STUDY

AIASGAMECHANGER

TheNewDrivingForceoftheAutomotiveIndustry

person

Authors&Contact

Lead

AugustinFriedel

SoftwareDefinedVehicles

Augustin.Friedel@

Lead

MatthiasBorch

ArtificialIntelligence

Matthias.Borch@

ContactPerson

StephanBaier

ArtificialIntelligence

Stephan.Baier@

Author

MarcusWilland

Mobility

Marcus.Willand@

Author

Dr.NilsSchaupensteiner

TransformationAdvisory

Nils.Schaupensteiner@

Author

PatrickRuhland

TransformationAdvisory

Patrick.Ruhland@

AIasGameChanger

Thestudy“AIasGameChanger“anditssummarywerepublishedby:

MHPGesellschaftfürManagement-undIT-BeratungmbH

Allrightsreserved!

Noreproduction,microfilming,storage,orprocessinginelectronicmediapermittedwithouttheconsentofthepublisher.Thecontentsofthispublicationareintendedtoinformourcustomersandbusinesspartners.Theycorrespondtothestateofknowledgeoftheauthorsatthetimeofpublication.Toresolveanyissues,pleaserefertothesourceslistedinthepublicationorcontactthedesignatedcontactpersons.Opinionarticlesreflecttheviewsoftheindividualauthors.Roundingdifferencesmayoccurinthegraphics.

3

4

Contents

Contents4

Tableoffigures6

12KeyFindings8

WelcometoChange!10

01.RevolutionandAutomotiveMarketPotential11

02.InvestmentinCompaniesWithanAIFocus15

03.PilotProjectsandImplementation19

04.AIModels,Levels,andUseCases23

4.1TheGameChanger:WhatCanBeAchievedWithAI26

4.2AutomobileManufacturersWithLowAIInvestment29

4.3AIModels:MakeorBuy?29

05.AIApplicationsAlongtheAutomotiveValueChain31

5.1AIOperationinVehiclesandintheCloud35

5.2AIMonetizationinVehicles39

5.3AddedValueofAIApplicationsinCompanies40

06.WhattheCustomerWants:TheUserPerspective47

6.1UseandUnderstandingofAIApplications49

6.2AdvantagesandDisadvantages–GenerallyandinVehicles49

6.3PurchasingDecision,TrustandWillingness

toPay51

AIasGameChanger|Contents

07.SuccessFactorsandStrategicApproach55

7.1StrategyandGoalPlanning56

7.2ThinkfromthePerspectiveoftheCustomer,nottheTechnology56

7.3OrganizationalAnchoringandOwnership58

7.4LocalDifferencesrequirelocalSetup59

7.5ReducingComplexity59

7.6Useand

MonetizationofData60

7.7ChecklistforsuccessfulImplementation61

08.Challenges,Responsibility,andRisks63

8.1CostsofTraining

andOperation64

8.2Dataand

DigitalizationasaBasis65

8.3BusinessModelsandCasesforB2CandB2B65

8.4EthicsandResponsibility67

8.5NewRisksandRegulatoryChallenges69

09.AIApplicationsintheAutomotiveIndustry:7RecommendationsforAction71

10.FurtherInformations75

LiteratureandSources76

Contact

International78

About

MHP79

5

6

Tableoffigures

Figure1:Technologysupercycles–artificialintelligenceasthenextrelevantplatformshift

(Coatue,2024)12

Figure2:AImarketsizeintheautomotivesector(PrecedenceResearch,2024)12

Figure3:TotalinvestmentsinAIcompaniesfoundedsince2001,inUSDbillion(Scheuer,2024)16

Figure4:InvestmentinAIstacklayers(Coatue,2024)17

Figure5:CompanieswithteamandbudgetforAI(Capgemini,2023)21

Figure6:InterconnectedAIconcepts24

Figure7:VisualizationofAIasapyramid25

Figure8:ClassificationofAIterms27

Figure9:TheperformanceofAImodelscomparedtohumancapabilitiesintheMMLUtest(iAsk,2024)28

Figure10:SchematicdiagramofthetrainingofAIfoundationmodelsforvehicles30

Figure11:UseofAIalongthevaluechain32

Figure12:SignificantimprovementsoffunctionsandfeaturesthroughAI33

Figure13:InterestinAIfunctionscomparedinternationally34

Figure14:Roleofon-premise,cloud,andvehicleforAImodels35

Figure15:Levelsofasoftware-definedvehicle(SDV)(Willand,Friedel,&Schaupensteiner,2023)36

Figure16:DifferentmodelsforADASandADapplicationsandfunctions37

Figure17:AI’spotentialatdifferentstagesofthevaluechain

(Capgemini,2023)40

Figure18:UseofAI-basedsolutionsbyregion41

Figure19:KeydriversbehindtheuseofAIinproduction42

7

AIasGameChanger|Tableoffigures

Figure20:Decisiveissue–fewerusersofsoftwareduetoAIorfreesoftware(Coatue,2024)43

Figure21:PossibleusesofAIinsoftwaredevelopment

(Wee2024)44

Figure22:UnderstandingofAIincars48

Figure23:AdvantagesofusingAIincars49

Figure

24:TheperceivedadvantagesanddisadvantagesofusingAI50

Figure

25:AIincars:purchasemotivationorblocker?51

Figure26:TrustinstakeholderswithregardtotheimplementationofAIinvehicles52

Figure27:WillingnesstopayforAIfunctions52

Abb.28:AssessmentofthefutureAIcompetenceofcarmanufacturersbyregion53

Figure

29:Customerandusecasefirst,andthenAIapplicationsandmodels57

Figure30:Dimensionsforvalidatingtechnicalfeasibility57

Figure31:TrainingcostsforAImodelsareincreasing(StanfordUniversity,2024)64

Figure32:Dataavailabilityandqualitybyregion65

Figure33:Customers’willingnesstopayisunclear;costsariseforimplementationandoperation66

Figure

34:ClassificationofAIusecasecategoriesandpossiblebusinessmodels67

Figure35:RisksassociatedwiththeuseofAI68

Figure

36:PrinciplesandpenaltiesoftheEUAIAct70

Table1:ThedevelopmentofAImodelsdividedintodifferenttimephases27

12KeyFindings

ThewidespreaduseofAIispredictedtobethenextrelevantplatformshiftaftercloudtransformation–originalequipmentmanufacturers(OEMs)needtostepuptheiractivities.

Morethan

50

ofrespondentsseetime-savingasthebiggestbenefitofAIapplications.

SkepticismaboutAI

applicationsisgreaterin

theUSthaninEurope

orChina.

Only

ofrespondentsinChinastatethattherisksofAI

outweighthebenefits;thisfigureisaround25percentinEuropeandtheUS.

Themostfrequentlymentioneddisadvantagesof

AIarefearoflossofcontrol,lossofdataprotectionandpersonalprivacy,andsecurityrisks.

8

CustomersworldwidewanttouseAIincars,butrarelypayforit.

InChina,morethantwiceasmanycustomershavealreadyusedAIintheircarsasin

Europe.

KI

InChina,AIfunctionsmostlyhaveapositive

influenceoncarpurchasing

decisions–only

ofrespondents

wouldnotbuy

avehiclebasedonAIfunctions.

Today,Chinesecar

manufacturersareregardedasleadersinAIinnovation.Infiveyears’time,JapaneseOEMswillbeattheforefront,followedbyChineseandGermanOEMs.

AIisnotonlyrevolutionizingthein-vehiclecustomerexperience

–theentirevaluechainis

experiencingdisruptivechange.

TraditionalcarmanufacturersarethemosttrustedwhenitcomestotheuseofAI,far

aheadofstateinstitutionsandnewcarmanufacturers.

SuccessfulimplementationofAIapplicationsisnotpossiblewithoutpriordigitalizationandaccesstospecificdatasources.

AIasGameChanger|12KeyFindings

9

10

WelcometoChange!

Dearreaders,

Artificialintelligencewillbethenextplatformshiftthatrevolutionizesallindustrialsectors.StakeholdersintheautomotivevaluechainhaverealizedthatAIischallengingmanytradi-tionalprocessesandwaysofthinking.TheintroductionofthePC,thestationaryInternetandthenthemobileInternet,andCloud/SaaSpreviouslyhadasimilarlydisruptiveimpact.Newbusinessmodelsandprofitpoolsareemerging,whileatthesametimetherearenu-merouschallengestobetackledwithregardtotechnology,partnerships,andethicalissues.Inthisstudy,wetracethegroundbreakingdevelopmentsinAIsofarandexaminetheop-portunitiesandriskswithintheautomotiveindustry.Accompanyusthroughpresentandfuturescenarios–withspecificrecommendationsforactionforyourownstrategywhenitcomestoimplementingAIapplicationsinproductionandinvehicles.

Whetherthenewtechnologiesmeettheexpectationsofdriversisdeterminedrightthereinthecockpit.That’swhy,inChapter8,weoutlinetheuserperspectivebasedonourowncurrentdata.OurinternationalsurveyprovidesinformationaboutwhichproductsandservicesfromautomotivecompaniescouldfulfillAIneedsandwhatthewillingnesstopaylookslike.Thatmakesthisstudyessentialreadingfordecision-makers,CIOs,andapplica-tiondevelopers.

InvestorsinAItechnologiesandAIteamsneedaconsistent,long-termcost-benefitratio.Wethereforeexaminethedirect/indirectmonetizationofin-carAIandlookatnewbusinessmodelsbasedonAIanddigitalization.

Ultimately,asissooftenthecase,itbecomesclearthatthejourneyintonewtechnologicalterritoryisbestundertakenwithexperiencedtravelguides.Gettheknow-howyouneed–andalwaysbecurious!

ENABLINGYOUTOSHAPEABETTERTOMORROW

Bestregards,

Dr.JanWehinger

ClusterLeadSoftwareDefinedVehicles

MHPManagement-undIT-BeratungGmbHLudwigsburg,September2024

AIasGameChanger|01.RevolutionandAutomotiveMarketPotential

01.

Revolutionand

AutomotiveMarketPotential

11

EveryonerecognizesthatAIisthenextplatformshift

Mobile Internet(Web2.0)

Cloud/SaaS

GenerativeAI

Desktop Internet(Web1.0)

Networking

PC

Mainframe

1960–19801980s1990s2000s2010s2015–20202022–...

Figure1:Technologysupercycles–artificialintelligenceasthenextrelevantplatformshift(Coatue,2024)

AI-Basedsystemsforautomotiveindustrytoreach

US$35.7billionby2033

35.7

26.6

20.015.2

...inbillionUS$

11.79.2

3.23.94.75.87.3

20232024202520262027202820292030203120322033

Figure2:AImarketsizeintheautomotivesector(PrecedenceResearch,2024)

12

ItishighlylikelythatthebigtechnologycompaniessuchasGoogle,Meta,andMicrosoft–whichgainedinimportancewiththelastplatformshifts(supercy-cles)–willalsodominatetheAIage.

Alongtheautomotivevaluechain,stakeholdersaresometimesaccusedofhavingrespondedtothelastplatformshiftstoolateorwithanineffectivestrategy.Inouropinion,therelevanceofconnectivityandcloudsolutionswasrecognizedtoolateandimplementationcouldhavebeenbetter.Theindustryisatthebegin-ningoftheAIplatformshiftandthereisstilltheop-portunitytorespondearlywithatargetedstrategy.CompanieslikeApplehaveshownthatitisnotneces-

Onefear,however,isthatartificialintelligencewillincreasinglyreplacepeopleandjobsmaydisappear.Currently,AIapplicationsareregardedmoreasacom-plementratherthanareplacement.AcademicssuchasKarimLakhanifromHarvardBusinessSchoolbelievethatAIwillnotreplacehumans.OnepossiblescenarioisthatpeoplewhouseAIwillhaveasignificantadvan-tageoverworkerswhodonotuseit.

RegardingthequestionofwhetherAIwillimprovetheeconomy,asurveyshowsamixedpicture.Worldwide,34percentofrespondentsbelievethattheuseofAIwillimprovetheeconomicsituationintheircountryinthenextthreetofiveyears.Thishopeisaboveaverage

“AIWon’tReplaceHumans—

ButHumansWithAIWillReplaceHumansWithoutAI.”(HBR,2023)

sarytobethefirstinnovator.WithastrongAIstrategy,acompanycanalsoexploitpotentialasafastfollower.Themarketforartificialintelligenceintheautomotiveindustryhasshownremarkablegrowthinrecentyears.ItiscurrentlyestimatedtobearoundUSD3.9billionin2024andisexpectedtogrowtoUSD15billionby

2030.SomemarketanalysesanticipatethatAIsalesintheautomotivesectorwillrisetooverUSD35billionin2033.Growthfrom2024to2033correspondstoarateof28percent.

Estimatesinothermarketreportsmaybeslightlyhigh-erorlower,butallshowthesametrend.Thismeansthatextensiveeconomicopportunitiesarebeingcreat-edalongthevaluechainformanufacturers,suppliers,andserviceproviders.

incountriessuchasThailand,India,andSouthAfrica.Atthelowerendoftherankingarecountriesinclud-ingBelgium,Japan,theUS,andFrance(Ipsos,2023).Overall,thereareincreasingsignsthattherearefarmoreopportunitiesthanrisks.Thetargeteduseofarti-ficialintelligencewillsignificantlyaffectourprosperityinthecomingdecades.AIboostsefficiencyandcancounterthenegativeeffectsofskillsshortages,demo-graphicchanges,andhighlocationcosts.Itisnowuptotheautomotiveindustrytotakeboldandappropri-atelyfastaction–andfollowastrategicallyintelligentapproach.

AIasGameChanger|01RevolutionandAutomotiveMarketPotential

13

AIasGameChanger|02.InvestmentinCompaniesWithanAIFocus

02.

Investmentin

CompaniesWithanAIFocus

15

16

Magnetforinvestment:TotalinvestmentinAIcompaniesfoundedsince2001inbillionsofUSdollars

16.5bn.US$GreatBritain

4.6bn.US$WashingtonDC

5.0bn.US$Germany

29.2bn.US$NewYork

6.1

bn.US$

France

16.6bn.US$Boston

★★

★★

★★

★★

★★★

39.6

BnUS$

8.4bn.

US$

Dallas

Diego

5.3bn.US$San

101.2

bnUS$

234.1

10.2bn.US$LosAngeles

BnUS$

55.8bn.US$SanFrancisco

7bn.US$Seattle

41.7

bnUS$SiliconValley

Figure3:TotalinvestmentsinAIcompaniesfoundedsince2001,inUSDbillion(Scheuer,2024)

AlookatthedistributionofAIinvestmentshowsthedominanceofthoseregionsthatalsodominatedthemarketinthelastplatformshifts(seeCoatue,2024;Figure1).Itcanbeassumedthattheautomotivein-dustrywillcontinuetobedependentonhyperscalersandtechnologycompanies.Collaborationsregardingsoftware,cloudapplications,andtheuseofAIareex-pectedtoincrease.

AnanalysisshowsthatalargeshareoftheinvestmentinAIcompaniescomesfromtheUS.Acloserlook(Coatue,2024)showsthatonlyapprox.3percentoftheventurecapitaldealshaveaclearlinktoAI,butthat15percentoftheinvestedcapitalflowsintoAIstart-ups.Fromthisimbalance,itcanbeconcluded

thatthemarketseesrelativelyhighvaluationsandcorrespondinglyhighinvestmentrounds.Thefinanc-ingroundsshowthatmostoftheinvestmentsin2024wentintocompaniesthatdevelopAImodelssuchasChatGPT,Mistral,andClaude.AtotalofUSD14bil-lionwasinvestedinAImodelsinthefirsthalfoftheyear.Thisequatesto62percent.

In2024,asmallerproportionofthecapitalinvestedinAIcompanieswentintofirmsthatdevelopsemicon-ductorsforAIapplications.Roboticsapplications,suchashumanoidrobots,garneredapprox.USD2billionincapital,whichcorrespondstoaround9percentofthetotal.

17

AmongthelargestinvestorsintheAIfieldarethemajortechnologycompaniesincludingMicrosoft,Amazon,NVIDIA,andAlphabet(Google’sholdingcompany).In2023,thesecompaniesinvestedaroundUSD25billionandwerethusresponsiblefor8percentofinvestment.

Carmanufacturers’investmentsincompaniesthatdealwithartificialintelligencearemoremodest.Belowaresomeexamples:

InvestmentsbyNIOCapital

Momenta:Start-upwithafocusonautonomousdriv-ingandonthedevelopmentoftechnologiesforenvi-ronmentalperceptionandhigh-precisionmapping

Pony.ai:Companyfocusingonautonomousdriving;itformspartnershipstodevelopmobilitysolutions

BlackSesameTechnologies:CompanyspecializinginAIchipsandsystems

InvestmentsbyBMWiVentures

Alitheon:SpecializesinopticalAItechnologyforob-jectidentificationandauthenticationwithFeaturePrinttechnology

Recogni:Focusesonhigh-performanceAIprocessingwithlowpowerconsumptionforautonomousvehicles

AutoBrains:DevelopsAIsolutionsfortheautomotiveindustry,particularlyinthefieldofautonomousdriv-ingtechnologies

InvestmentsbyPorsche

Sensigo:DeveloperofanAI-supportedplatformforoptimizingvehiclediagnosticsandrepairprocesses

Waabi:CanadiandeveloperofAI-basedsolutionsforself-drivingtrucks

AppliedIntuition:Providessoftwaresolutionsforthedevelopmentofdriverassistancesystemsandauton-omousdriving

Cresta:Specializesinreal-timeintelligenceforcustom-erinteractionsandcommunicationsolutions

WhereareAIVCdollarsgoing?

Funding~$14B

100

80

60

40

20

0

62%

AIModels

~$4B

20%

AIApps

~$2B

9%

AIOps/AICloud

~$2B

9%

AIRobotics

~$100M

<1%

AISemis

AIasGameChanger|02InvestmentinCompaniesWithanAIFocus

Figure4:InvestmentinAIstacklayers(Coatue,2024)

AIasGameChanger|03.PilotProjectsandImplementation

03.

PilotProjectsandImplementation

19

20

Without

comprehensivepriordigitalization,the

implementationof

AIapplicationswill

beaninsurmountablechallenge.Car

manufacturersandsuppliersshould

allocatebudgetsforAIandbuildup

expertisepromptly.

21

Intheautomotiveindustry,amixedpictureisemergingwithregardtotheacceptanceandimplementationofAIapplicationsalongthevaluechain.Thelevelofim-plementationislowamongsuppliersanddealersandinafter-salesservices.Automobilemanufacturershavemadefurtherprogressintermsofimplementation,butthereissignificantpotentialforimprovementhere.

Lookingattheautomotiveindustryasawhole,only4percentofcompanieshavebeguntoimplementAIapplicationsatselectedlocations.Thatisaroundhalfasmuchasinthepharmaceuticalindustry.Inretail,thefigureisfourtimeshigher.Some28percentofcompaniesintheautomotivevaluechainareworkingonAIpilotprojects,andthevastmajority(68percent)arestillatexplorationstage(CapgeminiResearchIn-stitute,2023).

Only30percentofthecompaniesintheautomotivesectorhaveadedicatedteamandanextrabudgetfortheintroductionandimplementationofAIprojects.Bycomparison,therateis62percentinretail,74percentinthehigh-techsector,and52percentinaerospace/defense.(Capgemini,2023)

Interimconclusion:Theautomotiveindustry’sinvest-mentinAIhasbeenbelowaveragetodate;thisaffectsbudgetsandspecializedteams.GiventhehugeimpactofAIontheindustry,itisadvisabletorectifythissitua-tionquickly.

ProportionofcompanieswithadedicatedteamandbudgetforAI

A

e

g

a

r

62%

ve

52%

36%

30%

40%

74%

CarHighTech

manufacturing

RetailAerospace/

defense

Tele-

communi-

cations

AIasGameChanger|03PilotProjectsandImplementation

Figure5:CompanieswithteamandbudgetforAI(Capgemini,2023)

AIasGameChanger|04.AIModels,Levels,andUseCases

04.

AIModels,Levels,andUseCases

23

24

InterconnectedAIconcepts

Eachconceptisaspecializedpart

oftheoneprecedingit.

Figure6:InterconnectedAIconcepts

AIcoversawidefieldthatcanbedividedintoseveralareasandtermsusingahierarchicaldiagram:

ArtificialIntelligence(AI):Researchareafocusingonthecreationofintelligentmachines.

Machinelearning(ML):BranchofAIfocusingonthedevelopmentofmachinesthatcanlearnfromdata.

Deeplearning:Asub-categoryofmachinelearn-ingbasedonartificialneuralnetworks.Examplesareconvolutionalneuralnetworks(CNNs)andrecurrentneuralnetworks(RNNs).

GenerativeAI:Aspecialtypeofartificialneuralnet-worksthatgeneratedatasimilartothetrainingdata.Examplesaregenerativeadversarialnetworks(GANs)andlargelanguagemodels(LLMs).

WithAIapplications,variouscategoriesofusecasescanbeimplemented:

Datamanagement:Thisinvolvesharmonizingdataandobtainingfindings.Itisessentialfortheefficientuseofinformation.

Patternrecogni