人工分析人工智能采用調(diào)查-2025年上半年 Artificial Analysis AI Adoption Survey – H1 2025_第1頁
人工分析人工智能采用調(diào)查-2025年上半年 Artificial Analysis AI Adoption Survey – H1 2025_第2頁
人工分析人工智能采用調(diào)查-2025年上半年 Artificial Analysis AI Adoption Survey – H1 2025_第3頁
人工分析人工智能采用調(diào)查-2025年上半年 Artificial Analysis AI Adoption Survey – H1 2025_第4頁
人工分析人工智能采用調(diào)查-2025年上半年 Artificial Analysis AI Adoption Survey – H1 2025_第5頁
已閱讀5頁,還剩45頁未讀 繼續(xù)免費閱讀

下載本文檔

版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進(jìn)行舉報或認(rèn)領(lǐng)

文檔簡介

PUBLICRELEASE

ArtificialAnalysis

AIAdoptionSurvey–H12025

HighlightsReport

FullreportavailablewithAITrendsSubscription

ArtificialAnalysisisaleadingandindependentAIbenchmarkingand

insightsprovider.WesupportengineersandcompaniestounderstandAIcapabilitiesandmakecriticaldecisionsabouttheirAIstrategies.

Ourdata,insightsandpublicationsaregroundedinourcomprehensive

benchmarkingofAItechnologiesandusecases.ThisincludeseverythingfromhourlyperformancetestingofLanguageModelsAPIstomillionsofvotesinourcrowd-sourcedevaluations.

Ourpublicwebsite,

artificialanalysis.ai

,iswidelyreferencedthroughouttheAIindustry.Todiscussthisreport,ourpublicationsorourservices,pleasecontactusat

contact@artificialanalysis.ai

.

ArtificialAnalysisAITrendsSubscription:ComprehensiveAImarketintelligenceforenterprisedecision-makingfromtheleadingAIbenchmarkingcompany

B|AIAdoptionSurvey

Real-worldadoptioninsightsfromthosebuildinganddeployingAI

?Enterpriseusecasepatterns

?Enterpriseadoptionbenchmarks

?Developerprioritiesandpainpoints

?Model,inferenceandhardwareproviderdemandbyindustry

OverviewofArtificialAnalysisAITrendsSubscriptionThisreport

C|TopicUpdates

Timelybriefsonemergingmarketsandbreakthroughproducts

?AnalysisBriefs(data-driven

overviewsofparticularproducts)

?StateofAI:China-ourdetailedbenchmarkingreportonthe

progressofthetopChineseAIplayersacrossAIsegments

D|Databooks&API

Directaccesstotheindustry'smostcomprehensivedata

?ComprehensiveAIperformancedata-sourcedataforallour

analysis

?Intelligence,performance,cost,

preferencevoting,surveydataandmore

?ExceldatabooksandAPIaccess

A|QuarterlyStateofAIReport

ThedefinitivequarterlyupdateonAImarketdevelopments

?TechnicaltrendsateachlayeroftheAIstack:agents,models,

infrastructure,hardware

?Deep-divesintoagenticuse-cases

?Marketmapsandperformance

rankingsforhundredsofkeyplayers

E|QuarterlyAITrendsWorkshop

Connectmarketintelligencetoyourstrategicpriorities

?LivequarterlybriefingwithArtificialAnalysisresearchteam(twohours)

?What'sworkingnow:insightsfromleadingSFstartupsandenterprises

?Deepdivestailoredtoyourbusinesspriorities(e.g.,codingagentbestpractices,inferenceeconomics,upcomingchips)

F|OngoingTeamAccess

Directaccesstoourresearchteam

?AccesstotheArtificialAnalysisteamfortime-sensitivequestionsandclarificationsbetweenworkshops

ArtificialAnalysisis

trustedbytheleadingAIindustryplayersand

publications

9roAMDdiCMeta上?MistralAl

satooT?TechcrunchWSJ

Feelfreetogetintouchwithusat

subscriptions@artificialanalysis.ai

tolearnmoreabouttheArtificialAnalysisAITrendsSubscription

HighlightsVersion(this)

AIadoptionmaturityoforganizationsandpopularinternalusecases

PopularAIChatApplicationsandCodingAssistantsinH12025

PopularLLMfamiliesinH12025,basedoncurrentuseorconsideration

Leadingspeech,image,andvideomodelsandselectioncriteria

Leadingfirst-partyandthird-partyinferenceproviders

PremiumVersion(AITrendsSubscription)

IncludeseverythingintheHighlightsVersionplus:

AIadoptionmaturitybysector,region,andorganizationsize

Expectedusescasesbysector,regionandAIadoptionmaturityPlannedadoptionofAIAgentsbybusinessfunctioninthenext

12months

LLM-familymarketsharebyregion,andsector,includingdemandforreasoningmodels

DemandforChineseLLMsbyregion,sectorandorganizationsize

MarketshareofAIChatApplicationsandCodingAssistants

bysectorandregion

DemandforLLMfine-tuningandpre-trainingincludingkeydriversandchallenges

Inferenceprovidermarketsharebysectorandregion,includingselectioncriteria

Trainingandinferencehardwaremarketsharebysector

ThisistheHighlightsVersionoftheAIAdoptionSurveyReport,thePremiumVersionisavailabletosubscribersofourAITrendsSubscription

Feelfreetogetintouchwithusat

subscriptions@artificialanalysis.ai

tolearnmoreabouttheArtificialAnalysisAITrendsSubscription

4

6keytrendsstandoutinourH12025AIAdoptionSurvey

AIreachesproduction

EngineeringandR&Disthefrontrunnerusecase

~45%oforganizationsreportusingAIin

production

66%areconsideringAIforEngineering/

R&D,CustomerSupportandSales&

Marketingfollow

CompaniesareincreasinglydiversifyingAIuse

OrganizationsaretakingdifferentapproachestoBuildvs.Buy

AveragenumberofLLMsused/

consideredbyorganizationsis~4.7

32%ofrespondentsfavorbuilding;27%buyingand25%ahybridapproach

Google,xAI,DeepSeekgainshare

since2024

~80%areusing/consideringGoogle

Gemini,53%DeepSeek&31%xAIGrok

OrganizationsareopentoChinesemodels,ifhostedoutsideofChina

55%ofrespondentswouldbewillingto

useLLMsfromChina-basedAIlabs,

ifhostedoutsideofChina

TheArtificialAnalysisH12025AIAdoptionSurveycollectedresponsesfrom1,000+usersofAItodevelopmarket-leadinginsightsonthelatesttrendsinAIadoption&use

ArtificialAnalysisAIAdoptionSurveyrespondents

Whereareyoucurrentlybased?Howlargeisyourcurrentemployer?Whatisyourcurrentrole?Whattypeoforganizationdoyoucurrentlyworkat?N=1036

Region

333

292

252

8673

RestofWorldEuropeUSAIndiaChina

Role

450

224

169

87

106

Product&

ProjectMgmt

Other(e.g.

sales)

ExecutiveLeadership

TechnicalLeadership

Engineering&Technicalrole

Sector

352

155

164

152

149

64

Technology

-Large

Technology

-AI/ML

Dedicated

Education

Government/PublicSector

Non-

Technology

Org

Other

OrganizationSize

335352

205

144

Medium(10-200)

Large(200+)

IndividualsSmall

(2-10)

6

Tableof

Contents

AIAdoptionSurvey–H12025

1.AIAdoption&UseCases

AnalysisoforganizationalAIadoptionstagesandplannedusecasesacrossdifferentbusinessfunctions

2.LanguageModels

OverviewofLLMfamilypreferences,Chinesemodeladoptionattitudes,andfine-tuningpractices

3.Speech,Image,&Video

ExaminationofmultimodalAImodelusageacrosscontentgenerationplatformsandapplications

4.InferenceServices

Analysisofproviderpreferences,accessmethods,anddeploymentchallengesincludingcostandperformance

5.Training&Hardware

Reviewofhardwarechoicesfortrainingandinferenceworkloadsacrossdifferentvendorecosystems

7

01

AIAdoption&Use

AIAdoptionSurvey–H12025

01.AIAdoptionandUse

Maturity:AIadoptionismovingpastprototypingtoproduction,with45%ofrespondentsusingAIinproduction;organizationsareoptingtoeitherbuyorbuildAI

AdoptionofAIbyOrganizations

WhereareyoucurrentlyinyourAIjourney?N=1036

DemandforBuildingvsBuyingAI

AreyoubuildingorbuyingAIbasedapplications?N=1036

9

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

01.AIAdoptionandUse

UseCases:Engineering&R&DstandsastheclearfrontrunnerforAIadoption,citedby

66%oforganizations,whileCustomerSupportandSalesarethenextmostciteddomains

ExpectedInternalAIUseCasesOverNext12Months

WheredoyouexpectyouwilluseAIinthenext12months(internalusecases)?N=960(excludesrespondentswhomentionedtheywilluseAIexternallyonly)

Otherusescasesindicatedbyrespondentsinclude

research&analysis,education&training,

administrative,contentcreation

10

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

01.AIAdoptionandUse

KeyChallenges:Intelligence,reliabilityandcostsarethemostcommonchallenges,facedbyatleast50%oforganizationsusingAI

KeyChallengesFacedWhenUsingAI

WhatarethekeychallengesyoufaceinusingAI?N=1006

11

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

01.AIAdoptionandUse

AIChatApplications:ChatGPTleadsasthemostpopularAIchatapplication,closelyfollowedbyGeminiandClaude

ShareofRespondentsusingAIChatApplication

WhichAIchatapplicationsareyouusingorconsideringusingthisyear?N=974

OtherchatapplicationsusedbyrespondentsincludeDeepSeekChat,T3Chat,andQwenChat

12

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

01.AIAdoptionandUse

AICodingTools:GitHubCopilotandCursordominatethemarketasthemostpopularAIcodingtools,withasignificantleadoverClaudeCodeandGeminiCodeAssist

DemandforCodingTools

WhichAItoolsareyouusingorconsideringusingthisyear?N=955

Otherselectcodingtoolsusedbyrespondentsinclude

Cline,RooCodeandChatGPT

13

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

02

LanguageModels

AIAdoptionSurvey–H12025

02.LanguageModels

Models:GoogleGemini&OpenAIGPT/omodelfamiliesareclearleadersconsideredby~80%ofAIdevelopers;DeepSeekisthepreferredopenweightsoption

DemandforLLMFamilies

WhichLLMfamiliesareyouusingorconsideringusing?N=591

AveragenumberofLLMfamiliesusedorconsideredbyrespondents:4.7

15

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

02.LanguageModels

Models:Overthepastyear,OpenAIhasmaintaineditslead,GoogleGeminiandDeepSeekhavesurged,andMetaLlamaandMistralhavefallen

DemandforTop10LLMFamiliesinMay2025

WhichLLMfamiliesareyouusingorconsideringusing?N=270(2024)and591(2025)

20252024

Changebetween2024&2025,p.p.

+1%+49%

84%83%

AveragenumberofLLMFamiliesusedorconsidered

80%

byrespondentsincreasedfrom~2.8to~4.7between

+21%

2024and2025,indicatingmarketmaturationand

increasedexperimentationasorganizationsexplore

67%

variousAIrequirements

+53%

-6%

53%

49%

46%

43%

-15%

+31%

37%

31%

+17%

31%+25%

25%

+14%

21%

22%

14%

4%

0%DeepSeek

0%

PerplexitySonarMicrosoftPhi

0%0%

GoogleGemini

OpenAI(GPT/o)

AnthropicClaude

MetaLlama

xAIGrokAlibabaQwenMistral

16

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

02.LanguageModels

ChineseModels:MostorganizationsarewillingtouseAImodelsfromChinaiftheyarehostedoninfrastructureoutsideofChina

WillingnesstouseChineseLLMs

WouldyoubewillingtouseLLMsfromChineseAIlabs?N=574

WillingnesstouseChineseLLMs,byRegion

WouldyoubewillingtouseLLMsfromChineseAIlabs?N=574

17

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

03

Speech,Image,andVideo

AIAdoptionSurvey–H12025

03.Speech,Image,andVideo

Speech:OpenAIandElevenLabshavethemostpopularspeechgenerationmodelswith

largegaptopeers;streamingquality,naturalvoice,&latencyarethemostimportantfactors

DemandforSpeechGenerationModelsKeyFactorsforSelectingSpeechGenerationModels

Whichmodelsareyouusingforspeechgeneration?N=279HowimportantarethefollowingcriteriawhenselectingSpeechgenerationmodels1?N=269

1.RespondentswereaskedtochooseifeachcriteriawasCriticallyImportant,ConsideredbutnotapriorityorNotconsidered.Eachofthethreeoptionsareassigned100,50and0pointsrespectively

19

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

03.Speech,Image,andVideo

Image:OpenAIistheclearleaderforimagegenerationmodels;usersofimagegenerationvaluepromptadherenceasthe#1mostimportantfactor

DemandforImageGenerationModelsKeyFactorsforSelectingImageGenerationModels

Whichmodelsareyouusingforimagegeneration?N=338Howimportantarethefollowingcriteriawhenselectingimagegenerationmodels1?N=312

Top3usecasesforimagegenerationmodelsinclude:

?Contentcreation(e.g.thumbnails,illustrations)

?Marketing&advertising(e.g.socialmediacreation)

?Entertainmentandgaming(e.g.gameassets,imagery)

1.RespondentswereaskedtochooseifeachcriteriawasCriticallyImportant,ConsideredbutnotapriorityorNotconsidered.Eachofthethreeoptionsareassigned100,50and0pointsrespectively

20

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

03.Speech,Image,andVideo

Video:OpenAIandGoogleleadvideogenerationbyalargemargin,withusershighlightingpromptadherenceandrealismasthemostimportantfactorsforselectingmodels

DemandforVideoGenerationModels

Whichmodelsareyouusingforvideogeneration?N=161

KeyFactorsforSelectingVideoGenerationModels

Howimportantarethefollowingcriteriawhenselectingvideogenerationmodels1?N=161

Top3usecasesforvideogenerationmodelsinclude:

?Documentaryandentertainment(e.g.recreationalvideos)

?Marketing&socialmedia(e.g.promotions,marketingassets)

?Prototypingandtesting(e.g.proof-of-concepts,productdemos)

1.RespondentswereaskedtochooseifeachcriteriawasCriticallyImportant,ConsideredbutnotapriorityorNotconsidered.Eachofthethreeoptionsareassigned100,50and0pointsrespectively

21

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

04

InferenceServices

AIAdoptionSurvey–H12025

04.InferenceServices

Inference:First-partyAPIsfromproprietarymodellabsandhyperscalersleadininferenceserviceswithalongtailofthird-partyinferenceproviders

DemandforInferenceProviders

Whichinferenceprovidersareyouusingorconsideringtoaccessmodels?N=480

23

Source:ArtificialAnalysisAIAdoptionSurvey–H12025

04.InferenceServices

Inference:Providerswithfrontierfirst-partymodelAPIs(e.g.OpenAI,Google&Anthropic)andchipchallengers(e.g.Groq&Cerebras)gainedshare;AmazonandAzure'ssharefell

DemandforTop10InferenceServicesinMay2025

Whichinferenceprovidersareyouusingorconsideringusingtoaccessmodels?,N=256(2024)andN=480(2025)

20252024

Changebetween2024&2025,p.p.

+12%

80%

+30%

68%

68%

+22%

-1%

50%

-12%

+19%

42%43%

40%

38%

36%

+10%

28%

28%

26%

+6%

+5%

17%

16%

17%

14%13%

11%

9%

NA1

OpenAIGoogleAnthropicAzureGroq

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
  • 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負(fù)責(zé)。
  • 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

評論

0/150

提交評論