Cheetah Mobile (CMCM) 13 Sep 242024 Q2 Earnings call transcript

Good day, and welcome to the Cheetah Mobile Second Quarter 2024 Earnings Conference Call. [Operator Instructions] Please note this event is being recorded. I would now like to turn the conference over to Helen Zhu. Please go ahead.

Thank you, operator. Second Earnings Mobile's Call. Cheetah to Quarter Conference XXXX Welcome and our With are Chairman CEO, us today company's Mr. and Mr. Sheng; Fu our Director and CFO, Ren. Thomas

management's script will note we an prepared Q&A Following Please section. AI conduct agent. that be will the CEO's presented by the remarks,

release, I today, earnings agent as this make statements. also we we in begin, safe time, statements our behalf of Chairman, on and At now speak AI you will to refer Before applies let we the to to harbor CEO our which call the Sheng. will conference forward-looking our Fu

you robotics QX, about driven sales of with AI accelerated Thank This of us joining of others by the and for revenue. key today. and primarily XX% OrionStar our total XX.X%. Mobile, of was growth for year-over-year acquisition In our showing revenue our international for enterprise-facing becoming accounted a total both Beijing a contributing in solid growth wheel growth. domestic robotics markets, pillar an service service progress our revenue company. The to Cheetah has made in

service robot three continues position gain areas. to our are market service from cases and and the among exhibition in used and corporate Moreover, hold dominant in Our use widely other centers, competitors. delivery voice-based in museums, a restaurants top ranks robots receptions share to

to better easily ability far thanks our to Customers previously. Customers voice enables recognition, versus best of apps services. have tailor our respond end-to-end product optimization. that including strong strong to support our make within AI and choose smartly LLM, capabilities, field requests because Strong our us experience customers' can software and robots, core their aftersales benefited inquiries robots understand

are robotics two growth wheel service focusing strategies. the of expand business, revenue further on To key our we

enabling innovations, to aim and on a focus voice deliveries. product core with robots Firstly, of continuous interaction use competencies, service capabilities including autonomous broaden our cases through our we the

relatively the delivery market. designed a robots received focus low Southern overhead performance Korea started introduced with and in the and reliable customers instance, most and centers, shipping on For payloads. in customers recently and from already our and autonomous pricing in overseas, Japan we orders providing offer Europe. fulfillment robots for China have for for to have South robots the factories We both We in

In hotel service progress. from industry, Hotel are gaining a robots, making we for our use is wheeled the case existing are robots players. and share market proven service

We are competitiveness. for our updating also further our hotel robots industry increase the to

inquiries. introduction approaching sale by In supermarkets, comprehensive providing ticket price product our service them, the facilitated robots successfully and low buyers, potential responding of clearly to have proactively products identifying consumer

Following our collaboration stores into supermarkets. with [indiscernible] local we expanding are to sell supermarkets, and hot sausages more in now

expanding overseas our robotics revenue business surpassing domestic we service Secondly, already revenue. globally are with

in Southeast America South presence Australia, cases centers. actively in North Korea use as such Japan, building particularly and Southern Asia, success our our we Following Europe, factories in are restaurants, and fulfillment and

one products Mobile's markets. are the offers electronic of demonstrated market with AI in their the compete in in ability confident to have We our to And these Chinese service robotics we for operations, extensive succeed potential overseas opportunities that experience believe international largest commercialization. globally. industry Cheetah

as serve AI and a at deployment generative for robots, and their accelerating LLM enhancing commercial machine intelligence scale. brain

and number early to to Mobile balance and is from best a authorities rollout. on they LLM believe industry enabling reduce for this by and through take to of scale full offer have enterprises costs company robots robots, performance approved at scratch, wheeled as solutions stage, an for to only between Cheetah cases, larger still affordable making cost this our robotics labor telco We we of make focus LLM its unlock China service across them more local use the the the will the use However, because working its time potential. service is trained growing in robots. hard a

We serving cases. in through [ synergy robots, focusing ] LLM hardware, software our use to various power customers on use the

We continue enhance have feedback products LLM creating the voice-based for use notably our IR This deliver enhanced will performance to us and the loop. and capabilities, experience, to our should robots generated and data product cases, market. in price a to allow reception-related global particular use positive for approach best we in

remain to striving innovation, the cost At committed always same we effectiveness. performance or time, product balance to

Applications making for earnings is because discussed useful call, Cheetah enterprise enterprises. how Mobile in address developing we LLMs enterprises. to build need are the avoid are applications. errors LLM industry-specific Gen last to enterprises we helping effectively company-specific and During challenges. is knowledge LLM-based This applications key or AI

are create efficiency. to apps our work to AI employees Gen or improve encouraging We tools

are we helping their closely them time, AI streamline app develop accounts, key Gen same the At working operations. daily with to

identify Gen goal Our where that features be is can and other to can and companies. enhance efficiency tools then cases AI to standardize scaled

optimize and in positive. a After For across it feedback helped a Cheetah reduce an helps AI platforms. China training we've also app, customer based at management for hotel We've it using enterprise cloud example, app costs, begun Gen ServiceNow, offering major LLM been successfully has to enterprises to Early users develop monitor clients. programs. cloud employee operator

gradually AI further products will portfolio. working a internally will approach experience their refine This developed customers us our business help we and ahead, product product comprehensive the closely Looking build we've with continue operations. behind key to Gen

turn to I over Thomas the financial for emphasize want I call the Before following. to highlights,

market and of and businesses represent a opportunity apps our First, stages early the development. robots service still each LLM-based in are enterprise huge

with capabilities, to on gains. focused products with Cheetah has in best opportunities, areas, than rather investing AI PC short-term pursuing team the the developing long-term these we growth Second, high quality, from are capture a Mobile, extensive achieving that since along experience mobile and strong

otherwise, Sheng. Fu RMB that money you, note Thank Please call. everyone, on terms. Hello, are amounts in the unless all stated

XXX in Cheetah Mobile In by RMB XX% revenues million. RMB gross solid a by quarter. year-over-year, this at reaching Non-GAAP delivered XXX.X total increased grew XX% QX, performance million. profit coming

the at margin gross remained previous to non-GAAP Our XX% compared stable year.

a improvement reflects from reduced on QX, our non-GAAP successfully initiatives. losses RMB our in business RMB loss In our Internet strategic This Despite million. AI decision X to AI, sequential by quarter-over-quarter ongoing reallocate our our to to resources operating million XX decreased basis. about investments our legacy operating we

Internet at our business. Looking

expense, in the remained margin quarter increased XX.X%, share-based of flat Excluding year-over-year the prior in a relatively QX segment increase X.X% the and compensation Revenues previous this X.X% from operating X% to up year. with quarter-over-quarter. from

by was acquisition OrionStar. As operating year-over-year a our we have driven in following Beijing indicated investments AI, increase in before, stake the controlling in of losses our

to well our for increased attributable these as higher Specifically, headcount increased losses service hardware-related R&D, as costs and G&A, are sales in robots.

in employees a XXX around we quarter, XXX had ago. and compared to the XXX previous As of June XXXX, XX, year approximately

We are are bear fruits. beginning that our pleased report to AI investments to

field as Our service key revenue a emerged business driver. robotics has

on large language valuable insights companies we and leading with collaborations industry through model currently testing are deals. gained enterprise their initiatives, for LLM-based Additionally, applications

incur achievements our generation almost cash continuing quarter's strong capability. QX, cash highlighting we activities to losses, of One XXX generation standout RMB capabilities. million in management our cash is Despite the generated from operating in

Finally, sheet our remained balance robust.

XXX June along XX, XXXX, we long-term have USD with such USD as OS. includes of investments. cash [indiscernible] cash This entities and As approximately above state well-known equivalents, investments in in XXX short-term million million and

the open call that, we will for With questions.

is will management's management's and Everyone, of AI the in for for our into line. purpose any will call, in agent discrepancies Chinese English management case comments In questions Chinese, today's convenience translate call, that the [indiscernible]. during another Please in answer only. statement translation note

in you questions Chinese we translation available Chinese. ready English be take our a now to Thank X hear working in unable Operator, to If website are the are will you. translation, within on days.

session. [Operator and for please Q&A the by the Instructions] Ladies translation gentlemen, of stand English

question about the robots. my forms is One, of product

humanoid robots. on XXXX and think wheeled the Cheetah in discussed focuses you industrialization more wheeled differences a Nowadays, and traditional have the the robots. also first future people How humanoid embodied as obstacles of Many and about year the you robots of the Cheetah's been intelligence do of lot. implementation of robot different do products? forms robots scenarios. robots humanoid plan similarities, of How define

become a Well, more it's essence humanoid, Yes. are attention work, humanoid Actually, whether humans human-related think type to in of right? publicity. machinery. traditional I attract machinery has like more can of This certain robots complete terms or human-related and the is wheeled

deliveries on to some arms complete of But complete in point-to-point use fact, or most our mobile bodies robotic robot indoors tasks. the robot real

So say this. to meet whether of today, most in because mobile -- very can steps situation wheels actually believe it's I scenarios, of how the hotels, The is that factories, always restaurants, including rare.

of In the lower can cost many places, wheels be Anyway, than is ones. and it much these bipedal fully there slopes cost has advantage. the are that wheels Then realized. themselves a of

be seen, made popular clear an said or whether comes although it just embodied humanoid opportunity, there that be point thinks practical. intelligence or when robots. because today, particularly can Thirdly, Tesla video robots is bipedal humanoid specific a cannot embodied mainly time of It should be very intelligence a is everyone to

driven. industry has this So been

My there view wheels some later work. the that movement, same lifting robotic has complete added always arms can is to be scenarios, most been in but

complete for some and voice our product and such simple tasks, pursue So putting right, handling. into present, of picking interaction up robots, some will machines, we terms robot as bipedal ones. our will, arms and like But of we form doing do especially humanoid form not at delivery the robotic or small things in something of

I think our plan. this is within

time, for years think maybe think a this personally entire or several least bipedal the be thing. we very for to won't ones that be at long And long consider It wheels. bipedal movement I mainly will must also time, for So industry I a considerable a in take commercialized.

the how Where and pay is ceiling if macro increase if their large is of do compare budget profits experience charge Is How monetization and do B, historic on Based large enterprises help on models? are of any to willingness cost, industry. Chinese in current have the application? to enterprises does the the enterprises really enterprises consider to of the the SaaS, clients from large budget affect to the how and charging experience, et which practical ability help models models? commission, overall the enterprise high? is of revenue, model very from? the for enterprises you in A, different cetera? do effects, large the investment reduction come relatively can environment, it budget it your be? low, Two, In based take model it may large of potential possible

before software. [indiscernible]

in pursuit historical such think many I China, environment efficiency In be today, But enterprise's fact, of may an factors. is in increasing. there

that, large that because not recruiting itself. by is can new, so model people SaaS for enterprises few especially do a the nontech very itself it's something is Another

personally, late that in think, I it China. emerged relatively

already the was few of talents Before China. there formed, software division relatively were labor in

enterprises and or for for free companies So do large could it some it need used it themselves many everyone.

enterprises. It's seen, of comes the implementation models, a very actual we today, as API few a that the the well. large do of can not But deep requires it to with people have when still have and you scenarios integration an take it application

seen are with clients, trend. currently several cooperating We have leading such we already a and

teams is which professional him part enterprise as answer enterprise, negotiating sales a we still needed client Then large enterprises really a willing. are can two them. to to the with save costs, quotation money. is the are questions, a This if very plan considerable save help to a help an it discussing take and the So budget. currently It of cost of amount helps

for a also we part disclose. is budget that disclose. can some But as can clients, of clients, the for we we taking promoted okay, So it including have some

your analyze on machines the we you example, can be cloud system for idle application and cloud management which how AI have many saved. can an the For help of developed rate cloud,

of have now to very popular XX% cost. combine the run enterprises cloud several client. and with to service This maybe save this accounts the XX% is can We of

is That say with to they X.X% promoted fee. usage cloud to entire his saved it's the are pay. The willing have have several charging entered deployment about model are period software about budget. Currently, negotiating as the of we the we clients and is

taking do from are the it's share to Essentially, cost. a willing Generally, enterprises this. saved

possible. or increasing We for one charge it is it's think effects the on also second commission? revenue, based The possible to

purchase the of successful worked them. to build have commission enterprise This stores for will large at We model this share with application will Later, the their present service, another a is also and we -- chain business franchisees with them. model okay. seems

really I difficulties huge is a think that think thing. the enterprise such very themselves, such the if I for you really about As for application, high, don't there technical talk ceiling is whether

being itself, in able but closed doing are difficulty test loop to achieving from a positive not job reputation. is understanding or that the deeply real ourselves. meeting doing product ceiling needs We applications achieving come a of does results, the technology truly of good withstand the of a enterprises, their applications the truly and The real from points, and needs their

of advantages. past one is our This

this it, the are a that think to how parameters not really spend also this in are and ceiling, model is many say conduct companies the such or what willing to Everyone with I stage, time. that about large is of because application So lot talking at model is the of things. in enterprises research still there

this I opportunity So think for an instead. is us

depth conducting to that such we the previous this is very into we believe with in standardize set. other promote much leading components good through product, difficulty that enterprises, them of enterprises other than opportunity Moreover, applications has or industries. AI And industries of a simpler crossing research such that the and can

amount relied also on it Previously, a large code alignment. of

be is so and and amount key smaller. the will loop amount of is So now new code Well, very codes generation, has understanding in achieve to the But a closed experience. much of in large. process a AI the own its the environment,

my Okay. This answer. is

has cash. are dividends? the generating management these to does on any books the buybacks plans company funds? there share privatization reserves still its use for a plan Three, cash is sufficient and plan? Is Or or How there net

you generated net raised. and cash said, cash this quarter. do was as Indeed, have question currently sufficient The also we reserves

economic However, reserves. uncertain, the environment, particularly that in company maintain cash we is overall actually, sufficient believe is for current the important which to it quite

we cautious that when relatively to So strategy. the resistance market flexibility has facing will company sufficient maintain and continue This financial ensure fluctuations. ability can a

because a has company maintaining as can company's the believe also is At [indiscernible]. we transparency provide plans the and to better announcement has market At plan, no the listed that Cheetah helpful of present, liquidity If the for possible. no soon for as same status also Board Directors approves the shareholders. plan, Currently, the will the level. governance enhance time, dividends. company make specific company privatization we relevant as buybacks share of to we and an currently

using to inclined towards We robot the idea How with most Or have does large with companies advanced end-model seen consider cooperation collaborating more implementation. integration? Cheetah models model product manufacturers? model it large companies large accelerate the is of

do At this excessive the that tokens we now. with to this powerful use most what of models stage, sharply Because cooperate and the currently, said will industry, large cost which of manufacturers level certain supply AI is that basically, it also can to models need and intelligence the large has in is the entire now. dropped be it enhance model robots, of we is

this our is and So of for kind advantageous very cooperation cost our for rapid implementation.

we period Our train idea experience is exploration of end. our gradually making will to is after good, own that for product if necessary through model a see it the our

have need basically, parameter release trained at now, model. we also stage, and We one, XX this billion the fact, think, enhance RMB still [indiscernible] first. In we shall a intelligence model. to But I product for product implementation already

implemented integration. for still take of sufficient is end-model model That or the for better back the you. model model cooperation now, obtaining path our configure train some carry manufacturers. we But After land. robo-application, experience, then will will better be term, in our and end. applications Maybe the that the long can with a come we Thank on large we out lease for

Five, that products? seen you How robot in are to similar many we are have robot projects Where in Cheetah's robots? product differences start-up the competitive the making view forms. very advantages Cheetah's do

mainly in lie is advantage I our think, three aspects.

is aspect that technological have still we sufficient first The accumulation.

of years. X or in accumulated field for a We have service X robots

this navigation and it So whether development is the our very based our secondary dialogue the level development can the current of and whole system. indoors, ability it's system agents including of on do and or take that the is our super customization intelligence that our of system ability robot, interactive autonomous robot,

concept the have we Currently, own trained next, industry. already be observable. model. there model is, the we This of will leading will are their And robot a in Today, That companies that is robot. few our large the should large models. in we industry, very introduce launch

also possibly We the at responsible of software the but the operating one team our for now models. them, is design least only hardware or of one and has but ability system, the maybe of are large the robot not manufacturing first it for

level our technological are ability. mentioned development in secondary the intelligence just and reflected advantages So interaction ability,

one Through a years, as that there productized can technologies second robot, because mechanical are such truly we The is and actually, so-called a various technology, technologies. many see, robots. various everyone navigation have robot is many launched several of the circuit such as combination electronic structure technology,

relatively product in usage our various through and team might fast. run really sample, scenarios of a make much terms it is make go than inspections and product this in aspect think to to higher capability, in user be quality quality But enough current stable before. many already So I to

or launch it not product general easy Our to product scenario, small start-ups for new a new to promote be ability a compare. should a in

one years have is explored initially many the ability commercialization or of third The channel. of of that establishment too the so accumulation, we

We are have in in they agents quite globally. That quite China and of are some of countries significant. agents companies. The dozens hundreds agents is large

are business quickly to users through our products We this to channels. network, build our enabling cooperating reach

today, and that for on our and us So users are development. was to finished we for our lightweight and delivery conduct including for feedback needs our the robot with prototype, channels also conduct the recently. It can now research research we the based looking inversely we from of require, development prompted factories they products that agents but the users the launched the

a loop from research a robots to the development basically user to think this our several loop for Cheetah I sale experience. and formed So is closed closed have points That today.

point that. technical also certain closed this It's at others do our point. real think loop because is a not I differentiating can

like short. not for loop, this the Tubi, But be of cycle the a especially will establishment channel network closed too

competition too worried the not projects. from start-up So such about I'm

and robot how After data attributed? data generated data. our this iterate we of still to Can terms are in exported products models is Sixth, continuously use overseas, the the products?

overseas, of stored is the access. general for and accordingly legal Then requirements is to on handle principle and necessary it of regulations. reviews process the of collection their encryption, going In server and on laws security undergoes AWS the it our strict data, different regions this based storage the and overseas

measures. Control

standards. the Management Information certification management Intelligence Artificial Certification. Management Security with ISO System certification privacy the We also to system completed information XXXXX, in XXXXX, ISO ISO reference Certification XXXXX, System the

obtain the are to company We systems Asia in first certification. management artificial this intelligence

the standard security usage our overseas systems this In went establishment international certification. data because actually complies these we of before. ensures this with early, that So regard, Cheetah problem encountered

We to these when went problems earliest. also regulations, Europe great in APPs overseas importance inheritance. such also other have we the encountered and compliance, and our experience Including such GDPR attached

Then in of and not machine follow involve data, is just use of about, principle necessity problems. the not to of that personal that basically, you does that information asked places. principle usage is the of nonessential the in terms analysis many minimization ensure we of This operation definitely

will For basic example, in some some just achieve functions countries, function. off turn some to we

cautious part model We is, navigation, data not question the most the model? of regulations. overall, are usage data of language foreign actually, whether by this Actually large laws things like [indiscernible] the of Well, this, comply need involve already AI the of like them another very and very TI in not be those and this this involve large data it for local do does we capabilities training. And the do this and So reverse training for used real not in provided or being. time with train are of data the good. to today, can

very are attach security we today, importance and to data Thank general, overseas great in So compliant. you. also

a enterprises such cetera. capabilities, product variables because and models, large et when innovation. entrepreneurship, My only starting is about organizational need data, in requires users, as question era of considered, business not developing customers, many to about the are technology also but products, organizational is models, Especially be end

organizational In than much integrated spend you results company, the this has best half more year, on Cheetah the significant a the into recruitment? does do innovation? what your period recruit talent was opinion, time achieved since during talent? listed How Orion Cheetah in How Orion of

Okay.

acquired Frankly things the routine a innovative particularly after of capabilities because Cheetah mentioned, this. speaking, organizational work, organizational has our me the Let start done you innovation is entire in many we with not building organizational Orion, especially, long-term and reform. is as

whether [indiscernible] Cheetah today, has on gradually into they In business, our a right, , sales centered XB are core fact, international our the of it's all advertising around company the with transformed a robots the to or agency business, focus sales.

organizational capabilities, year of myself organized company. trainings to is in the sales in training. So by thing participated first the in this training I various terms strengthen

in effort of and a We lot lot a also have money. spent put of quite

Tomorrow, we still have training.

position not area sales. The an -- sales I good in in was second is Cheetah called before. have But since started at XXXX, where the management. was This myself our #X

So we change management XXX the to important an Huawei's made and in approach the be from learned in of iron many enterprises learned triangle from also system, sales and our organization.

In the team organizational XXX have innovative various work. In and this because should some mature why in The I of the say the weekly regard, XB of itself applied is, management think, it's references. a report methodologies of XB we it's set lot management our has it XB for approach own sales first and reports a reason we process entire done AI, to hard management. sales daily to of implementation management, the for

from We implementing and more models. learning are advanced

today, strengthened year, [indiscernible] many key form key change to in or is for has reached our obvious before. do we certain accounts This different truly agreements customers. or compared cooperation CA industry-leading that with to some or ability definitely have businesses, customers, One accounts business

which are also mobile an the we And some for automobile have leading manufacturers, phone phone give also manufacturers ones. some AI, example. leading cooperation with Let's like Yes, for mobile manufacturers, in hotels.

with all organizational on have changes cooperation by about cloud. be These should brought some the them We reform.

not easy XC to I XB. from Of to that it's say to ship particularly have course,

of explored efforts for We long taken time, a have lot some spend a and detours.

compared comparable XB of to sales, integration Now sales entire that system, it made and this progress before. the management has of set industry seems including

time. some related something week Regarding or I recruitment, Every spend to I recruitment. interviews also have myself

mainly because have Of to say a currently. XB have sale such on in sales and today, We some that some XB system, I several have years building this talent focused don't excellent gap finding talent course, spent putting effort. we we big

training. recruitment talent. on depends also them. will the year, to addition it also myself top to company may our and importance entire not in of time attach campus Overall, campus We spends recruitment the internal training This young lectures cultivation to give the we lot of great restarted a of I recruits, people program,

is Thank our this So current you. situation regarding organizational reform. management and

But model cannot in effects these better truly the Based Chinese experience models if meet shows being to on see level, Historical practical comparable this you them At developing which far applications and from are how demand. better at a enterprise have be implementation stage, applications that companies are particularly based useful? applications can underlying overseas? on effects intelligent large those the the reach of applications. China experience

is effect vacation models. today's The large XB first of one the

like first usages wave to large the XB text model the In of to image, of effect fact, text video.

this that there's is the seen companies grown application. rapidly. demand and creating a for This field text. all writing have have simplest We graphics because in large actually It's

I Now entering think stage. it's a

what questions. for training, need you assessment them. with some example, All read doing Previously, are client system and training practical For an new could answer internal employees the documents and scenarios. only is we

AI, this an aspect training with then This can and Now conduct and of you interact is assessment. customer, a right?

quite also satisfied. Currently, seems users and the effect good are very

assessment document a enterprises type aspect is within has training employee This of internal skills. one is market considerable content related new that training, for which the to skills. is this training So and and

the seems internal make knowledge models you of the simply of randomly it I as this think enterprises, system the truly of useful. product said cannot and truly capable at It because used requires have to useful. requirements a Well, they present, be themselves. are, and large building of enterprises AI large and achieving to hallucinations understanding right, models the

So lot this takes of actually a effort.

take some this time to it So enough make I the think, progress. for good experience will

when traditional effects That be actually buy. is are second also buy data organization which related The types, the to AI combined one with is much some will better.

like and every just buy to amount cost, application code. I intent for according the requirements of requirement, to had What some the the mentioned, then to of cloud implementing also of Previously, written, good I saving needed our code write see a example. that large large changed be it the it. to enterprise. and Using such understand of be clients is are models an

large use the ourselves, but we question, think Actually, the gap In not mentioned, is underlying For to particularly level today, you gap that I models level the not top model scenarios Chinese large. compared model many when models, second has foreign is the in experimental of the foreign intelligent. a level large usually the stage.

difference foreign not requirement To ones. to with need fact, services not the really do the China. switch this to we we no the usage. there be experiments because final demand. it in intelligent enterprise first and to as will significant Basically, after the feasible, But models there in in essential will you clarify the as is in finding conduct a to domestic provide in is many after Because performance difference scenarios, imagine, but compared replacing be domestic models,

popular very for a essence agent. is term agent is to type of this demand. there write something The Today, called of

combine mistakes model For to find by -- the and by make technologies. adjustments, example, let itself, large magic. some call quote do think this itself right then This to traditional

in really users those real advantages in advantages In good. to detail, engraving when it application in great have to comes have definitely ones, can demand actually But the overseas is and development full very this applications overseas companies fulfilling way, level companies in I the algorithms. confidence details now. think companies achieve Chinese compared the because of applications Chinese in have combination effect this I

I development, limited by model good. real so think right? underlying level, for application be It effect be will So the called not large will

for new I one. of, is like By just this released mentioned as wrote our one Today, an usage, a of scenario, It and application specifically collection can that optimization. through to version that some information the wrote that math you mobility. equivalent is is the it internal something Open way, problems they seems this. which They in codes see not AI improvement

when self-reflection model will answering in internal math like things problems, debate, various their have So refining. the

has The improved I is but involve entire underlying when application So of the model. not intelligence level only so-called level its level, model. the example good. the particularly performance improvement does the of gave a at will intelligence it this to lot improves, It that say the not level be that the was

As the in can not be the adjustments too the application. as is of bad, through large basic long ability achieved model it continuous

see scenario demand In this the now. can as we user done case, be quite can well

you. Okay, thank

perspective talk question. From about research and Cheetah's plans. The competitive the R&D advantages of and last development,

Okay. for question. Thank your you

Let repeat. me

technological in the [indiscernible] this the started think very wave advance in is a way strategy, that step-by-step AI than super a AI, more then products it be system at to we my more that I robots sweep we the to system, to which AI of first rather think somehow it a through complex now and update thinking was because X.X, technology pragmatic the to a inclined mobile experienced is also today's Up time. when and refined invest a of phone. more I current market. way, complex instead hot -- is

when artificial So such the from scenario right? it mobile integration takes a different is involves software, phone The they in pocket. operate as various user from different scenarios, technologies hardware, completely their intelligence of and a out

or the in front runs library. in in desk restaurant, it a hotel, Sometimes a a at

the of processing are is the different. your experience applications it inside, voice they all Whether or navigation,

different in the a in does think lie efficient in and is our this robots I chain ones, certain development technical say is an So really humanoid everyone are but this on ideas implementation. sales research customers, is whether about That not promoting some has making and today, point, can real advantage and think focus today. people to form I research end with linkage development and what the today, but

right? implement your occur when a to we After team be encountered, are the time, we scenario engineering these definitely another Then run after your batch, it and will solve are advantages, for this After problems. the or problem at efficiency enable problems can respond team and you there some the R&D in of capability first one problems your quickly.

an you and others it single and a very to business out the think industrial pushed can enterprise in this and your technical loop. the a compared I This achieve is is can't, can be certain today's of that this environment. point ability if say such the and customers. difficult close to group competition terminal, Today, is This technical from quickly the of do that the product, actually back product to to to point we close

have R&D. So what in we years through this I is these explored think

customers and find this Amazon First, innovations a among said is is there problems like what advantage, If advantage. the called then from bit an customers. Customer find It's

or the we found made restaurant owner we reach was mistake right. ours. it thinking quite of saw concerns you product we from Sometimes, out that good. kind turned the before when something users, everything But of may different owner when a and this hotel The about fine was that be

So the a for only such realization our R&D can this process. in thing, establish ensure the future to system right, we of own

one. first the is this University,

this. like new on today technologies view my is Secondly,

of new have to in and may do size, will for today's model added huge. one embodied be and technology asset I think how right? is we we we embodied be is think time cost for inside, advanced particularly technologies, be cannot the in no need I of to Sometimes, arms advanced, robots what intelligence, intelligence, the there our to happening Even error a company conduct now, understand intelligence. robotic it. experiments But some to definitely of artificial trial including large

and So them apply new in follow we technologies gradually products. our

the to into with which robot, be with to large high-speed and completed This the model capability. following now model robots will of large yes, step-by-step is call is the Combined by being framework. the of or core quickly to ensure are doing [indiscernible]. particularly problems it make What the entire one, respond R&D not we technologies model and the R&D users new team, scenarios, user our iteration and large pay printing also specific real high attention is to to a entire large sales turn done the we feedback

is we can regarded I think this as be an you today. what advantage. What have think

conference the Thank team. the can. the is we will much Helen earnings English have so from for call. for translation as the Thank as we available joining transcript being soon Bye. line purpose, participation. [indiscernible] you you This in all our This is and for

by the attending now present were you today's spoken may for conference lines. [Statements on concluded. now English presentation, Thank this you The interpreter in transcript live has an on and your call.] disconnect