NVDA - Earnings call transcript Q4 2020

Due to licensing restrictions, you must log in to view earnings call transcripts.

Operator

Good afternoon. My name is Christina and I'm your conference operator today. Welcome to NVIDIA's financial results conference call. All lines have been placed on mute. After the speakers' remarks, there will be question-and-answer period. you. Thank Instructions] [Operator

the Relations, to to turn President Jankowski, of Investor begin call now Simona over Vice I'll your conference.

Simona Jankowski

afternoon, fourth President Financial XXXX. Executive Officer; Good call and are Jensen call Chief you. everyone, Colette and to NVIDIA welcome Thank Chief President NVIDIA's today for from the me Kress, quarter Officer. With fiscal and Huang, of on the and conference Executive Vice

our webcast I'd Investor call be remind like without that is NVIDIA's until The property. you prior live results for replay to Relations consent. webcast for of the website. call transcribed first of our written content It is discuss will today's The financial being available NVIDIA's on be conference can't quarter or to the XXXX. call fiscal our reproduced

may expectations. differ During of number actual risks based forward-looking make subject are this to results statements call, our These and a on we may and uncertainties significant materially. current

and Securities All are as based assume future no our today of any statements. Exchange For with update our made such factors Except discussion may results XX-Q earnings forms XXXX, today's and available disclosure most a refer on and February affect recent by statements release, XX, to currently law we X-K of that Form information us. the required XX-K Commission. as obligation to that our to file reports we in and the financial business, the could on please

During this call we financial discuss will non-GAAP measures.

these in our CFO posted You website. on is non-GAAP measures measures a financial commentary, can financial find reconciliation to of which GAAP our

Colette. the turn me With that, let to over call

Colette Kress

from was great data was excess in above outlook, up channel pause with in gaming Full center revenue and reflecting year down recovered $X.XX upside gaming inventory the revenue up billion, Thanks, exited and well the momentum. our earlier XX% our billion, We year spending X% an and hyperscale businesses. X%. in and $XX.X sequentially, Simona. year-on-year QX

our up was end XX% for $X.XX gaming enjoyed of $X.XX Full Revenue year was gaming. XX% year. billion revenue our notebook sequentially. from down strong year-on-year with GPUs. Starting prior XX% We and and down desktop demand billion,

some me Let more you details. give

Our for channel GPUs performance the price Singles for of incredible demand positioned at capabilities strong well with season point. in West, was stack. gaming Christmas the tracing Day ray RTX was and From unique China, every our event lineup in shopping through the our entire exceptionally the the holidays,

low for Duty: momentum this and Fueling SUPER Gaming of products. Modern blockbuster is new $XXX, the GeForce. sweet now as spot esports as tracing and like were new continued prefer With is RTX games Call ray points gamers thriving price Warfare, gamers. RTX PC

Championship momentum World in keeps Super XX% Bowl. phenomenon global Legends gaming just Newzoo. million, exceeding The than brought XXX par to this million XXX with League more of The viewers on of up two audience now esports according years, with over an month's

titles come to GeForce tracing this to important support are continue the RTX only Ray that GPUs ones technology. market and

Sampling technique, were Youngblood support Super The as well performance. which quarter Deep NVIDIA's also AI to This as uses the tracing, boost ray Moon latest Deliver to Wolfenstein: and Us Learning titles

games performance continue Gamers seeking the top-to-bottom to GPUs, solidly ray ever move Architecture higher-end and With of RTX-enabled enter upgrade support the proliferation we our better Turing cycle. performance, tracing. for to are best and

expand, up NVIDIA retailers XX the up laptops form season holiday for GPUs, graphics performance. factors Gaming The based Xx. driven light consecutive fantastic appealing growth Max-Q double-digit a eighth record with to year-on-year and year, laptops with from our posted on last thin quarter. XXX designs by category stocked continues gaming This

ASUS. GeForce XX-inch, launched we CES RTX At first with the world's laptop

at fast-growing of CES. with laptops Studio creators' XX new our and designers to population continue expand RTX introduced Studio Youtubers of also lineup for systems We the freelance

XX applications, our creative over Powered we with tracing optimized service. NOW AI. accelerated GeForce by ray RTX launched gaming Turing GPUs, and and/or week, cloud these design for are Last systems

the first can open having GeForce enjoy store use they service already existing only the repurchase NOW games to to deliver gamers also GeForce, by cloud have. Powered ray gaming is And It's so trace accounts their platform, games. without games. the

GeForce enables on membership mobile Windows, that Mac, PCs, in soon NOW TVS, games Chromebooks. model includes a premium two devices GFN has PC plans. and business

support. the which a with with $X.XX gives membership month Founders priority free ray A standard access. per RTX price access tracing of tier starting And and

Our expand goal more gaming to GeForce with is gamers. to GeForce NOW

About on cloud. to playing PCs able GeForce on GPU underpowered With OS NOW, PC devices Android. XX% of they are GeForce gamers gaming a NOW enjoy or are GeForce or with the Mac in

expand, into collaboration can largest today. gaming the entered to beyond largest world's roughly million we to XXX bring now GeForce reach with China, the PC Separately, in well world's platform we gaming GeForce the cloud gaming the market. Tencent, gamers,

Tencent's, cloud will testing NVIDIA power in gaming service, stages. Start GPU technology which is early

sequential XX% ever Moving strongest in terms. record a up sequentially, year-on-year to our and data XX% revenue up center, $XXX dollar growth million, was

fueled year Full record revenue Strong up, and data prior $X.XX X% XXXX growth year fiscal hyperscale year. vertical billion by was the and industry was from center a customers.

purchases and language inference of Hyperscale as, products, support such both, AI key demand our conversational was deep and in natural AI training understanding, driven workloads recommendators. of by

TX a gaming, computing. excels workloads, region. including computing array by the of at the every desktop. driven inference, underscores and also wide now rendering, was versatility high-performance in of This demand virtual cloud available AWS the AI which TX, Hyperscale makes, cloud

driven engagements. logistics such primarily and growth as, foundation companies. deep health from Internet with retail, consumer was early-stage Other care, by, strong to verticals, grow build-outs, a continue learning industry Vertical of

deployments, strong of set shipments applications. inference up our year-on-year computing computing, And another cloud TX, we as see well times, quarter, video applications. edge high-performance driven by as X had AI public platform expanding with data and opportunities across edge science an analytics

training IVA's AI reflecting both for revenue. leading remains set TX inference and and VXXX, demand model the this respectfully, platform records strong and shipments for NVIDIA quarter for training.

and leading the enterprise some and PayPal, Express, consumer inference companies, wide of Pinterest, is world's use, gaining Internet American including NVIDIA's platform Twitter. Snap by Microsoft,

groundbreaking work continues NVIDIA. AI for do The to industry

inference, and improvements language biggest For for software Microsoft's past from search understanding year, example, NVIDIA models. use in over training Bing, the its its quality natural of GPUs of engine, and stemmed its made

These major are AI models, DNN in that high computational magnitude workload, than, models. latency. of Conversational new a achieve, inference the higher requirements is have the order within BERT, by for requiring popularized image-based desired transformer to low earlier throughput for training GPUs

with cited Microsoft per this BERT to throughput X compared week, Indeed, second enabling an announced up Microsoft for advances inferences it worldwide. language summarization over on ever And as a largest GPUs of million researchers inference assistance XXXx to with increase NVIDIA new processing serve NVIDIA CPUs the trained tasks model questions, such AI natural and language the state-of-the-art publicized generation. in and on answering natural DGX-X. breakthrough just This

model which Internet learning feature powering enable learning industry with second. queries inference and GPUs leverage to for is and Running Recommendators compared dramatically latency to are an also wide on The throughput billions moving automatic to can learning deep deep reduce per content. as important demodel, the of unstructured models support recommendators, CPUs. machine increase and such these

achieve For Deep example magnitude NVIDIA Alibaba's beyond on rates. throughput and orders Alibaba Baidu's enabled inference click-through recommendation engines CPUs. by increase in XX% to their AI recommendators run of boosting

time on take the months Europe, also the AI that Eni in on industrial fastest NVIDIA entire a Microsoft large of from customers deploy. can on-premise supercomputer to their energy the the desk, supercomputer capabilities announced supercomputers new rent enables supercomputer based We to first demand Azure. GPU-accelerated an availability world's And matching company of for It on announced GPUs.

customer customer AI At a helping we data order our deploying with accuracy come the Annual deep announced month, for and Domino's Conference engagement as has to last prediction. learning and applications pizza even delivery. Federations National Retail platform science

have More GPU-accelerated shrinkage, AI partners, NVIDIA's platform reference operational ARM, logistics broadly and SCXX conference, applications by we that reduce design edge ARM-based servers increase Fujitsu Ampere a introduced optimize with and in offerings create retailers Marvell. for large of along retail, At computing a seen significant in adoption we efficiencies. the for the ecosystem powering Supercomputing computing

libraries accelerating cloud of supercomputing. the to development tools growing made available of This ARM-compatible exascale support new CUDA-X software applications consisting computing. development We and our for hyperscale opens from NVIDIA to innovation floodgate kit

introduced processing AI modeling massive computing nodes a out complex researchers. computing sets financial data carry multi-server, other with software and to We Magnum storage scientists, up workloads for IO and suite delivers optimized NVIDIA and Magnum for input/output bottlenecks. climate analysis, when working to to multi-GPU data faster high-performance also IO, of eliminate XXx data

below introduced TensorRT of software in components seventh our we interactions inference conversational AI millisecond of for latency to real-time supporting considered XXX running Finally, the growth threshold comparing the speeds X, which This AI. generation our drive conversational development helps up XXx on CPUs. necessary by kit,

year. prior more record cycle Moving accelerated Revenue from is year and year-on-year applications million, an revenue driving rollout increase X% the of of $XXX ProVis. Turing reached XX% the a up QX Full in sequentially. X% upgrade a GPUs. record up strong billion, $X.XX was to as our RTX-enabled for ProVis

also segment Pro, and RTX market opening joined SolidWorks, is with V-ray leading our creative creatives. up Blender rendering began as for software technology. design Dimension, and and and studio including opportunities renderers RTX Arnold Maya. In These freelance November, new shipping applications Premier such CATIA

real-time With RTX, is over million. versus gains. and these than user for XX leading by supported a boosted rendering productivity combined a RTX creative notable Xx base design applications enable In reaching CPU. more applications performance now enhanced Blender XX is Cycles creativity example, and of

year a revenue year million, ago Finally, a record reached up turning year-on-year. $XXX to Full and automotive. up X% $XXX Revenue sequentially. from million, X% was flat

AI-assisted allowing of from During and investment product Orin the delivering and plus multiple scales SoC. to driving Orin, performance by the DRIVE, for next-generation generation quarter, Xavier the nearly Orin our autonomous across their leverage platform robots, vehicles compatible, SoC we previous platform fully and level The powered is AGX Xavier the with to five new operation. two Xx software-defined generations. level developers driverless announced

largely expenses contribution Moving data QX was products. of GAAP the billion and XX% gross $X.XX higher were non-GAAP rest up QX million, expenses center non-GAAP to margins XX% up were and was sequentially, GAAP operating X% $X.XX XX.X% reflecting was operating record GAAP operations $XXX EPS billion. P&L. cash the XX.X%, from and cash year-on-year Fiscal $X.XX, $X.XX up a year flow was year from operations Non-GAAP billion. up XXXX year EPS was was a XXX% ago. a QX earlier. respectively. QX from $X.XX, from of

the the XXXX. not fiscal quarter outlook of turn include acquisition outlook me of first let the that, for from The With any Mellanox. contribution pending does

China acquisition part with the progressing first and XXXX. regulatory likely approval believe engaged and of will the calendar the on close are in We

numbers me of comment we -- to the coronavirus. impact on get new the Before let the the

early estimate, reduced by million account is outlook to to and ultimate for it difficult revenue still effect impact. our While the potential QX $XXX is the we have

X%. respectively XX.X% plus be and expect gross or margins are plus be XX XX% to GAAP points. to and non-GAAP minus billion minus $X We or revenue basis expected

respectively. be GAAP both million plus expenses non-GAAP GAAP website. excluding In $XXX let Capital both and non-GAAP minus operating $XXX income for $XX upcoming financial included be GAAP items. $XXX details approximately expected are OI&E expenditures to to and highlight billion non-GAAP of are are the be other expected community. on information IR to approximately and million. million. the in available are be million Further, approximately to $X.XX discrete X% the X% rates expected commentary an and financial and or event CFO tax are closing, me expected to

at will We Conference Technology San Telecom March Morgan in Media be Xnd on Stanley Francisco. the and

With will you questions. for the for Operator, that, open we poll now questions. will call please

Operator

first line Toshiya question Hari And our Instructions] Sachs. with from the Goldman of [Operator comes

Toshiya Hari

helpful some Jensen, can talked having quarters, question. in TX and or very guess the Thank upside also a if center What expectations. about the the ago, years you the for areas the were of and expanded. inference of Hi breadth you. I your relative of businesses that quarter? record the profile but the can of in speak well. how would the internal much some You Thanks relative on be That to guys. to your customer that's drove you both VXXX that Colette to speak today upside to the couple data drove as

Jensen Huang

growth driver four Yes. fundamental AI. There Toshiya your for thanks are dynamics. lot question. for is The primary a

amazing quite models natural recommendation is that created systems. AI the and fact in breakthroughs achieving breakthroughs are first are AI in being The language conversational understanding, in and that

reason And trillions you so content of or it for and the others or of are of know this, hundreds are there the in whether the because engine you it's the but world, And information of is items the audience, are or world. store that in it's are Internet many systems essentially querying, depending the you for recommendation in websites how the there billions, on count trillions whether items hundreds a today.

the contexts. the in own what And provide billions and and location affects of looking the weather environment those users the items kind there they are answer each what's And between or and people query the countless there with that happening that search people we're in contexts are the for of you're -- and their the contexts characteristics various you.

just is foundational search. system recommendation to The now

some is last done being end of And have world this year, is the done to Work around be recommendation said beginning search recommendation systems. time the the it's in very advancing era systems. the people And in and the able over in deep of learning. been first everywhere

the And thing breakthroughs first so just in is the AI.

inference. models second is continue model production are larger means into production seeing these training we're AI going that in because and more that to The of significant opportunities the them had is business we have we a which significant and whereas, called of getting and lot is there

is now. our times one X, was is, entire the about an complex of Inference statistics year-over-year. over an enormously TensorRT a grew a computational download mentioned, interesting part XXX,XXX, doubling of ago. people business substantial software Colette year What the this complex it's problem, don't most But it's as a about but year understand problem. TensorRT Inference It's X incredibly

going moving dynamic and is training growing training called models Inference. so, or production from into the And second from

is and developing start AI of works own because software in The thousands but younger. their cloud become just the thousands that as now infrastructure anymore, AI they growth better they vertical start-ups the much but could the are When Public OpEx cloud lives not industries. public in center them hyperscale cloud. third in model cloud. for in the to the they're And build of on-prem, decide of start-ups their data in larger

We're verticals. also seeing in really great success

I Walmart they're amounts analytics companies that or last enormous data this have do on end of predictive vertical trying announced analytics One data logistics; do large think the is Express, exciting just retail, American We most -- quarter last quarter logistics, of warehousing. and of who to USPS, on.

really the and lot as then and a dynamic is And And growth is in vertical hyperscale AI excitement the exciting We're industries well now so and AI. called -- being is and beyond a industries. action industry to dynamic about is. But used last AI. is the of third as Edge We public where it the called this cloud Edge there. it's call the really, seeing talked lot

need And to have latency. low that the for you reason is, inference.

reasons cost need the data the you or stream the reasons way response to able be not might You all time. to cloud for sovereignty data and

four around growth. really those our AI dynamics so And drove

Operator

Your with Joe next comes from question the line of Moore Morgan Stanley.

Joe Moore

Great.

Just following up on that.

you back first to business at that XX the in place? just last As perspective cloud the you thoughts the you look in digestion? months driving on and the drove seeing, slow any what's that rebound HPC deceleration the what it Was of now your saw it have down

as guess help we business have a that from of newer handoff just to these And a kind this there? of pertains sort it happened you it? to image will of Was future I about? applications talked Just recognition that do the lumpiness think as what us to it that

Jensen Huang

really a That's question. good Yes.

incredible. you AI The systems, language models only breakthroughs which continue the evolve what the the really the advance transformer In here if last history. the all from the this and benefits captured to happened understanding was natural deep these have breakthroughs backwards velocity so-called industry And quite in we year. recommendation of conversational fact and by benefit look

grown all have breakthroughs tremendously grown of number the has of so and And in tremendously. a models the AI sudden just these

X.X Megatron understanding the this in has Just size's has XX.X just of did call a with their parameters. the net that collaboration natural accuracy model increased the really billion announced of -- been Microsoft we And parameters that to training boosted. language week last neural from billion work they're we it model

models is are so are there And finding them. AI more really bigger fantastic of breakthroughs models the -- and getting are and

when And essentially you look when last happened happened at look back year. it breakthroughs and these this

those inference. growth into this on second, the inference been models all And until learning really working in across few of of saw The last production. went in now year year some models we've hyperscalers great for this inference we we have And And very time. production. deep last

is able now of there's they're over out production. All AI of TXs The dynamic started about X,XXX starting And are develop clouds. and third with put their to the every them startups have being all world be in into AWS models the that public to these geography. scale single we

of the of sales And So the deploying the the kind of TXs. combination that and a and just GPUs the And came combination availability happened cloud in result together. this models last into of all year. our we production startups had all industries and record VXXXs of as that AI their vertical of

so developments powered all by And we're really it's AI. with the excited quite and

Operator

question Securities. comes from next Bank of Arya the Vivek with Your line of America

Vivek Arya

How on perspective my year? growth. wanted business about running Your to from taking foundry coming for congratulations pretty strong how main you for ask you the to for? to growth supply and are returning give positioned I Thanks question are positioned some tight. that Jensen, the looking a will back is you're could supply If this a many very the here how investors us be XX% will you on perspective so that be or support able from just helpful. commentary

Jensen Huang

across won't nodes the multiple process tight shape we're industry pretty combination different think and It of We've -- We different of the on got factories different a true and that supply. of lot of is I Well supply. our surely have good lot is partner processes, ample multiple in a qualified. TSMC. fabs supporting several

good in shape. think I we're

of very a foundry and closely all it forecasting TSMC. we job -- that of we're our better closely. partner and course forecasting. us our do working have us visibility and between closely customers as just very all And And have customers better with our gives of watch so well but and to to And working we're

Operator

of Arcuri UBS. comes line Timothy from next the with Your question

Timothy Arcuri

Hi. can how a -- us and give I if maybe a Thanks. the much give us can something hyperscale? Colette, think or in wondering mix than little up if between the that? was XX%. data little hyperscale you I'm it idea give Can mix center industries you Thanks. what was quarter us of bit you less was like last

Colette Kress

still growing for growth of continue question. moved all in this with seen of terms thanks last Tim quarter to we we Similar split to little We're tad a expand maybe looking overall the same XX vertical hyperscales things our quarter our the and terms and in total cloud industries our hyperscales bit into what hyperscales. industries in Yes. even the instances. at had XX-XX and terms vertical the those around between in below of of as

Operator

Your next question Rakers the Aaron line of Wells Fargo. comes from with

Aaron Rakers

you the basis taking When seen on data or in Through would north sequentially of the Xx numbers that as the pretty was sequential absolute were in I we've growth past question center for as results. quarter. and absolute Xx change. what on course Thanks clear Yeah. see the almost an at of the quarter an congratulations acceleration look of far December this the you in expect to growth the the

just going that about this into any how If framework. And the we give rate course about of through think can year? quarter? April of I'm that you the of kind you curious how us growth think should

you you we the the for mean attach stand looking next bigger think I or think the might from of years curious out be so? where market? over as or amount acceleration do rate just GPUs in in the you. Jensen Thank about is of server perspective that three terms And today think where an industry the And picture you do

Jensen Huang

you go Thanks want Colette do first? Aaron. to

Colette Kress

still go into QX expecting about move well we're going overall and our expect do think we we to and our continued When into that into going We and data growth, positive we remain it Sure. forward. QX. believe as quite visibility see both growth center

Jensen Huang

will recommend the and from have a at sort Aaron And language queries have I query recommended down through to that it. handful queries. filter be make Internet the recommendation on answers to Yes. it will to natural all of understanding the core and trillions it believe very some every of all accelerated queries a -- possibilities almost your component of someday. most will Almost

That it's or the down. of even answers needs be answer possibilities or best deep process locations filtering is movies asking query first recommendation Whether accelerated. involved -- will really or is believe number to just system to all and learning the called asking be that is is question, is be shopping I the filtered of all every the complex versus thing. that. that recommendation. a in what going That's And

has The ended. CPUs called just without for really to about marketing second is It's increasing the end of continue there's a increasing And thing. thing. Dennard no a two scale slowed as you It's And to know it. the ways has physics ability it's not scaling and cost or power CPU scaling.

has And be another so there to approach.

the use with and learning artificial of of benefit acceleration. but suggests necessary of for single be the there computation that query, every is that comes emergence to the and that scaling needs combination to intelligence the The amount another the along of approach that believe that's deep Dennard approach and and we of end

evolve programmable. software-defined fundamentally been data is accelerator. the will And for different accelerated data get acceleration accelerator, have evolve for to is say be believe approach a our models why software-defined we Notice AIs needs larger we successful. and continue that never reasons say of different a The It that and all will we reason center because than to kinds center computing. Now, the will so one continue of AIs. the to we've is an

the of the of creating. that ever-growing and programmability richness models, are questions go been the last the asked have that and four over these software, the think do me of this the answer the with you or of compilers, models of architecture, the difficulties if has the advances around of And consistency the models, of years size to the about the the all diversity area the back three

of And era. a so beginning the new computing we're seeing

is accelerator function fixed the not answer. A right simply

these so to is require believe It's that we be centers data defined. an software so And software and vital accelerated. accelerated that platform future is going really the could richness be going computing to

of the to new think the And is early more early become innings, early this computing think I innings, that in innings that very going important. And more so very, accelerated future. we're I and

Operator

line next Cowen. comes from the Matt of Ramsay Your question with

Matt Ramsay

you -- wanted Thank bit units of something little how of about One, a your business gaming it much. thinking a is from of I $XXX $XXX revenue -- you coronavirus, congratulations that and demand what ask at in I data about you in China? disruption us, little sales? how within might the chain to guidance, percentage the pieces. a Good or to took and thinking the to bit success. are about million are you Really limit Colette your And could Thank million that if ask for very two wanted on supply put afternoon obviously Or center the remind terms got number. that as maybe look perspective? out you. you you that about you that from out you

Colette Kress

the Matt. Sure. Thanks for question

overall what out in really impact figure from to it's trying the of maybe. terms still quite early the coronavirus So,

be. much estimate. So, that not also might understanding in still our split challenges those Yes, impact we data and a between The maybe overall and get estimates may our are on $XXX terms demand. But looking pretty or we're be split gaming think necessarily possibly what million center, at reflects of we're equally. supply to where precise an better

of terms our and important gaming business business yes, In makeup, overall piece. for business China is our our an

as our gaming business. gaming percentage We a of have about overall our China of overall XX%

a the it for For They end just customer moves data but important quite overall a us, choose. well builders bit. from are market center, the that very as mix based they as system moves on may it quarter-to-quarter

So, it's a little harder to determine.

Operator

Sur the of JPMorgan. Harlan comes question next Your from line with

Harlan Sur

no and recent the on launch cloud with to infrastructure the be mainstream? standalone own the already But model. congratulations But have products? -- service. largest to yeah, the Good the NOW should your place. GeForce your see have gaming with you gaming, deal are be Good going smoother expect seems They On afternoon cloud is the Tencent Tencent the like they results you many the just GPU this in of them GeForce company going They own center so it on transition a their will framework? and world, in guidance. strong to team do gaming problem. hardware how partnership? the your of on supporting also service to games. to NOW Or go be NVIDIA when And supporting Is partnership data

Jensen Huang

Tencent is about transitioning is the cloud that world's is the long-term going reason the third publisher. the connection be a gaming of see. throughout journey. Internet to world's not the largest for represents China is entire And consistent Let's to because might and market. onboarded extra be reach. very with billion we talking needs We're lot gamers and we're to If working of still And long closely term, application a successful enthusiastic we're we're them. an super able that it. be about one about to a

And so, I think that journey. this is an exciting long-term opportunity just the

around hubs the as connections West, the and more to the through beta the local world as WiFi people's opportunity routers the lot well working refine in home. here a we've Now had at centers,

process And allows we're We're now, cloud. open. here been West onboarding to in at know. another on And as the And games quite us for data cloud you some and that's of the games. Windows, several X,XXX games in so, we have only to platform is to hundred the we've and time be platform able our bring based PC

had our you And here guys in -- with been saw we so been Overall, is NOW, about GeForce and have we've more we West a -- that the the we have with Tencent. had the obviously the has enthusiastic can reviews games have we've But fantastic. more reach reception super I'm launch, the partnership lot reach onboard. experience fantastic. The

GeForce the components. three has service NOW strategy we Our provide that There's ourselves.

lastly, example platform of done the just publishers, engineering to course well, of our in world refine to that offer have a a like the that. telcos alliances We excited with is And GeForce regions around the world presence also I'm and have don't software collaboration that service. around super and of large we to them NOW great deal in. going we that partnerships be then with for And reach about the And has Tencent. and a lot

Operator

next Muse from of the comes Your question Evercore. C.J. with line

C.J. Muse

a taking the Good Thank for I on that? you better you speak my overall your guess trends look guess Yeah. guide, to into side. you're seasonal guess, for April. to gaming And suggest question. I would looking typically, I seem If can it that afternoon. revenue question at I

should about you just available? desktop modeling then July going now more April as third as and part And how XXXX? for of How we how demand think trajectory With really are the you. into a Thank revenues, with seeing of growth should of the about question becoming seasonality then business? through gaming that is content recreation your of And notebook we gaming. part think kind the

Jensen Huang

Yes.

C.J., So it then Colette to take going go first want I is here. and home

first gaming at -- sorry. has So is of our part the the I'm end this, business it

Okay.

business super market that is feature It's the year. RTX graphics. tracing end doing It's clear now And really of healthy. -- the new next-generation is the been clear gaming healthy really throughout fantastic. terrific. demand whole And Our important ray most it's pretty very it's it's is that

out been have games games over going is XX that are -- XX games just coming of have or really been so XX exciting. The pipeline shipped. have announced, to that be really, We tracing that ray

finally and one we have down -- RTX, The second RTX about $XXX. more to factor taken thing

So at spot it's of gaming. the now sweet

fantastic. doing is fantastic world. RTX so And is all the sell-through The over

profile really of sales part our of gaming is success The Switch overall the changing and of amount business. of is in Nintendo that gaming the the changed second our business notebook has

reason game that device. seen world be category. and of as is growth And people for is earlier, any notebook has consecutive is This Colette are this there eight in the unquestionably Like a the double-digit Our believe. game largest than more gaming quarters laptops console. new it's to because console for there the going with are I new a business mentioned other

And so RTX to in it's really that we're great such the into a been we've thin reasons seeing notebook light notebook. of and fact is a able and notebook, get why breakthrough. And a light the success thin one

and largely it and hubs because build QX. of QX largely the of And our seasonal, Switch it more in the the and business two systems to has Nintendo are Switch QX them QX, them gaming changed overall they quarters season. and around notebook put to the Between and more build a tend holiday ship and they It's takes them seasonal. become world. ahead built in And the and build into has devices like notebooks profile because business while

of But one than is fantastic. result it to Colette. tend notebooks. seasonal hand success past. the why of QX going and part the the a of larger the to be great. to will I'm And end QX tend of will just is to and seasonal over will that's be it demand more doing is so to tend And reasons be more QX our RTX

Colette Kress

Yeah.

more course from coronavirus. desktop, that the be of we sequentially in of as pieces QX. the those background a gaming it and with you have are all QX terms outlook of components year with from and the QX So seasonality gaming pronounced overall look at our Switch to within think a bit those all three normal notebook, down to the tracing all will little usually we different QX down probably for about is that due sequentially assumes This the that ray

to we're QX So, in in at probably in the looking gaming. sequential be low decline total, double-digit

Operator

comes question next line of Atif Your Citi. Malik with from the

Atif Malik

on And for and good job Hi. my taking you guide. question. Thank results

historic. surprised the why wider went same versus million the guidance talk I'm Colette, guidance the -- range is the you from not about just coronavirus? topic on coronavirus. range? Can what widen a the that the hit On that $XXX And into not bit

Colette Kress

the early still very thanks it's So Atif, for Again, question. coronavirus. the regarding

are out others are both with Our thoughts China. families that employees and the in the

time. more our have thoughts our best supply overall prominent as as our well at to and be. about timely see So on customers as the gave we'll our that our get overall We'll And million. But with that this that and to as overall into comes was our went in we our estimate just this region, through the we that the as how our AIC is when can at as quarter well we guidance very $XXX Asia is chain makers, discuss it. as into time both discussion discussions

Operator

Your with line Stein the next William of question from comes SunTrust.

William Stein

taking and the playing all X.X the last my to years market little essentially Historically for to thoughts training out. of Thanks you the your NVIDIA's hear Jensen, inference the in market how anticipate Great. so. had or question. love I'd of market inference as

I think that's in changed better where you've done much inference.

have edge. in growth? anticipate called how at have How at EGX portfolio the you And inference it's you endpoint you TX cloud, is do I the you device. the developing your aligning And for positions? have Jetson, what you think for that are Now market the that those across various

Jensen Huang

Will. lot, a Thanks Yeah.

a for just not machine AI small necessary Deep AI and particularly GPUs weren't Let's particularly still came being inference It learning the that AI, has historically of business along I amount developed. -- suited part learning historical -- computation until because learning deep of think for extraordinary. is the wasn't been was weren't our is suited see acceleration. classical

And and the The recommendation instantaneous related Eventually were these the is is answer, going the conversational type type AI AI of likely require the on and to language of answer. systems the second developed. someone understanding models factor faster models response. the that more click natural to

it's matters And measurable. and a so The directly on measurable. latency the you great deal effect know is business that

And so latency several with we've processing for of voice able the the take the the entire come That AI, -- and language to for recommendation could from synthesis. been speech entire such fix to for speech errors example text up pipeline to reduce to example, pipeline recognition a to to the seconds. conversational

the of of that We realm a to entire XXX in couple it's pipeline for the now milliseconds. so us run conversation. within interactive is it process That fast XXX possible to

too Beyond that it's slow. just simply

production. sensitive of us conversational a latency think and where combination you're models for now I are AI in AI that moving models the to see and essential and complex to are inference. secondarily, GPUs forward, And are then moving our lot inference more applications so And going large opportunities to that moving

And that it long-term is technology way and phenomenal to on that approach. about GPUs that Dennard suggests a learning, think that going in evolved to deep acceleration way of Process The has end compute we demand really of the is essential different is acceleration scaling. AI way. because is just a

data centers to the programmable I believe accelerated. models and computing all I be all running defined I I question, software-defined. And will accelerated in future platform will as essential. It earlier AI mentioned mentioned, be an be center going software think think it have be having another the to will is that and Data be and are to will I

move it something software-defined edge, inference out many, to As your many in temperature platform models it could detecting work could still based be inference fix be There as the sound simple devices in or deep type It's It whether or programmable is the you to as really changes Those it's specific. has on detecting whether depends on are where motion. functioned. is or whether learning. are – changes very of function-specific. be

to change running time. models it and given very you're two don't any or point have one You in at often,

devices that And be so you're to chips, believe incredibly going you're into put detections. I chips those AI those to AI are be something going to going cost-effective. magical it's $X $X.XX, going it just are and have and doing to

ideal so get and the complicated the Those The platforms much of us. platforms it's we're in, come really so such better. evolution that is there's for to software-defined robotics, type self-driving constantly and and to targets software yet cars going are as

devices. it AI And edge so call edge, the at we computing

what the with edge edge. data One call that going That and very XG of people recently devices, edge I'm we announced to center programmable. we're be the basically be telco excited Ericsson We mobile is, accelerating about will XG computing stack. or partnered

software-defined applications to those And be all those applications, It are going – XG. kinds be runs including so that are for center. And a us. fantastic opportunities to data needs of

Operator

Your with next question Lipacis the Mark line of comes from Jefferies.

Mark Lipacis

within question is make today staying investment And hear in a and to your made I sustainability sure call took your to now. is I that's my a investment here strategy for demand seeing the simplistic in question markets these R&D way for position view de year. even you're the are Thanks and for years you your ago you guess, my consensus Neural developed markets I in we And – XX market standard who last to distribute like, that investing products me ahead are invest how it guess your is too? And of think Jensen, CUDA you about your how about do you from facto interesting the the you. had Thank taking your in to about question. were universities. about out that competition networking the are of data software you and right off just the that what's of my think center. market customers the

Jensen Huang

Mark. Thanks, Yeah.

has to of company XX years Our the market. live ahead

in world is And to the five work to so XX in we and time, years imagine backwards. be have time where way years going our

computing It's systems singular company is one thing. of it one high. The the throughput on that is that our is of architecture is accelerated Now, all is about we're And and computing. because focused incredible. complicated Accelerated course. singular thing about simplicity the computing accelerated in,

networking's a bottleneck. compute faster of when a the bottleneck, sudden Memory's bottleneck. a XXX bottleneck, we a CPU, our we becomes a XX, everything acceleration XX, times storage can than can is a everything – all XX, bottleneck, When is

system – some course it the the be over of so But supremely our the really – software be to NVIDIA of stack last a really has the paid have is good to complexity And where designer. has above investments is now which years off. we the XX stack

was frankly an been NVIDIA the born. has company computing accelerated since it day

our trying we so can – number the company constantly is to we expand applications accelerate. that And of

tracing. graphics we're Of an and course, with it was one computer ray original real-time reinventing

We a brand-new we're progress have application rendering, which making is that in. great

We mentioned just talked genomics Recently, computing. – acceleration. we XG announced I just

more to we're natural progress are And computing. and tremendous in robotics really understanding, animation, to artificial intelligence, language the area complex. to of one the recommendation systems, of so field are In number and the of conversation, broad. recognition them and going to AI making applications future accelerate important that really, from more of is to that image those of the applications are And to each getting new really

the question so of to two And dimensions. comes about company the down really our sustainability

computer are Whether it's science graphics forward. continuing of that expand the of or we path right of continuing physics fact we example. basically One to assume – AI suggest believe to laws laws to a two comes would XG really for that is can for surely applications and the expand? that or dimensions. from computing for And Are forward to Let's number the computer of assume But down is that let's accelerated there's path computing we lot the we is, now evidence accelerate? the the `accelerated or this genomics so. dimension

more to impactful And be the is industry? ecosystem, And -- and the getting are they Those and are then they adopted those second continuing more number complex? by the then the of number applications, dimensions, the a applications, to impact and growth then continue grow, of to those sustain. to those of We're of applications the dynamics the think we're do job. I going going the good those evolution and if complexity applications

when of I and -- of so think basically equation way, And the growth spelled it I out our company. it's that

think ahead. I fairly exciting it's opportunities are that clear the fairly

Operator

comes next line Curtis Your question from Blayne of Barclays. with the

Blayne Curtis

Thanks squeezing the on you in. for me just auto Jensen, to I ask wanted side.

might customers out slowed think at their least I have your of one program.

be out anything Thanks. if has years, helpful? look OEM on challenges you regulatory moving of kind just Just couple as the changed there then curious the of side perspective the next the any slower? And is would

Jensen Huang

the struggling, all but for everybody that knows. industry I think is reasons automotive the of that

to And company car is with this be future capability. now However, to would really to and autonomous vehicle do to and their And a company know to reinvent racing That software-defined. them, they've They and wants autonomous They need redefine for Every do tech platform that. software-defined. and known by so company. single which their autopilot electric get one need the model to surely to been the never the some -- to car they're has single be be a has tech Every vehicle to capabilities an there. Every be is business them greater. one enthusiasm time needs of tech every they is be of car companies. company.

and never is term, is autonomous The automotive The clear. than whether whether has very benefits near AV safety, reduction whether in The utility, real. it's industry of of for in more opinion. future it's so cost is more better struggling important the opportunity although their productivity And has opportunity my been it's been is ever. never

the we're projects on as so, think enthusiastic ahead. autonomous as I'm that moving ever that about And I working vehicles the are and

as And challenges is China I the as sales importance so about experiencing, in they're of or AV slowdown industry the feel the the badly near-term clearheaded that industry whatever But ever. that. automotive about of

Operator

I for any over will back call now to Jensen turn closing remarks. the

Jensen Huang

a NVIDIA had response revenue. AI graphics an is quarter for market's We record driving creative upgrade cycle NVIDIA is community in huge for the serve opportunities and to rendering computer excellent graphics both powerful with data center opening the of workers and gaming strong cloud and markets new and gaming. platforms RTX and demand professional RTX and in independent new NVIDIA social excellent, graphics, while whole content reinventing us creators and

AI Conference rapidly. these next GTC and momentum. gaining Jose. business Internet. key recommenders understanding, more our much from the Our Come a natural way production, changing data San more key center public shifts with of AI lot me. And deep join the interact The is in business is is AI, three demand trends new by wave cloud month's people powered conversational as enjoying be trends about AI in We'll inference for development language growth at a talking to are growing

disappointed. won't everyone. You Thanks be

Operator

concludes call. this gentlemen, Thank participating. and Ladies today's conference you for

disconnect. now may You

NVIDIA (NVDA) 13 Feb 202020 Q4 Earnings call transcript