiconductorindustryoutlook.42

ButtheemergingATfacilitiesfacedistinctchallenges.Newadvancedpackagingtechnologiesandtestsolutionsshouldbedelivered

withhigh-qualityperformancewithinstringenttime-to-marketconstraints.Also,suchtechnologiesoftenrequiredistinctskillsandexperiences.Forexample,packagingandtestingengineersneedtohavespecializationsinelectricalandelectronicsengineering,materialsciences,capacityplanning,andyieldprocesses.43

Additionally,back-endplayersarechallengedtoprovidearangeofnovelbutintricateadvancedpackagingoptions;forexample:2.5D/3D,fan-outs,chiplets,SiP,andhybridbonding.44

In2024,IDMATunitsandpure-playOSATscouldlooktoshortlistfromthoseseveraloptions,andgainmasteryofspecificpackagingtechnologies.Theyshouldbeagileandconstantlyinnovateto

helpallowbrandedsemiconductorcompaniestolaunchsuperiorproductsmorerapidlyandatcompetitiveperformancesandprices.OneotheraspectthatATfacilitiesshouldconsideristheenergy,materials,andotherresourcesusedinassembly,test,shipping,anddistributionoperations—whichareoftenequallyimportantpartsofthesemiconductorsustainabilityequation.45

TostaycompetitiveinthedynamicATlandscapethroughout2024andbeyond,OSATsandcaptiveATfacilitiesshouldstrengthentheircoreenterpriseITsystems.Additionally,integratingAIandMLintotheiroperationscanhelpdevelopadvancedpackagingtechnologiesandfeatures,improvedemandplanning,manageinventoryeffectively,andstreamlineinformationflowacrosstheextendedsupplychain.Testingisalsoexpectedtogainprominence,ascomplexchipand

moduledesignscouldrequirecaptiveATandOSATstoadvancecapabilitieslikesystem-leveltest,adaptiveordynamictest,andAI/ML-basedbinprediction.

Strategicquestionstoconsider:

?Whichpeercompaniesandacademicinstitutionsshould

companiespartnerwithtotakeadvantageofgovernment

incentivesandsharedresearchinfrastructuretocollaborativelybuild,test,andpilotnext-generationinnovativepackaging

approaches?

?WhatpartsofATprocessesshouldbereshoreddomestically,whichonesshouldbeoutsourced,andwhatcountries(withinSoutheastAsia,SouthAsia,EasternEurope,LatinAmerica)shouldbeconsideredtobuildnewback-endfacilities?

?Inback-endassembly,howshouldmetals,plastics,andothermaterialsbehandledaspartofbroadersustainabilityeffortsandtocomplywithESGregulatoryneeds?

?ShouldmorefoundriesandIDMsintegratepackagingintotheirvaluechain?Willtheincreasedstrategicrelevanceofpackagingmanifestitselfthroughnewalliancesand

partnershipsthatfablessandIDMscouldconsider?

2024globalsemiconductorindustryoutlook

10

Bolsteringcybersecuritytocombatintensifyingcyberthreats

Thesemiconductorindustryfacesadifferentlevelofcyberthreatscomparedtootherindustries.Inadditiontotheusualprofit-seekingransomwareattacksthateveryindustrydealswith,semicompaniespossessunique,valuable,andrestrictedIP.Duetotheincreasing

importanceofsemiconductorsformultipleindustries,it’soften

targetedbystate-backedactors.Asaresultofgeopoliticalissuesandrestrictionsonadvancedchipmakingtech,theIPofsemicompaniesisoneoftheworld’smostimportanttargetsforcyberattacks.46

Ifgeopoliticaltensionscontinuetoescalatein2024—resulting

infurtherrestrictionsaroundIP,chips,andrawmaterials—

cyberattacksmayintensify,disruptingproductionintheindustry.

Cyberthreatactorsarenotonlytargetingthecoresemicompanies,butattacksareincreasinglydirectedatextendedchannelpartners,too(e.g.,suppliers,distributors,andcontractmanufacturers).47Sophisticatedthreatactorscoulduseadvancedmethods,like

maskingasransomwaregroups,48tofabricateattacksandcausebusinessdisruptionsacrossthesupplychain.

Here’swherethesemiconductorindustrycouldexperienceanasymmetricbattlein2024:Theindustrymayneedtofaceoff

withsophisticatedthreatactorsthathavemuchmoreadvancedresourcesthanotherindustrialcyberthreats.Thisasymmetricbattlebetweenstate-backedactorsandbroadersemiindustryparticipantscouldpresentnewchallengestoindustryexecutivesin2024andbeyond.

Therefore,in2024,semicompaniesshouldconsideracceleratingtheireffortstobolstercyberdefensecapabilitiesoftheirown

aswellasthedigitalandcyberinfrastructureoftheirextendedsupplychains.

Companiesshouldbemoreagileinrealigningtheircyberprogramswiththeseemergingthreatsandinvestincustomerdataprotectionsolutions.Forinstance,Deloitte’s2023GlobalFutureofCyberSurveynotedthatmorethan75%ofrespondentsbelievedthatinvestingincybersecuritystrategy,cybercloud,anddataprotectionsolutionsgeneratedthehighestvaluefortheirorganizations.49Inaddition,

semicompaniesshouldlookatusingartificialintelligencetohelpcombatcyberthreats:AIcannotonlyenablecybersecurityteamstorespondfasterthanthepaceofcyberattackers,butusinganAI-basedanddata-drivenapproachcouldalsohelpsenseand

anticipatethreatsandcountercyberattacksproactively.50

ApartfrombolsteringtheirownITinfrastructure,semicompanies

shouldhelpsuppliersacrossregionsbolstertheircyberdefenses

aswell.Thiscouldallowthebroadersemiconductorecosystemto

operatewithagr