BUILDINGABUSINESSINTHEAGEOFAI

Whatventurecapitalistsarelookingforfromcutting-edgeAIfounders

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAI

TABLEOFCONTENTS

TableofContents|2

TheFutureStartsRightHere 3

AIInvestmentin2025:Bubble,RisingTideorBoth? 5

ARRvs.ERR:WeighingExperimentalIncomeforAIStartups 8

PricingSoftwareintheAgeofAI 10

ProductRoadmap:SeparatingSignalfromNoise 12

ThePullofDataGravity 14

TheKPIsThatMatter 15

GTMandPartnershipsforAI-NativeStartups 17

HowAIFoundersNailtheVCMeeting 19

EarlyDays,DistantHorizons 21

NextSteps 22

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAITheFutureStartsRightHere|3

THEFUTURESTARTSRIGHTHERE

EIGHTPERSPECTIVESONAISTARTUPS

AKASHBAJWA

Principal,EarlybirdVentureCapital

ROHINICHAKRAVARTHY

ManagingPartner,NewBuildVentureCapital

PATRICKCHASE

Partner,RedpointVentures

SAKIBDADI

Partner,Stage2Capital

MATSOLOVMAGDESJ?

OperatingPartnerforAnalytics,EQTVentures

LIAMMULCAHY

Go-to-Market,KleinerPerkins

SHRAVANNARAYEN

Partner,IVP

SAMTEDEN

Investor,AnthosCapital

We’retold,repeatedlyandfairlyconvincingly,thatthevanguardofadvanced

AItechnologybeingledbylargelanguagemodelsandgenerativeAIisgoingtofundamentallychangetheworld.Theworkbeingdoneinlabsandofficeparks,thefundingdecisionsbeingmadebyventurecapitalfirms,willcumulatively

catapulttheworldintoanewera.

ButifAIisgoingtochangeeverything,surelyitwillstartwiththatveryprocess,thewaythebusinessesbringingAItomarketarebornandnurtured.Itseemed

likeasmarttheory,anditmatchedwhatwewereseeingfromthenewbusinessesworkingwiththePoweredbySnowflakeStartupProgram.Sowedecidedtodo

alittledigging.

TounderstandwhattheageofAIhasdonetotheperpetualpursuitoffunding,weinterviewedagroupofventurecapitalistswhopartnerwithSnowflake’s

startupprogram.EightexperiencedfundersintheUnitedStatesandEuropeexplainedhowtheirapproachhaschangedwhenitcomestoconsideringandfundinganAI-basedstartup—andhowmuchithasn’t.

Thefirstthingwelearnedisthatit’skindacrazyoutthere.Everytimeyoublink,theperformanceandcostofartificialintelligencehastakenasignificantleap

forward.PatrickChase,apartneratRedpointVentures,sayshesawthatwhenheandhiswifehadtheirfirstchildinmid-2023.

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAITheFutureStartsRightHere|4

“Iwasonpaternityleaveforabouteightweeks,andwhenIcameback,thingswerecompletelydifferent.Llamahadshipped7B,13Band70Bmodels,OpenAIhadshippednewmodelsandsodidAnthropic,”

hesays.“It’sjustcrazyhowfastthingsaremoving.”

Notonlydothespecificsofthetechnologychange

onanall-but-dailybasis,buttheplayingfieldhas

“OneofthebigdifferenceswithAIisthatmore

executiveshaveownershipoveritandbenefitfromit,”saysIVPpartnerShravanNarayen.“Inthepast,abusinesscouldtheoreticallyisolateitsdataand

analyticsstrategyunderafewpeople.ButyourAI

POWERED-UPSTARTUPS

The

PoweredbySnowflakeStartupProgram

providesstartupstheresourcestheyneedtobuildandgrowdata-intensiveapplications

andproducts.

Hundredsofstartupsarealreadybuilding

applicationswithSnowflakeforawidevarietyofusecases,includingmarketingautomation,monitoringandobservability,IoT,machine

learning,embeddedanalyticsandmanymore.

strategyencompassesfinance,product,engineering,IT,sales—everysinglepersonaffectsit,isaffectedbyit,andhasastakeinit.”

effectivelybeenleveledbyasheertidalwaveofnew.

“ThereareallthesepeoplewhohavebeendoingAIforalongtime,butintermsofthestateoftheart

rightnow,thetransformermodels,it’sonlyexistedsince2017,”Chasesays.“Soinsomeways,themostexperiencesomeonecanhaveisacoupleofyears.”

It’snotjustthetechnologythat’sinflux.HowanAIstartupsellsitsproduct,andwhoitsellsto,ischangingaswell.

“AIisanincredibletrendtoinvestbehind,”Chasesays.“Theworldwillprobablybereallydifferentinone,three,five,10yearsfromnow,andIthinkthatcreatesalotofopportunityforstartupsandventurecapital.”

Theinsightsandopinionssharedacrosseight

in-depthconversationslargelybreakdowninto

eightcategories,rangingfromwhetherexperimentalrecurringrevenueisactuallyathingtowhat,

specifically,excitestheseinvestorswhenfirstmeetingafounder.

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIAIInvestmentin2025:Bubble,RisingTideorBoth?|5

AIINVESTMENTIN2025:

BUBBLE,RISINGTIDEORBOTH?

AIisturningupeverywherethatsomethingcanbesoldtoanenterpriseclient.HalfthesandwichshopsinSiliconValleyareprobablyspellingthenameoftheirfancymayonnaiseas“AI-oli.”Thereisbroadpublicexcitementandanticipation,ofcourse,aroundnewgenerativeAIproductsthatprovidedirectinteractionwithlargelanguage

models.Buttheinvestorswetalkedtoarealsolookingat,andfor,AIcapabilitiesbeingappliedtoexistingproductsorbusinessworkflows—less“AI”andmore“AIunderthehood.”ButinadditiontothequestionofwhetherStartupXisagoodinvestment,theyalsohavetoweighalotoftalkaboutbubblesandcrashes.

2024was,inpart,theyearofAIhandwringing.

GoldmanSachs

and

Sequoia

weretwoprominentvoicesintheinvestmentworldscreaming,“Where’stheROI?”and“Where’sallthismoneygoing?”O(jiān)urinvestorsagreethatthestakesarehigh.Butventurecapitalisalwaysahighwireact,balancingthewinninginvestmentsagainst

theinevitablelosers.WithAI,in2025,thewireishigher,butsoisthepayoff.Whichiswhythey’rematchingstrongfaithinthepotentialofAIwithprudencearoundwheretheymaketheirbiggestbets.

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIAIInvestmentin2025:Bubble,RisingTideorBoth?|6

LIAMMULCAHY:“Therehasdefinitelybeenafrenzyof

investmentinAIcompanies.Becausewe’restillintheearlydaysofAI,theopportunitytoapplyitwithincompaniesisseeminglyendless.Wherewelookforinitialsignalisinuseradoption,

satisfactionandfrequency.It’sprettyunbelievablewhatthese

companiesarebuildingandallowinguserstodo,andhowquicklythoseuserscouldnotthinkaboutdoingtheirjobwithoutthem.ROIbecomesadownstreameffectofadoption.

“Thisisafoundationaltechnologyshiftthatwillgenerate

significanteconomicgrowthonaglobalscale.Whenitcomestoearlystageinvestingandcompanybuilding,theopportunityisgreater,sothestakesarehigher.Thoughitdoesn’t

fundamentallychangehowweapproachinvesting.Attheendoftheday,we’reinvestinginpeople.”

SAKIBDADI:“We’reinabubble.It’snotevenaquestion—it’sabubble.Thequestionis,whathappenstodeflateitandgetustoreallongtermvalueatreasonableprices?I’mnotsayingthat

therearen’tcompaniesbuildinggreatproductsthatarevaluabletotheircustomers,butit’sprobablynotthesmartestthingtobepayingridiculousmultiplesforcompanieswherechurnisjustthismassivestructuralissuethatnooneseemstowanttotalkabout.

“I’mgenerallycautiousaboutinvestinginthespaceofAI

startups.Hopefullysomesanityresumeswithintheinvestorcommunitytopreventusfromgoingalittletoofar.Maybethat’salreadyhappening.”

MATSOLOVMAGDESJ?:“There’sstillalotofhype.ButI

thinkwearemoresoberwhenitcomestoitslimitations.We’vealltriedAIproductsthatdon’tliveuptotheirpromises.Soweareinthatawakeningphase.Andthat’sgreat,becausethat’s

whenyoustartaddressingtheproblems.We’retalkingabout

AIinamorenuancedway;wemightdistinguishbetweenwhenyoushouldapplygenerativeAIandwhenyouneedML.AndIseesomeinterestinbuildingplatformstomakethishappen.”

AIInvestmentin2025:Bubble,RisingTideorBoth?|7

AIWILLBEABROADERINNOVATIONFORCE

ROHINICHAKRAVARTHY

ManagingPartner,NewBuildVentureCapital

“Tomymind,atechnologyinnovationrailhasthreecomponents:First,itpowersaneweconomiccost

point:capabilitiesbecomecheaperthaneverbefore,

democratizingaccesstohightechnology.Second,there’softenanewbusinessmodelattached,likewesawwith

theco-developmentofSaaSandcloud.Thethirdpiece

istheanthropologyofit:Whatisthecommunitylike,

areothercustomersbuildingontopofthistechnology?

Manydevelopersacrosscompaniesusingandtrustinga

commonplatformmakesitmorevaluabletoeachofthem.”

Whenallthreeofthosepiecescometogether,she

says,youhaveaninnovationrail.ItisclearthatAIwillbeanewrailasthereisaneweconomicargument

andfantasticcommunitysupport.Andweareseeing

experimentationonnewbusinessmodelsforAIfrom

servicestoagenticpricing.Withthisfoundationinplace,entrepreneursandenterprisescanfindcountlessuse

casesacrossindustriesandfunctions.

“GenAIfoundationalmodelsletyoudothingsthat

youcouldn’thavedonebefore,andthereisalotof

communityinterest—thesmartestpeopleintheworldareworkingonthis,”shesays.“Certainlytherewillbepocketswhereit’soverhyped,butourbetisthatoutofallofthis,we’regoingtogetanewrail.”

RohiniChakravarthy

ismanagingpartneratNewBuildVentureCapital.Across20yearsofinvestingexperience,shehasshepherdedcompaniesthroughexponential

growth,successfulacquisitionsandIPOs.Shecurrentlyinvestsinearly-stagecompaniesattheintersection

ofAIandsupplychains.HertakeontheAIboomis

thattoday’sangstwillbeoverpoweredbythedeeper,structuralpotentialofastill-nascenttechnology.

“I’mlong-termoptimistic,”shesays.“Intheshortterm,thehypecyclemightoverestimatetheimmediate

impact.Thatcanleadustounderestimateamajornewdevelopmentinthelongterm.”

Shepointstothecatalyticpowerofmobiledevices,whosetrueimpactwasn’tseenuntilyearsof

infrastructureworkhadbeenaccomplished.

“AIisdevelopingintoaninnovationrail,whichiswhen

newtechnologiesandtechniquesenableinnovationbycustomersandecosystempartners,”sheexplains.“CloudwasaninnovationrailthatmadeSaaSandopensourcecompaniesviable,ultimatelycreatingtrillionsofdollars

inmarketvalue.

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIARRvs.ERR:WeighingExperimentalIncomeforAIStartups|8

ARRVS.ERR:WEIGHINGEXPERIMENTALINCOMEFORAISTARTUPS

Annualrecurringrevenueisaneasy-enoughconcepttounderstand.SaaScompanies,oranysubscription-basedsoftwareprovider,dependontheongoingflowofcapital,andtheARRmetricprovidesdeepinsightintothefinancialhealthandgrowthpotentialofthebusiness.It’sinvaluableforstrategicplanningand,forinvestors,understandingthevalueofthecompany.

ButAI-basedproductsaresonew,withsomanygrandbutunprovenpromises,thatinitialadopterscan’tbecountedonaslong-termcustomers.Newtechnologieswithnewcapabilitiesbringincompanies—

orindividualteamsofusers—whoarewillingtodropafewbuckstokickthetires.Thatmeansthatchurn,akeythreattotheARRmetric,willbeexceptionallyhigh.

Inablogpostlastspring,AltimeterCapitalpartnerJaminBall

proposedanewmetric

:experimentalrecurringrevenue,orERR.UnderstandinghowERRisdifferent,andparticularlyprevalentontheAIfrontier,couldhelpfoundersandinvestorsavoidmisinterpretingflashesofexperimentalspendingasdurable,long-termadoption.

Madesensetous,butwewantedtoknowifBall’sthinkinghaspenetratedtheindustry.Soweasked.

Thetakeawaywasthatit’simportanttounderstandthedifferencebetweenERRandARR,butthatARRremainsthegoldstandard.

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIARRvs.ERR:WeighingExperimentalIncomeforAIStartups|9

LIAMMULCAHY:“AIisbrandnew.Alot

ofpeoplearetryingtofigureouthowtheycanuseit.SoERRisunderstandable.Butit’sthestorybehindERRthatmatterstome;

howcompaniesareapproachingaproblemspace.WhyareyoudoingsomethingwithAInow?Haveyoutriedbefore?Isthisneworrepurposedbudget?Andultimately,howquicklydoesthatconverttoARR?It’stheconversionpaththat’smoreimportant.”

SHRAVANNARAYEN:“We’relookingat

howtheAIfeaturesareactuallybeingused,andbywhom.Ontheapplicationside,isthisagreatproductthatwilldeliverROIfromatimesavingsorefficiencystandpoint?And

aretherethingsthatcustomersraveaboutindependentofAI-enabledfeatures?”

SamTeden:“Beinggrowth-stageinvestors,wethinkERRisOK,aslongasit’salongsidethe

traditionalproductARR.Broadlyspeaking,AIforusmeansit’saccompanyingaproductinacategorythathasalreadyseenorganicsuccess,

andAIisafurtherconduittoadoptionand

successofthatbusinessovertime.SoifERRisbeingaddedon,we’renotgivingitanywhereclosetothesamemultiplethatwewouldgiveatraditionalARRlineitem.”

AKASHBAJWA:“Becauseexperimental

ERRdoesn’tspeaktothedurabilityofthat

revenue,youneedtotakemoreofacategorylens.Isthestartupcapableofseizingthat

ERRperiodasameanstothengraduatingontoacorebudgetthat’smoredurable,orwillcustomersultimatelyoptfortheirincumbentplatforms?”

PATRICKCHASE:“AlotofAItouristswill

tryeverything.Soyouseealotoffast-

growingAIstartups,whereaconsumerAI

startupracestoamilliondollarsinrevenueinacouplemonths,butwithlowuserretention.We’realwayslookingfordurableenterprise

apps,sousage-basedonesthatarepay-as-you-gowithgoodnetdollarretentionarethemoreexcitingopportunities.”

CHURN,CHURN,CHURN

SAKIBDADI

Partner,Stage2Capital

SakibDadi

isapartneratStage2Capitalandpreviouslyspentnearlyadecade

withBessemerVenturePartners.Mostofhisfocusthesedaysisoninfrastructuresoftwarecompanies.TheERR/ARRquestionisfascinating,hesays,becausethe

highturnoverofexperimentalrevenueisabigchallenge.

“Everyoneistryingtounderstandthechurndynamics,”hesays.“AndthisisthebigquestionaroundalltheseAIcompanies.”

Henotesthatnetrevenueretentionratesareexpectedtobemuchhigherinenterprisesoftware,whichiswhythemarketvaluesARRsohighly.

“Ihaven’tseenanythingtosuggestthatchurnwillmagicallydisappearinacoupleofyearswhentheseAItoolsbecomestandard,”headds.“Itfeelslikethetrend

maybethatfortheseAIcompanies,thequalityofrevenue,intermsofstickiness,isjustnotashighasthequalityofrevenuefortraditionalenterprisesoftware.”

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIPricingSoftwareintheAgeofAI|10

PRICINGSOFTWARE

INTHEAGEOFAI

Costofgoodssoldisafundamentalbusinessmetric,andit’savitalconceptforsoftwaresales.Becausethe“goods”arenotaphysicalproductassembledinafactoryandcarriedoutofashop,itcanbehardertomeasure. Softwarepricingmodels,whichmostlychargeeitherbythenumberofusersorbytheamountofactualusage, havetoaccountforactualcosts.AImakesthathard,becauseAIiscompute-intense.Alotofcomputational

resourcesgointotrainingamodel.Andinference—theprocessofusingatrainedmodeltoanalyzeitsdatatogenerateoutputs—isalsointenseandexpensive.

Wewonderedwhethertheseissuesaregivinginvestorspause.Dotheyrequirecomplexformulasandhoursofspreadsheetnumber-crunching?Turnsout,notreally.Ingeneral,theinvestorswetalkedtohaveaneyeontheissuebutarewillingtoletitshakeoutovertime.AIasafieldisstillgrowingup.

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIPricingSoftwareintheAgeofAI|11

AKASHBAJWA:“We’reinaverynascent

phasewhereseat-basedpricingisstillmost

common.We’veseenearlyeffortsatoutcome-basedpricing,buttherealityisthatforalot

ofLLMpromisedoutcomes,reliabilityisstillaquestionmark.It’sabitofaleap,givenwhereweareontheadoptioncurve,totransition

soquicklytotheoutcome-basedmodel.

“Ithinkthere’llbeaninterimjourneyhere

wherewe’llstillbeonthehybridmodel

betweenseat-basedandconsumption-basedpricing,andthenoverthenextfiveto10

years,outcome-basedpricingmayendupbecomingmorepalatableandwidespread.”

SAKIBDADI:“Usage-basedpricingwill

becomefarmoreprevalentinthemarket.

Ittakeswhatsoftwaredoes,whichisto

enhancework,tothelogicalconclusion,

whichischargingforworkdone.Thatmay

betoexplicitlychargeyouforminutesspentinteractingwithanagent,orabundled

platformfeethatincludesausagecomponent.”

SHRAVANNARAYEN:“Wedon’tover-indexonCOGSrightnow.Butwealwayswantto

payattentiontothenarrativearoundhow

itbecomesasteadystate,propermargin

businessoverthecourseoffiveorsixyears.It’snotaone-yearjourney.”

MATSOLOVMAGDESJ?:“It’sarenaissanceforthegranularP&L.Manyofthecompaniesthatdidn’tusedtocareaboutgrossmargins

aregoingtohavetostartcaringaboutgrossmargins,andmonitoringCOGSisobviouslyextremelyimportantwhenyou’reanAI

companyusingalotofcompute.Soit’sgoodoldfinancialanalytics.AndaGPU-intensivebusinessreallyneedstofindamodelthat

works,becauseaplainsubscriptionmodelmightcostyoualot.”

DISCOUNTINGTHEFUTURE

PATRICKCHASE

Partner,RedpointVentures

PatrickChase

isapartneratRedpointVentures.Hestartedhiscareerasan

engineeratLinkedInbeforeearningbothanMBAandanAI-focusedmaster’sincomputersciencefromStanford.HesaysthatwhilethecostsaroundAIcanbehigh,he’ssurethey’llcomedown.Infact,theyalreadyare.

“There’salotoftalkaboutAIbeingexpensive,butwedon’ttalkenoughabout howfastthecostisdropping,”hesays.“Costforinferenceandtrainingarebothdown10xyear-over-year.”

Hesayshe’slookingforstartupsthatarebettingonthingsgettingcheaperandsmarter,because,likesomuchintheAIspace,thosecostsarechangingquickly.

Andheseespricingmovingtowardusage-basedmeasures.

“Thetrendwe’reseeingattheapplayeristochargeforworkbeingdone,ratherthanforseats,”hesays.“Youseeincustomersupport,chargingforworktobedone,suchasresolvedtickets.Infiveyears,someapp-layerSaaScompanieswillbechargingforseats,butmanywillhavemovedtoamodelalignedwiththeir

costsandthevaluetotheuser.”

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIProductRoadmap:SeparatingSignalfromNoise|12

PRODUCTROADMAP:

SEPARATINGSIGNALFROMNOISE

Therecontinuestobefocusontheinfrastructurelayerrightnow,buttheinvestorswespokewithare

turningtheirattentiontotheapplicationlayer,wheretheyseetremendouspotential.Atthesametime,theyseetremendoushype,soidentifyingrealvalueandpotentialamidallthechaosisnosmallproblem.

Inevaluatingstartups,theseinvestorslookcloselyatwhetherthey’reseeingaproductthatusesAIoranAIproduct—thelatterbeingmoreopentodisintermediationandcommoditization.Yetoverall,theyagaintellusthattheyfocusonthesamefundamentalsthey’vealwaysprioritizedwithSaaSandinfrastructureventures.

interestingtous.Opportunityfordisintermediationhappensalotquickerinthosescenarios.”

MATSOLOVMAGDESJ?:“Ifit’sasemimaturecompany,we

lookathowcarefullythey’vethoughtthroughtheirgo-to-market,becausethattellsyoualotaboutifit’sjustnoisetryingtobuildAIforAI’ssake,orifthey’reactuallythinkingaboutaproduct.”

SAMTEDEN:“Weask,isthisAIonly,orisitaproductwithanopportunityforAIintegrationorAIusecasesthatwillcreate

efficiencyaroundthatproduct’sadoption?WhenIseesomethingthatseemstobebuiltjustfortheAI’ssake,andisleveraging

publiclyavailabledatawithoutanythinkingabouthowtocreateaproprietarymoatovertime,itbecomesdramaticallyless

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIProductRoadmap:SeparatingSignalfromNoise|13

PATRICKCHASE:“Ourbiggestfocuslatelyhasbeenintheendapplications.Whatwelookforissimilartowhatwelookedfor

withtraditionalSaaSapps:Alargemarketandtheopportunity

tobuildastrongcompetitivemoatovertime.IntheAIworld,wethinktechnicalmoatswillcomefromworkflowcomplexity;appswherethesolutionis80%workflowand20%model,ratherthantheopposite.ComplicatedenterpriseworkflowslikeERP,CRM,customersupport,procurementfitinthisbucket—thingswherethere’sareallybigopportunitytousethebestandbrightest

modelsandbuildalotofvalueontopofthem.”

AKASHBAJWA:“Wewanttohearwhythisisatechnology

that’suniquelyplacedtosolvethisproblem.There’ssomuchyoucandonowwithcompoundsystemsusingretrieval

augmentedgenerationwithyourowndata,fine-tuning,betterfunctioncalling,etc.Butaretheresultsofthatcompound

systemjustincrementallybetterthanthestatusquoorastepchangeimprovementthatwillactuallydriveanupgradecycle

ofbuyers?Forexample,withpayroll,youhavetoprocessall

thepayrollpayouts.That’snotsomethingthatyou’lldomoreofbecauseAIspeedsyouup.Butifyoucandomoresales,youwilldomoresales.Ifyoucanwritemoresoftware,you’llwritemoresoftware.”

AIAPPROACHES:OFF-THE-SHELFVS.ARTISANAL

LIAMMULCAHY

Go-to-Market,KleinerPerkins

LiamMulcahy

isonthego-to-marketteamatKleinerPerkins,whereheadvisesstartupsonhowto

commercializetheirinnovations,withafocusonB2Bsales.Beforethat,atUnusualVentures,heworked

withearlystagefoundersgoingfrom0-1.

Attheinfrastructurelayer,thekeydifferentiation

isdeepinthearchitecture.“Youcan’tjustbuild

acompanyasawrapperontopofafoundational

model,”hesays.“Therearesomecoolcompanies

thataresecuringLLMs,providingobservabilityforanAI/LLMworld—thesearen’tnewconcepts,buttheyrequiredeeparchitecturalchange.”

BetweenthesetwoextremesarealotofcompaniestryingtoinfuseAIintofamiliarSaaSworkflows.

“There’salotofnoiseinthatspace,alotofcompetition,”hesays.“Sometimestheresultisjustaraceto

thebottom.”

Hesaysthatthebiggestopportunitieshe’sseeingattheapplicationlayerforAIstartupsisinindustriesthattheSaaSrevolutionunderserved,orflat-outignored.

“That’smainlyserviceindustrieslikehealthcare,

constructionandlegal,”hesays.“Theservicesectoristhelargestpartoftheeconomy,andwhatwe’reseeingnowistechnologythatdoesn’tjustmake

workflowseasier.Itcanperformtheserviceitself.”

Suchstartupsareamongthemostexcitingplacestoinvest,hesays.BecausetheseindustrieswerepassedoverbytheSaaSrevolution,there’salotofpentupdemandfornewsolutions.

“Theothercoolthing,it’snotjustsoftwaretalking

tosoftware,”hesays.“Thesesolutionsaregoingto

changehowbuildingsgetbuilt,howlawyersworkwiththeirclients,howpeopleinteractwiththeirdoctors.

It’sexcitingtoseeAIjumpintotheworldinways

beyondconsumerapplicationsandsocialnetworks.”

STARTUP2025:BUILDINGABUSINESSINTHEAGEOFAIThePullofDataGravity|14

THEPULLOF

DATAGRAVITY

Theefficiency,securityandgovernanceadvantagesofbuildingapps