EMERGING

TECH

RESEARCHPitchBookData,Inc.John

Gabbert

Founder,

CEOGenerative

AI

VC

TrendsNizar

Tarhuni

Vice

President,

InstitutionalResearch

and

EditorialSegmentandcategoryanalysisPaul

Condra

Head

of

Emerging

TechnologyResearchPitchBook

is

a

Morningstar

company

providing

the

most

comprehensive,

mostaccurate,

and

hard-to-find

data

for

professionals

doing

business

in

the

private

markets.InstitutionalResearchGroupAnalysisBrendan

BurkeSenior

Analyst,

EmergingTechnologybrendan.burke@Key

takeawaysData?GenerativeAIisnotexemptfromtheVC

recession:

Outside

ofoutliercorporatepartnerships,

VC

investmentisonlyon

pacetomatchthetotalraisedbyGenAIleaders

in2021,with$7.9

billioninconventionalVC

roundsinvestedYTD.Matthew

NacionalesData

Analystpbinstitutionalresearch@?AbarbellisforminginGenAI

startup

dealflowbyproduct

type:

VC

deal

countissupporting

theunderlyingAIcoretechnologiesand

theirultimateverticalapplicationsinsteadofgeneral-purposemiddlewareacrossaudio,language,images,

and

video.PublishingDesigned

by

Megan

WoodardPublished

on

November17,

2023???Closed

deals

feature

highrevenuemultiples:

The

medianlate-stagevaluationhasaccelerated171.7%YoY

to$250.0

million,whileearly-stagevaluationshavegrownas

well.ContentsKeytakeawaysOverview1Among

segments,

code

stands

out

for

M&A

probability:

Companiesinthissegment

featurehigh-qualityinvestors,consistentfundraisingfrequency,andimprovingheadcount.23YTD

fundraising

and

deal

countbysegmentIn

IPO

probability,theAI

core

and

healthcare

segments

havethehighestvalues:

Companiesinthese

segments

stand

out

withtheirscale,

paceoffundraising,and

high-qualityinvestors.YTD

valuationsCorporate

VCactivityYTD

exits589Exit

predictionInvesting

outlook1011We

have

launcheda

pre-seed

reportmethodology

to

more

accurately

andcomprehensively

capture

deals

from

theearliest

phase

of

venture.

Going

forwardwe

will

sunset

“angel”

asa

specified

stageof

venture

in

all

of

PitchBook’s

venture-focused

reports.1PitchBook

Analyst

Note:

Generative

AI

VCTrendsOverviewOurQ3

2023

Generative

AI

Emerging

SpaceMarket

Map

isavailabletoclientsonthePitchBookPlatform.Generative

AI

(GenAI)has

created

a

new

class

oftechnology

leaders

across

productsegments,

yielding

unprecedented

VCdeals.

The

multibillion-dollar

partnershipsbetween

tech

giants

and

startups

are

changing

the

VClandscape

and

presentingpromise

for

continued

startup

outcomes

during

a

tech

recession.

Discourse

aroundGenAI

suggests

that

weare

in

the

next

hype

cycle

and

that

VCsare

eager

tofund

anycompany

building

in

the

space.

To

validate

this,

wehavebeen

tracking

the

space

sincethe

emergence

ofdiffusion

models

in

Q32022,

continually

adding

companies

that

arebuilding

text

and

image

generation

models

toour

GenAI

emerging

space

market

map.This

list

includes

both

new

entrants

and

pre-existing

AI

companies

developing

newtechniques.

We

believe

this

list

has

become

the

most

comprehensive

ofits

kind

in

theVCecosystem.

As

a

result,

it

has

been

leveraged

byleading

corporationstomap

thestartup

landscape.

We

debuted

this

market

map

in

our

Q12023GenerativeAIVerticalSnapshot

and

havesince

added

more

than

double

the

number

ofcompanies

includedtoover

1,500

in

our

GenerativeAIEmergingSpace.

With

this

foundation,

wecan

gaininsight

intothe

initial

tendencies

ofinvestors

in

the

emerging

space

that

will

shapeinvestment

opportunities

going

forward.GenerativeAIproducttaxonomyAI

coreAudioCodeHealthcare?

Data?

Audio

search?

Application

development?

Coding

assistants?

Data

analytics?

Biotech?

Deployment?

Model

architecture?

Orchestration?

Music

generation?

Synthetic

voice?

Healthtech?

Voice

recognition

&transcription?

Documentation?

TestingNatural

language

interfacesVertical

applicationsVisual

media?

Chatbots

&

personal

assistants?

Customer

support?

Neural

search?

Corporate

strategy?

E-commerce?

Edtech?

2D

images?

3D

models?

Content

suite?

Gaming?

Sales

&

marketing?

Translation?

Fintech?

Industrial?

Video?

Writing

&

editing?

Information

security?

Legal?

Proptech2PitchBook

Analyst

Note:

Generative

AI

VCTrendsAtotalof$23.2

billionhasbeen

committedtogenerativeAIstartups

in2023throughOctober

15—alreadya250.2%

increaseover2022’stotal.Outside

ofoutliercorporatepartnerships,

VC

investmentisonlyon

pacetomatchthetotalraisedbyGenAI

leaders

in2021,with$7.9

billioninconventionalVC

roundsinvestedYTD.

Weexcludedealsmade

byMicrosoft

and

AmazonforOpenAI,

Anthropic,and

InflectionAIgiventheiroutsizeddeal

valuesand

commitmentofcloudcomputingresourcesoveramultiyearperiod

tostand

inforcash.

These

commitmentsdo

notresembletraditionalVC

investmentsand

comefromcorporatebalancesheets

insteadofcorporateventurearms.

Seen

thisway,

conventionalVC

investorshavenotexceededtheirpriorpaceofdeployment,and

especiallyso

when

consideringtheinvolvementofcorporateinvestorsinthese

rounds.

GenAI

isnotexemptfromtheoverallVC

recession,thoughthefield

isnotfacingdeclinesseen

inothercategories.GenerativeAIVC

dealvalue($B)$25Inflection-MicrosoftAnthropic-AmazonOpenAI-MicrosoftConventional

VC$20$15$10$5$02020202120222023*Source:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023YTD

fundraisinganddealcountby

segmentAbarbellisforminginGenAI

startup

product

types,

withVC

deal

countsupportingtheunderlyingAIcoretechnologiesand

theirultimatevertical

applicationsinsteadofgeneral-purposemiddlewareacrossaudio,language,and

video.In2022,visualmediabrokeout

indeal

countwiththesuccessofStableDiffusion

and

Midjourneyinimagegeneration.In2023,

leadinginnovatorMidjourneyhasnotrequiredVCfunding,

and

deal

countisnoton

pacetosurpass2022’stotal.AI-coredealshavebrokenout

thisyear,withmorethan100

dealscompletedafter

barelysurpassing80

in2021and

2022.This,

inpart,

owestotheinfluxofnewstartups

formed

toorchestratelargelanguagemodels

(LLMs)and

aconsistentflowofmegadeals

forfoundationmodel

traininglabs.

Verticalapplicationshavegrownthemost

as

newuse

cases

emerge

infields

includingcorporatestrategy,edtech,fintech,and

legal.GenAI-nativestartups

incorporatestrategyand

fintechhaveraised23

roundsin2023,

after

onlyraisingfourin2022,demonstratingthenoveltyofuse

cases

infinancialplanning.3PitchBook

Analyst

Note:

Generative

AI

VCTrendsLagging

categories

demonstrate

challenges

toclosing

rounds

in

smaller

markets.

Dealcounts

in

audio

recognition,

music

generation,

and

video

havefallen

precipitouslyYoY.

While

models

in

these

spaces

haveimproved,

they

havenot

attracted

as

muchattention

as

text

synthesis

and

knowledge

search.

Deal

counts

are

also

not

growingquickly

in

large

categories

such

as

data

preparation,

neural

search,

and

sales

&marketing—areas

in

which

incumbents

havestrong

and

fast-developing

offerings.

Ofthe

33

categories

in

our

taxonomy,only

18

havereached

their

deal

count

from

2022,suggesting

that

investors

are

exercising

discretion

when

faced

with

new

use

cases.AnnualgenerativeAIVC

dealcountby

segment1201008060402002020AIcore2021Naturallanguageinterfaces2022Visualmedia2023*Vertical

applicationsAudioCodeHealthcareSource:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023AI

core

drives

the

bulk

offunding

given

the

megadeals

for

foundation-model

traininglabs

and

database

enablers,

including

Databricks.

Looking

past

that

category,

wesee

that

VC

funding

is

not

breaking

out

across

segments

above

historical

totals.Most

segments

are

not

on

pacetoexceed

the

level

offunding

committedtothespace

in

2021,

with

exceptions

in

code

and

vertical

applications.

The

record-breakingperformance

ofGPT-4on

code

prediction

tasks

has

led

toan

expansion

ofVCinvestment

in

code

tools,

along

with

the

creation

ofnew

categories

in

data

analyticscoding

and

automated

testing.

Coding-assistant

VC

funding

has

increased

457.4%YoY

to

$369.1million.

In

vertical

applications,

fast-growingapplications

includeedtech,

fintech,

legal,

and

industrial,

all

ofwhich

are

on

pacetomore

than

doubleVC

funding

YoY.

In

slower-growingsegments,

such

as

natural

language

interfacesand

healthcare,

2021’s

high

funding

enabled

some

companiestodevelop

new

GenAIproducts

with

existing

runway,lowering

their

need

or

ability

toraise

further

VCfunding

in

2023.

These

totals

suggest

that

large

companies

willbe

created

in

AIcore

and

code.Lagging

categories

demonstrate

that

VCshavenot

supported

all

use

cases

equallyand

haveleft

opportunity

for

new

startups.

We

havetracked

YoY

declines

in

largecategories

including

data

preparation,

customer

support,

voicerecognition

&transcription,

video,

and

multimedia

content

suites.

Some

concern

mayarise

duetothe

rapidly

improving

capabilities

ofcompetitors

and

the

high

valuations

ofcompanies

already

addressing

these

opportunities.

We

believe

that

investors

may

notfully

appreciate

the

data

needs

offine-tuned

custom

models

and

the

market

sizeofcustomer

support,

both

ofwhich

willencourage

larger

deals

in

the

future.4PitchBook

Analyst

Note:

Generative

AI

VCTrendsAnnualgenerativeAIVC

dealvalue($M)

by

segment$2,000$1,500$1,000$500$02020202120222023*NaturallanguageinterfacesVertical

applicationsVisualmediaAudioCodeHealthcareSource:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023YTD

valuationsForthedealsthatarecompleted,valuationshaveincreasedrapidlyin2023.

Themedianlate-stagevaluationhasaccelerated171.7%to$250.0

million,whileearly-stagevaluationshavegrownas

well.Pre-seed/seedvaluationsarenotnecessarilyoutlandish,withamedianvaluationof$14.6

million,whilethehighestearly-stagevaluationshavebeen

inSeries

Broundsforpre-establishedcompanieswithleadingalumnifromresearchlabs,

includingGoogle

Researchand

OpenAI.

The

highmedianlate-stagevaluationowestotheprevalenceofunicorncreation,with12companiesreachinga$1.0

billionvaluation,whilewetrackedonlyfourdownrounds.

Highvaluationsstemfromlargedeal

sizes,withoutstanding

AIcorecompaniesoutsideofOpenAI

and

Anthropicraising$150

millionto$300

milliontopursue

aggressivemodel

trainingand

inferenceroadmaps.

These

valuationsarenotdirectlypredicatedon

revenue,though100x

annualrecurringrevenue(ARR)

multipleshavepresentedaroughupper

bound

inthesample

ofcompanieswehaveheardabout.

Applicationshavebeen

limitedtoARR

multiplesof50x

to75xatthehighend.

Clusteringoccursaround15xto25xARR

formost

applicationcompanies,accordingtoVC

investorKelvinMu

atTransLinkCapital,who

hasexposuretoover50

closeddeals.guidelineissupported

byour

conversationswithinvestors.1This1:“Here

iswhat

I’mseeing

in

terms

ofvaluations,”

LinkedIn,

posted

byKelvinMu,

October

20,

2023.5PitchBook

Analyst

Note:

Generative

AI

VCTrendsGenerativeAImedianVC

pre-moneyvaluation($M)

by

stage$250$200$150$100$50$250.0$92.0$59.4$100.0$14.6$10.02022$0202020212023*Pre-seed/seedEarly-stageVCLate-stageVCSource:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023Early-stagecompanieshavebeen

abletogrowtheirvaluationsbased

on

outsizedexpectations

fromAIinsiders,

whilelate-stagecompanieshaveexperiencedmoremutedgrowth.

The

top10

valuationstep-upsinour

datasetexcludingpre-seed/seed

dealscomefromacrossallsegments,

exceptvisualmedia.

The

topvaluationstep-upcame

fromsynthetic-voicestartup

ElevenLabs,led

byAIheavyweightsAndreesen

Horowitz,NatFriedman,and

DanielGross.

DanielGrossand

NatFriedmanco-createdtheAIGrant

programwhilealsoraisingajointfund

tosupportAIstartups.

Other

specialistinvestorsgrantingoutsizedvaluationsincludeRadicalVentures,RedpointVentures,and

Scale

VenturePartners.

We

believethese

fundshaveuniqueperspectives

on

thefutureofAIthatgivethemconfidenceassigninghighvaluationstoearly-stagecompanieswhilewinningdealswiththemostpromisingstartups.

The

founders

thatreceivethese

high-end

valuationsstand

outfortheirexperienceinAI.6PitchBook

Analyst

Note:

Generative

AI

VCTrendsTop

post-seedgenerativeAIvaluationstep-upsYTD*Valuationstep-up(post

to

pre)Deal

value($M)Post-moneyvaluation

($M)CompanyElevenLabsReka

AIClose

date

(2023)May

1SegmentAudioCategoryLead

investor(s)AndreessenHorowitz,

DanielGross,

Nat

FriedmanSynthetic

voicePlatform$18.5$58.0$100.0$315.010.2x8.0xDST

Global,

RadicalVenturesJune

27AI

coreNatural

languageinterfacesSales

&marketingTypefaceYurtsJune

29$165.0$16.0$18.0$1,000.0$70.0Salesforce

VenturesNava

Ventures6.6x5.4x5.1xSeptember

19October

10AI

coreAI

coreInfrastructureInfrastructureAmplify

Partners,Redpoint

VenturesModal

Labs$156.0VerticalapplicationsRobin

AIRegie.aiJanuary

7Legal$12.0$16.0$39.4$64.0Plural

Platform4.8x4.6xFoundation

Capital,Khosla

Ventures,Scale

VenturePartnersNatural

languageinterfacesSales

&marketingFebruary

9VerticalapplicationsGutter

Capital,Stage

2

CapitalOpusMay

25Edtech$8.8$34.8$75.0$500.04.3x4.0x4.0xModerneFebruary

7September

7CodeTesting$15.0$100.0Intel

CapitalInceptiveNucleicsAndreessenHorowitz,

NVenturesHealthcareBiotechnologySource:

PitchBook

?

Geography:

Global*AsofOctober

15,

20237PitchBook

Analyst

Note:

Generative

AI

VCTrendsCorporate

VC

activityCorporateventurecapital(CVC)

participationhasdrivenvaluationsoverall,withcorporateinvestorsstrikinground-tripdeals

thatpromisecloud

usage

inexchangeforupfrontcommitments.

AIisthemost

computationallyintensiveworkloadthatastartup

can

bringtomarket,makingsupport

forthecommunityaleadingpriorityforcloudgrowth.

As

aresult,CVC-involveddealshaveaccountedfor89.7%ofdealvalueafter

onlyreaching41.2%

in2022,showingthechokeholdthatcorporateinvestorshaveon

themarket.CVCshaveonlyparticipatedin26.7%ofdeals,

downYoY,

focusingon

themarketleaders

thatcan

contributesignificantcloudspending.CVCshaveinvestedatleast

$365.9millionacrossallsegments

exceptforaudioYTD,

leavinglittlegreenfield

forconventionalVC

investors.The

involvementofthese

investorscreateschallengesfortraditionalVCstolead

roundsand

offer

valuewithoutcomputingresources.ShareofgenerativeAIdealactivitywithCVC

participation100%75%50%25%0%89.7%41.2%28.3%26.7%2023*20202021Deal

value2022Deal

countSource:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023Corporationsrankamong

themost

activeVC

investors.Alphabet,Microsoft,

andNvidiahaveallmade

atleast

10

post-seedVC

investments,rankingamong

thetopinvestorsbydeal

countalongwithAndreesen

Horowitzand

SVAngel.Thiscountincludesallaffiliatedventurearms

foreach

corporation.Other

leadingVC

investorshavedefined

investmenttheses

inthespaceand

featureAIinsidersas

partners,includingSequoia

Capital,KhoslaVentures,MadronaVentureGroup,and

CoatueManagement.

Solo

GPs

areessentialparts

ofearly-stageinvesting,includingNatFriedmanand

EladGil,among

others.

The

highparticipationofcorporateinvestorsshowsthedepthoftheircommitmenttostartups

when

VCsarestrugglingtoraisefreshfunds.8PitchBook

Analyst

Note:

Generative

AI

VCTrendsTop

post-seedgenerativeAIinvestorsby

dealcountin2023*InvestorDeal

countEarly-stage

VCLate-stage

VCVenture

growthInvestor

typeVCAndreessen

HorowitzNvidia

(including

CVC)Alphabet

(including

CVC)Microsoft

(Including

CVC)SV

Angel1212111010965576846362255433112201CorporationCorporationCorporationAngel

groupVCSequoia

Capital0000000Khosla

VenturesNat

Friedman841VC7Angel

(individual)VCFirestreak

VenturesElad

Gil630446Angel

(individual)VCMadrona

Venture

GroupCoatue

Management66PE/buyoutSource:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023YTD

exitsVC

exitshavedemonstratedbimodaloutcomes,wheresome

promisinginitialtractioncan

redeem

thevaluationsgrantedbyVCs,yetotherresearchgroupswithouttractionmayexperiencealow-valueacquisition.We

havetrackedthreeacquisitionsover$500.0

millionso

farthisyear,led

byDatabricks

committingtoopen-sourceAIleader

MosaicML

fora$1.3

billionheadlineprice.

The

acquisitionwasjustifiedbytheopen-sourcedownloadsofMosaicML’scompressed

LLMs,alongwithtechnicalprogresson

LLM

inferenceacceleration.We

believethatinfrastructureinnovationwillbringadditionallargeacquisitionsfromtechleaders.Notallsegments

areexperiencinggreatoutcomes,withonlyone

verticalapplicationyieldinganacquisition—legal.Searchand

AIcorehaveseen

thegreatestdemand,

withtechchallengersacquiringgenerativeAIstartups

thatcan

efficientlyparse

customerdata.

Tech

giantsremaincontenttostrikepartnerships

withleadingstartups

and

host

smallerstartups

on

theircloudmarketplaces,thoughtheback-and-forthbetween

cloudhyperscalers

and

startups

mayencouragetakeoveroffersinafavorableregulatoryenvironment.9PitchBook

Analyst

Note:

Generative

AI

VCTrendsTTMgenerativeAIVC

exits*CompanyCasetextClose

date

(2023)SegmentCategoryExit

value

($M)

Acquirer(s)ARR

multipleMOICThomsonAugust

17Vertical

applicationsLegal$650.0Reuters65x9.6xInstaDeepJuly

31July12June

29AI

coreAI

coreAI

corePlatformPlatformPlatform$680.9$1,300.0$285.0BioNTechDatabricksMeituanN/A65xN/A6.3x38.5x1.2xMosaicMLLight

YearBeyondNatural

languageinterfacesChatbots

&assistantsCardinaDiagramNeevaJune

26June

21May

24N/ARampN/AN/AN/AN/AN/A2.4xVisual

media2D

imagesN/AFigmaNatural

languageinterfacesNeural

search$185.0SnowflakeRecursionPharmaceuticalsValence

DiscoveryMay

16HealthcareBiotechnology$35.1N/A3.4xCodigaApril

4CodeCoding

assistants2D

imagesN/AN/ADatadogN/AN/AN/AN/AClipDropMarch

8Visual

mediaStability

AISource:

PitchBook

?

Geography:

Global*AsofOctober

15,

2023Note:

Annual

recurring

revenue

multiples

are

approximate.ExitpredictionThe

abilityofGenAI

toproduceinvestorreturnswillbe

conditionedbythescalecompaniescan

achieveand

thesuitabilityoftheirbusinessmodels

forpublicmarkets.We

havecoveredtheindividualmarketsizesofAIcategoriesinour2023ArtificialIntelligence&

MachineLearningOverview,whichwillbe

adeterminingfactor

intheaddressable

marketsfornewAIstartups.

We

areseeingsome

GenAIinfrastructureand

vertical

applicationstartups

find

product-marketfit,

yetGenAIrolloutsremainintheirinfancyand

enterprise-widespendingremainsfar-off.TheinitialprogressofVC

fundinginthespaceprovidesuswithenough

datatocomparenewstartups

againstpriorwavesofsoftware

startups.

Whilethecurrenteconomydiffers

fromthatduringthelastwaveofAIexits,

webelievethatconsumerinterestcan

support

similaroutcomesas

monetary

policy

improvesinthemediumterm.Among

segments,

code

stands

out

forits

M&A

probability.Companiesinthisspacehaveraisedsignificantfundingyethavenotgrownintoindependentlylargecompanies.

Our

model’spredictionsaredrivenbyfactors

includinghigh-qualityinvestors,regularfundraising,and

improvingheadcount.

We

havebegun

toseeM&A

activity

incodingautomationwithDatadog’sacquisitionofCodigaandAmazonWebServices’

acquisitionofFig.GitHubitself

looms

as

ahighwatermarkforprivatedevelopertoolsafter

its

$7.5billionsale

toMicrosoft.

The

successofGitHubCopilotshouldencourageITcompaniestomakesimilarinroadsintocodingautomation,withMicrosoft

presentingapotentialbuyeras

well.10PitchBook

Analyst

Note:

Generative

AI

VCTrendsInIPO

probability,AIcoreand

healthcarehavethehighestvalues.

Our

model

seesfoundation-modelunicornsas

probabletoachieveIPOs

giventheirscale,

paceoffundraising,and

high-qualityinvestors.The

continuedvaluationgrowthofOpenAIinsecondary

transactions

isnotfactoredintothemodel,

yetitunderscores

thepotentialofapubliclistinginits

future.Biotechisone

ofmost

matureuse

cases

forgenerativeAI,

giventheprogressinmoleculardesign,and

hasalreadysupportedmultiplehigh-valueIPOs

suchas