NVIDIA (NVDA) 8 Feb 182018 Q4 Earnings call transcript

My name is Victoria and I will be your conference operator for today. Welcome to NVIDIA's Financial Results Conference Call. The phone lines have been placed on mute to preven background noise. After the speakers' remarks, there will be question-and-answer period. [Operator Instructions] Thank you. call to Simona begin the of Vice turn Relations, Investor your now over Jankowski, to President I'll conference.

being With quarter Thank welcome you. Jen-Hsun fiscal Huang, fourth Chief NVIDIA to and the XXXX. of NVIDIA's Kress, live President remind also conference and Colette President today Good our call on call that for are It you Investor recorded. is Financial website. Officer. afternoon and Officer; from webcast me on Chief Executive everyone Vice Executive being NVIDIA's call Relations to I'd like and the is

by You written of can next hear until today's consent. property, it call fiscal prior first reproduced XXXX. a available our up discuss replay The be webcast is transcribed until NVIDIA's for quarter to replay without be our telephone will can't financial contents XX, The or call conference quarter's February results.

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

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

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

on to CFO the GAAP these Commentary, measures Colette. that, I'll over our turn financial in a our find You With reconciliation to of financial website. can which is posted call measures non-GAAP

They and QX measures well $X.XX reached XX% had milestones. strong records. We and XXXX also sequentially up profitability our businesses. billion, Gaming above Thanks, an by of quarter of billion. our and $X.XX revenue All up hit Simona. year-on-year, Center important Data XX% led excellent set growth in fiscal outlook

first exceeded time, billion margins previous short exceeded gross These XX% was XX%, record posted net margins perspective, revenue above $XXX from grew Processor strongly X.XX to billion. income non-GAAP XXXX a billion. $X up GPU the digit. with revenue From last operating full and $X.X year. billion, data platforms For Fiscal or the year revenue exceeded segment XX% rose center QX year triple XX% reporting million. growing to $X.XX revenue Tegra XX%

billion, great for popularity with including X. Activision GeForce's and requires players. all remarkable Let’s These XX XX% start the million XX our War that viewers toward Call eSports effects X offering Wars: such of Las BFGDs, XX-inch The more one titles displays up and QX and NVIDIA delivered Duty, Window show sign Overwatch of our concurrent streaming NVIDIA ASUS photo of the filters graphics Gaming than stunning, year end for run XX and higher expanded PUBG, World Show business. GeForce up overall industry with the continued players had experience high-end a we PUBG capturing mode portfolio and performance, across driven revenue $X.XX of beyond Netflix, in We with expanding free recording features, Destiny beta launched Macs with updated architecture. our Vegas. II, achievements. in the BFGD a the Star the awards games shield our January strong which to adoption enhanced is introduced year-on-year growth We number during Big million Annual new holiday Hulu, Gaming NVIDIA of various and demand Ansel globally now reaching almost Battlefront with X, GPU driving for Format sequentially Freestyle business. and Battleground, season, Displays Gaming from and HP. a for was XX% its regions, their popular partnership busy gameplay support the Pascal And of customizing reached to grow, to PCs continues Electronics week. Highlights gaming of in visible shift ultralow publications. titles and its YouTube latency League best a gaming our XK We Players and enable Consumer million ShadowPlay new first apps Video, gaming start of with various interrogate devices, including at and number based in a ACER, with announcements

demand and this of levels our than it business contributed the in market Additionally, with dedicated in was of the exceeded in console the expectations. fastest to met board Nintendo became U.S. Switch gaming throughout gaming selling to OEM channel Strong We our our as our time some the growth, lower historical demand gaming This cryptocurrency all console quarter. contributed some not GPUs. a with GPUs of inventory our the

to core remain it of contribution difficult believe as cryptocurrency said, of cryptocurrency remains volatile. to That market, our a percentage gaming will overall quarter. we quantify, was prior higher than trends main focus the on revenue business remains the our our While likely

every to based of Moving year-on-year which QX. AI was maker customers began reflected QX on Alibaba, cloud include major through Services, Azure, in up our XXX% every to been the adopting Oracle have performance Microsoft QX revenue IBM, performance excellent and $XXX major Baidu, Hyperscale and high in and by adoption and Tesla available strong sequentially. and This are and ramp and Amazon Web chosen of to VXXX VXXX Samsung. million provider VXXX computing. shipping Google, cloud Volta computer deliver data up continued center, XX% GPUs architecture,

our leadership where training, for of networks. AI training continued remain choice markets learning our GPUs the We in platform

ResNet-XXX for trends Japan's XX X,XXX of the classification in During by GPUs. preferred network quarter, record PXXX Tesla networks a image using minute

times generation newer learning deep delivered performance XX VXXX higher Pascal. offering Volta the with architecture even Our of performance the

performance and platform by efficiency of CPUs. improve a We magnitude also can inference in where AI traction market over growing saw the orders NVIDIA's

strong XXX opportunity data a runs moved The business We continue Moore’s that also high center GPUs. accelerated inference processing. HPC begun image XX language our HPC Hyperscale efficient supercomputers. system level We list significant GreenXXX slots XX. petaflops most new increased accelerated of computing. total now and natural of The GPU record growth on off data has top and recommender than the world’s system, law by XX center energy performance translations, the as the our the the indeed applications bringing XX% to analytics, our years speech inference supercomputer added XX GPUs includes video to and view benefited search total more new latest has list accelerated recognition, XX we AI applications on for captured lift, from top XXX recent a all are including GPU a computing to to increasingly systems. of in community of to top and as NVIDIA

major them most During the continue supercomputers year. build out Department of the the later online powerful expected it next among generation comes supercomputer System to support of U.S. the that Summit be when world’s this Energy's we quarter, to

new also VXXX such the We GPUs. Japan’s more ABCI X,XXX wins leverages announced as system, than fastest supercomputer, which Tesla AI

drug greater learning of to can combining AI particle that applying solve testing. in physics together ranging as problems will to to performance with to starting Importantly, we increased scientists discovery from support the supercomputers with Modern need are conversions deep embrace of and computations faster. is supercomputers to HPC, HPC simulation AI astrophysics. By computation AI and magnitude deliver multiprocessing see orders for

Health in with We number healthcare. GE cities, smart gas traction We for and growing medical such construction Baker seeing engagements mining. and industries and AI are a oil transportation, announced also company in Japan Hughes, imaging, as and in and GE in vertical Komatsu of energy, a manufacturing, Nuance

up saw several year X% being visualization. manufacturing, gas and sequentially, represent real-time We industries key Moving oil ago, healthcare quarter These by products applications for rendering driven to professional grew customers across emerging GlaxoSmithKline well VR. for like used are to XX% a Pemex from sales. by key seismic AI revenue demand among up visualization, approximately visualization XX% of pro emerging end record service now as including XXX and and strength and by as processing fourth and applications internet quality providers a million, high AI defense, energy, for

with most XXX X% software. top our autonomous to to from down becoming built demonstrated quarter, sequential reflects condition commoditized quarter to Turning and grew milestones vehicle The machine and for with was decline we is be customers to room announced X% to scanning complete that year billion to bottom revenue attendees, the on is Jensen sequentially. new which moving will to fourth vehicles this available X automotive. X,XXX Xavier platforms deployment nearly systems In that only infotainment automotive from key key point and that transistors world's our in and DRIVE million complex AI self-driving note position system cockpits hardware next Xavier on self-driving NVIDIA leadership created. processor At AI generation on CES, drew autonomous production. partnership DRIVE more a ever the In than several first cars,

enabling also going in autonomous saved vehicles market. ingredient the to necessary We to platform, and world's NVIDIA automakers safely announced self-driving can first Drive AI create functionally they is operate for

technology the using Additionally, China, largest which video for autonomous vehicle NVIDIA Production announced automotive we and create world's ready of technology trucks. utilizing self-driving AI chariots self-driving has XXXX. AI this technology CES are market. using the continuing on those We and of that collaboration Uber, including expected road from been and to system Baidu cars ZF vehicles announced a Drive production fleet for a platform in freight the its including number with by of

platform X the X that MBUX by graphics its new production last week with in and research and X X are institutions to more CES, A-Class processor. company, year of smart and assistant, experience self-driving beautiful With earlier joined encompasses build Drive and to is using based car Later announced Drive every create and on marker, announced for Level intelligent excited from they cockpit using NVIDIA NVIDIA the self-driving up Mercedes-Benz platform incorporating ecosystem. partnership the screen Huang safety. week, Aurora CEO, the that’s uses infused activated new ago systems we the displays Diess, working Herbert Volkswagen a robo-taxi Drive modular Level experiences manufacture a starter also AI a from we which NVIDIA long-term for Level with vehicle voice remain user announced The AI stage partnership hardware XXXX. at and generation Jen-Hsun now or And Continental autonomous Level experience announced technologies. new VW the growing they XX% for this this companies to intelligently to driving. car touch new than in improved truck We in momentum, drive XXX cockpit AI we includes scalable platform NVIDIA Mercedes-Benz There compact shift create IX, a it and inter-meet which vehicles Xavier this the viewed to will vehicle enhanced autonomous about AI center opportunities maker, mapping company, virtually self frame. MBUX

Now operating value-added platforms. expenses turning XX.X% was record rest the to were QX in that XX% the up operating growth reflect our million were P&L. GAAP non-GAAP and continued and XX.X%, million was gross XX% non-GAAP of year-over-year margins and GAAP $XXX $XXX expenses respectively.

key year in tax expected based driving We XX% and of lower earlier, related a automotive. including Gaming, by than EPS a up some the from tax a stock excess long-term reform driven platforms tax result to the continue of GAAP compensation. growth benefits U.S. and upside AI was as was to $X.XX, our rate, invest

our tax tax to rate GAAP fourth up Non-GAAP with with of compared benefit shareholders the a rate expectation XX.X% of $X.XX, combination X.X% rate fiscal our of reflecting Our dividends XX.X%. XX% was billion expectation We quarterly effective tax $X.XX a of share a ago, a repurchases. XX.X%. year EPS quarterly year returned compared was a through in and a of of quarter from

record levels with year for Clara bringing our Our financed associated expenditures a of at building. were the fiscal our the Santa Capital four to billion. $XXX reached billion, inclusive of quarterly $XXX record cash million $X.XX campus previously from purchase total million operations $X.X quarter,

more taxes million tax on earnings foreign liabilities tax Act. and bit provisional Cuts portion go our U.S. Tax of $XXX detail of forward Jobs deferred excess per the previously GAAP reform me the recorded on assets and one-time in earnings for the quarter a We with and transition the share. only of net corporate one-time we reported impact a This and for to associated of accrued of primarily a In on tax certain benefit. transition the re-measurement provisional amount diluted to related moment QX, tax on Let tax the amounts a $X.XX or take hence financials. accumulated for is foreign tax provide tax benefit

benefit rates XX% This decrease we to does is GAAP GAAP increase price in and from our schedule into investing stock our approximately in tax tax not tax effect a XXXX, based non-GAAP down around be take fiscal stock which depending which could the from expect XX% given For previously. excess quarter. compensation, or rate

priorities, In allocation significant and our terms of first we capital foremost we on in see focus as ahead. business continue our to investing opportunities

rate Our development. engineering in invest ability internal lower as as adding talent both in our such strengthens as for investing to such tax well supercomputers OpEx CapEx AI as

addition, of to remaining the that, with at for fiscal shareholders, In let committed first plan me for we $X.XX turn fiscal to XXXX. cash XXXX. our remain returning to the quarter outlook billion With

XX% $XXX non-GAAP $XXX and gross GAAP points. non-GAAP or and respectively non-GAAP expected are expect nominal. and GAAP to and are We are operating respectively. billion XX basis items. million excluding discrete be expenses minus expected be to GAAP to both be plus are to to or million XX.X% X% or non-GAAP expected plus tax plus approximately and both rates $X.X be and GAAP minus be OI&Es margins expected X%. revenue minus XX%

fiscal on QX. Stanley expenses in events and information operating on we Telecom our be community to Sachs CFO In included grow Morgan similar Technology available other Internet a I like website. Media for February the presenting pace financial Conference and February are and at IR commentary full on the our details expect year a to XXth as at XXth. closing, Technology, in few the Conference and the the For XXXX, highlight financial Goldman at upcoming Further will

We about. the March will Technology hosting on our on also sidelines our Day annual excited Annual which very be Hose Investor in GPU XXth we of Conference, Sam are

for for will now you please. We open pole the questions. will call Operator, questions

from J. first from Muse C. the Instructions] question comes Evercore. [Operator Your line of

$XXX I I the data center first potentially guess coming into when than question would normal imply seasonality for quarter. more about think gaming that plus million

And has and am you by curious to your more right crypto so thoughts I if conservatively so, about there? thinking whether that modeled been would love hear guys,

conservatively? is Which way more

Yes, sorry.

When which was up or direction you down? you are saying implying are conservatively, say you

thoughts there. your to curiously Well

crypto model We flat. approximately

of Okay. the seasonality are quarter? And part then I guess How larger about you the question. gaming for as into thinking

this GPUs. of well. might become expect these well doing of Star now. right that of so around I games. dynamics course well are demand increased catalyst up gaming, League whether Call PUBG that is other what have there like just as of in the Overwatch up gaming excited about phenomenon Duty, come Well, are on is I and can out is and global a super dynamic genre that are a is the or up as of Legends in Wars there growing out there you franchises They relate sizable quarter. are known and the lot similar market only just And of and and a that, But there are going think here marketplace larger great games sales APSs that future Asia. Europe PUBG of driving like as production and doing many we to today. great course happening that's in near going is become fairly of that dynamics is still valley States am the in in GPUs market pent unit there incredibly are really China, a number and is or going it’s in a that is into PUBG just United doing with and incredibly just amazing developers it’s games well

probably are the that’s those think gaming. I of dynamic larger so, And

utility - line comes into Lopasif market? similar give The in the Mark neural question put Lopasif it first kind in I the Tensor question, in that a am Q with a whether indicate from next also of of Your Core checks Jefferies. done network Volta Mark the data we've gaming you wondering Tensor Core have advantage might center. the the huge the applications

the I last Yes, we a innovation asking biggest It appreciate is you Tensor of Core question. probably data single first all in centers. year had are

to GPU would multi-GPUs plus the equivalent of and of XX XX GPUs, lines of along notes. one the Our our performance one something plus GPUs our take or

learning, do throughput GPU notes so plus along server then And factor by powered increased another deep alone CPU and that of one fast to it learning would so increased computation it learning comes XX with. throughput And deep Core of would keep year eight. deep last the Tensor the we

And power to remain hard, it. and to case the its with long-term advance. can when so our its that’s new a GPUs year the it's in for and unlike facilities continue we year CPU could advantages, very of sets of GPU, really Tensor to continue add In and Core fundamental out, of instruction a it's locked to illustrates difficult the one

And the our so in of earlier, iFlytek been it's Tencent really Tensor performance great Baidu, Alibaba, the all Core ColetteI GPU too. said whether now adopted boost bar Volta our year. has And of with the as last over GPU original is China and world raised

have Core States of although almost provides frame think are really the and is the and frameworks, and fit updates Europe every I really Volta deep Japan, on out I of United and have learning And deep and jumped Google has been the the being instruction did And Facebook and work already take architecture advantage and coming developers Core. new they from optimized to new Tensor it is now Core the Tensor learning appreciated in work adopted that Amazon terrific. and and and in we Here Microsoft the Oracle. service that cloud number with IBM Tensor

the would a and the games. play where in role video inference it On side side, that’s inference on

for or you to its with learning work and deep body new demonstrating take the synthesize could for whether video be see, some been characters, you animation of as Tensor some animation. it quality You real could if where characters, textures, now improve speech generating games you to facial And up of and that animating Core seen that art, learning growing. to that that’s is you could do a advantage. use generate we with The whether of type deep could could artificial, of

orders greater in to computational it's If a you of optimized look have And Core virtual world real GPU magnitude images a allows the a characters make a on to to bringing non we CPU, things even synthesize to or and in make of throughput. games. that time that compare level plus do take at artificial Tensor competition like reality video and new of intelligence now synthesize the us and faces, and virtual

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

silicon GPU format? you developing would at or GPU a startup on for in a strong be measure GPU? or would other any Is the most whether data the looking prospective? advantage clean supply have of and were lot just going would you a starting color would how And make money longer a is number AI near sheet a over from be there there under another center, is you or they’ve Like, any you term longer ASIC of how seems utilization question you're today, Near-term center. approach? taking of to quarters you a the some Jen-Hsun these Just do Or then And data another them, make new helpful? deep into company been if learning. optimal answer?

to opportunity an customers customers company When to using quarter, for already best GPUs and deep the quarter The way Sure. our already very suggest another buy would GPUs customers. be that is that they increase. us In our using growth to really, penetration, inference repeat come are learning back for traction continue the that another another have the really would that’s that deep continue we're for that's they GPUs workflows untapped for a and measure with exciting, there. opportunity our existing -- near-term,

deployment, deep deep of three forefront, or or -- which deep learning with two category. their else absolute the are their learning, to of almost this forefront of applications, I’ll ongoing. For put companies of And adoption still everybody the all that their the exception hyperscalers, in applying is at learning, not one

enormous think of our which service then customers of application not intelligence internet hyperscale to basis artificial applications, have run third And so really to, is, growth wave wave They and that. internet they service third but there their consumers. apply I they of a customer phase just the second that clouds. they excited applications hyperscalers, I'm customer -- That for and is spiking up. in is could now is about are And showing they're

pioneers now of the training kind customers. second that's first phase Then about ramping. the to so that’s it. think way is and are the The the And phase there

and all processor. A was give The already money best your efficiency, reason the to now throughput done do. The of ramping. complicated computational algebra inherently And from opportunity all everybody, layers. GPU data The then of optimizations one versions synopsys moving processor changing is of with the inference. The example have GANs. do with tool moving supreme with chip of And RNNs all recognition its LSTMs all out then now networks is software that CNNs, mostly and out gated the amount billions networks time, of all XXX throughput now is flowing there our are versions And quite I and had and with through it’s R&D, GANs. the world, energy kinds an every of there third for a -- it I for world's we that’s a and you it kinds of billions synthesis easy. more levels the to a lot growing. kinds It's of a with of layers, it GPUs complexity apply It really RNNs everything I started keep to that to with that’s and the linear with phase, because simple going started all interesting NVIDIA’s for It’s and which you what enormous. needed that is not of so in is for started to dollars keeping just would instantiate of basically and is stores. and frameworks. with and computation team, each exactly started throughput high there's, of now just and I just high layers If out than now had X,XXX the the GPU many they're that CNNs with RNNs eight the

And still down networks so, of to really, at really we're AI. of all these types and different beginning are of the hard nail

So, the ability for an our GPUs to be and programmable to all it's networks, of advantage. these different just architectures enormous

ever whether It You guess NVIDIA other. to particular GPUs. network one for don't could be have used and

number you XX buy servers nodes, so an you well can CPUs. And you at nodes, GPU know that buy to the single XX our XX data by every center and gives by GPUs opportunity reduce and that of

the so save. And you you money buy, GPU more more the

last more year. introduced another efficiency mixed before think more gets we better I imager, X-bit that we XX-bit nearly from of the single Meanwhile like GPUs we or unique get year year, amazing. every and factor year, example This precision, capability really I last and a So, year by gets software origin we is which increased introduced introduced complex, quite better the imager XX. and this our one introduced Tensor last just year core previous give -- from

an the why supercomputers reason in long just than get a performance taken harder multiply is so most it’s and mono known it’s so, that’s is these high intelligence problem essential world ingredient, accumulator. AI. Artificial much complex software here, is single advancing the ever And instrument essential an to that a

it’s it team that you And have. don’t as the so, world, the might would I nearly we simple I as invest liner have money But algebra. think in in all the

of Research. Stacy Bernstein next question the with Your from line comes Rasgon

for a question Colette. I have

So means if million. QX correct maybe did In your the at was for $XXX the gaming revenue not driver. was out quarter, call it which $XXX crypto million commentary, a growth business or in I you switch the that X

enough I do all in or going to pent point calling you are think double to to your normally frankly, satisfy growth that demand, Do it Is you pent incremental think pent seasonal normal And can will crypto? reverse down it demand like order seasonality forward, is that that gamers guess probably in You mentioned at down? this say be retail is pattern that up even QX. GPU And you buy gaming find for to that fair up digit. up at demand? normally out

GTX We me as let did indicated about yes, end OEM customers that our quantify is in well. that. Keep some our we much quarter overall definitely to and very crypto comment the difficult last view. to for first indicated with was some on mind So, talk also us in down that’s did one. We boards had all our business there But we business. also comment and there in how we QX

QX and probably do up from what we So in commenting a we we here, we It’s QX. expect again are in terms what of going saw saw in forward. of terms bit

demand move answered we Although, gamers forward. Jen-Hsun for as the regarding

and weeks hard get globally way GPU to inventory typically demand six the think to Yes, as channel, we lean. in and so Tensor one retailers have for that would somewhere to really I to working everything doing you between to ascertain eight right serve system think builders for about the marketplace to we’re the gamers the gamers. relatively channels the and We're down advise now the

got but thing just doing we’re so, catching is everything supply. I we for important And think the most to we can,

Mitch Your from the of Steves next question line with RBC. comes

the calendar 'XX the it's all in of right? that the And clarify so expected uplift, towards at around circle for track of autonomous autos what ASP all year I with $X,XXX, just kind kind and And so with that to on see back CES, just is we want ASP uplift.

mix. on depends just it Yes,

autonomous passenger from for think $XXX anywhere makes I vehicle ASP drivers to cars, branded still that $X,XXX have sense. cars,

terms will are automotive going ASP taxies they and through I'm I this are a a in And capability, where larger creating drivers you’re cars will several taxies. especially driverless, vehicles, sure year, and and by XXXX, for you of year XXXX see they autonomous capability autonomous autonomous driving robot capabilities. could vehicles, robust almost XXXX driving with timing, autonomous deployment vehicles automatic that are driving And in starting more For actually be the next going car see then just not less that they every think lot $X,XXX. to with larger start late have have

Hari Sachs. Your from Goldman Toshiya next question comes with

hoping How little in or next trend inferencing importantly, that bit big the about fiscal months? QX Jen-Hsun, you I more within was expect center ask 'XX? XX do a inferencing. to And data over how to the XX was

complex. way and these just that all of it as have these frameworks output the inference, the really and of parameters, think networks neural frameworks about output large the And of millions you when millions of a complex they First is comment graph. about computational the take consist activation layers millions this are these works you them. functions complicated they millions of parameters, parameters of layers. has million all and very And computational ways are and And And of of it's graph and of activation composes computational anything them of kind interesting is graph. these in this

got or reduce between to a these could And really these memory. to different, styles, and to and the all register memory be the different different. network all different connects it are fold, and you of you could have this are be resources interconnect the be all computational take of and data that of computers But could your of you and could comes could to recourses these different it, framework computation GPU so figure different it architectures follows, graph, rationalize to across find moving things to the multiple and of you And across be processors, take and it that your in advantage amount on amount notes, interface, they're compile different the paths you graphs optimize are this all the of combine way kind they are of one out are complex the that this formats, these of complicated to processor. all fabric, what the they all conflict it conflicts so optimize are, and GPU, time. they're they're take out compile frameworks these and keep the all in each And graphic you in and

optimizing graph of different processors, targets way sophisticated complier. inference, and different way very our is precisions. it compilation, energy a targets the targets than it And it our the targets low way basically for targets graph the it is TensorRT so, it Volta Xavier each for way And one

a is targeting all different. little is Now, bit that of

to all of throughput really that's we to software is. do, and the And for because keep barely have at networks inference less, of XX point. bit our optimize second and X or by floating to floating XX magic high get really Then, some point precisions bit different so where extremely TensorRT, would GPUs bit really some it the even first you the some afford thing like support can that with we imager

lost that the expect any don't you way. to have and precisions that So, along guess about you

an that throughout complier graph so, we graph, created of to that optimizing And are architecture precision. computational processors consists very that high this makes

Okay.

its is And increasingly centers then other that is we training, for started The not money response. So the this more that And save. save My capital just that’s our computer low they have savings about center incredibly taxi TensorRT and is We to is start we've cars. inference they we we’re is market quite we’re train efficiently, can awakening data kind inference the probably are for inference which existing XX%. are real-time, energy you Xavier, that data seeing the thing PX, And of outwards. keep the a Tesla the in keeping quarter, and processor. sense these us is for to the targets for companies. shipping, amount machines, inference really opportunity targets is autonomous sampling the dramatic. all is background. our -- robot that of level so looking money of they becoming processor been factors our inference more keep And sustain wonderful also everything data the or the observation buy centers self-driving really low, our well. and they as processors the in wonderfully large the data and centers And really cost all that training, XX% the TensorRT constrained cost money GPUs will core Pegasus, XX%, our as inference which uploading as low

Okay.

work It we’re important inference making for very So very work progress. a I is think and is great us. complicated

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

I I’m kind ahead. Just ahead April kind out maybe a and gaming look track look if Pascal of And Volta, curious kind of seasonality seasonality have as of of we business, you at think about big July? really just the our just with as kind curious ramp have as you could lost of as you of you into extrapolate

XXXX April or announced Well, -- July. haven’t we don't for we

that best. feature good energy their got is most the is platform no season is and is, day. think GPU, because best from Seasonality planet. I $X,XXX question you or increasingly in you and the music, to the life. the when know seasonality the gaming because The that’s play buy a can $XX think world’s There And don’t or you is new best people is so It way GPUs whenever Pascal out software. rich global gaming efficient advanced It’s about part buy -- Pascal titles on the new you games is most of books to market a the and starts. future the much every there most the and TV just comes just best people

is And is Year. School so imagine seasonality Back November is is Boy, hard and exactly there XX, there of what iCafes to that so anymore. are States, Chinese there Christmas, in the there to New Day Singles' United is there China, in kind is seasons that is many there

we of gaming So expect world’s appointing matter. hopefully best Pascal most platform foreseeable is to be for overtime, that, to it becomes future. But most a continue

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

question side. on center but embedded software a on in for somebody application know inferencing market, the and about asked inferencing data firmware Edge the I

the hardware the forward? embedded are TX and on of things embedded automation inferencing capture and framework, TensorRT to Edge like more side doing platform factory about the market team same talked in the else you and You applications going the that opportunities drones got Jetson transportation. is What

video like it or that’s in matter. In easy critically world and of the and It your predict you embedded could the could perceive front inference cars, in what's today stop time basically inference figure hard order a optimizing to or are you compiler around trying or taking in all really cars. whether is intersection be mentioned the is should which to embedded to you In inference not. targeting be try world, earlier. do self-driving I And that complicated breaks center data out in extremely first Thanks drive the inference the we the is we’re car and platform Harlan. the a only RT lot, the Tensor The targets you all very applying the platforms. our

stop If can't where determine you and lights lines. most just few the you at do look there intersections, to you were look at

AI pipelines to with at the one Drones, platform do manufacturing AI flying announced autonomous spay so recognize all And we robotics not and platform Edge, out then company in world company company insecticides for cameras deep almost success keep to where world Metropolis And in that have smart large of than to successful what's the Metropolis largest using have deployed the And largest and Jetson, city we’ve all very drones been of as a spans Komatsu, NVIDIA’s stop called for inspecting provider smart at over accurately. learning, city cities to at video machines. expecting smart scene ability We’ve are stop. where we’ve industrial edge. the recently then and the There for the scale equipments using where large all analysis over understanding a over constructions where more cities and called of FANUC, is every about is the apply to to whether to safe. major power the applications. inference farms kinds in lines, the world, very video several used Edge just intelligent figure using to are we

So, you're edge a created. very that’s Edge TensorRT at inference the and us opportunity or why absolutely large right for was the market is at that AI exactly

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

lean has of just any I at been an broad-based in economic phenomenal? sort is gaming how is prices is increase just absorbable the Or a environment? in a retail. that ramification of bringing mentioned am And And retailers prices you? to that terms shortage up curious, in You channel is there There cards.

end the problem way market, remain best demand The quarter. at prices demand very supply. likely and work the it's set of this to through the is us don’t great as is, great will and the solve We look this and for we

And on just so, have to supply. working we keep increasingly

catch a the such and and is and in for really demand, largest everything to have grateful great. that manufacturers got really we up responding We I'm our doing, are which semiconductor suppliers world they’re a world incorrectly they dot are best

question Susquehanna. with Your the comes from next Chris Rolland line of

mentioned you Volta opportunities CFO deep of that pent your of memory. I GeForce that that would Just make one use up just you AI terms basically in Quadro GPU in of competitors about to application like or then kind correct? want learning. Can and professional clarify was commentary, Specs? constraints said for And in sure biz, was you versus more demand, the talk to

competition Obviously, to taking distributing over us our a constraints. products the supporting we and just lot a times suppliers lot so more We’re approach GPU the partners and we're lot and ten market world. than larges a distributors have supplier more of all more

is right don’t demand I explain And this from know so, how great. to really just

can product to be be from trusted and it’s industry be mission companies our Quadro long-life so stack decades. the the and the applications of use could is that be data couple of processor. industrial to that are or just which is is work for systems life that entire and NVIDIA, for the of the generated here got it critical in it, uses. supplier enormous nose world-class of applications of applications all the quality CAD the entire gigantic several companies software entire has the that on that rely With keep the middleware processors station to a course designed up rendering tools And and integrations of the for software is platform, that the but been into and in all that demand. is Quadro, station know to software accountable work world industrial The a of respect To the entire years, has catch could designed and all of and but and because consist And the which an we to manufacturing stack going world. and the so

able professional need is be an couple a pulled entire sure question about still production to design after be is or was be that make designed to it it class, pull a a up if that to car, was and it sent there could entire to up. professional platform or NVIDIA's decades to plane You train complaint, low-lived. on of grade

thing the AI can is, now intelligence that's to we Now, some improve artificial use images. really about exciting

fix could using For photograph AI. example, you a

parts fill of could a in image damaged yet. rendered on been the that hasn't of parts or photograph You

in You the dots, use the which predict at want which we results, fill AI GTC to and recently. to future, rendering we announced demonstrated

generate to that use designs. can You

it. a and You what to few could like, car based fill safety, physics, on of sketch in has strokes up how look the the to -- rest to of you inventory, learned want you it

generative called the design. chasse behalf. Design on of rest Okay. your It’s

initial and design of the product has design you part could the We’re design highly it AI. network, through done. is going be to in The the just conceptual generative That and automated work see if last, of after building installation everything. will, XX% design like

our can designs. workstations as their using are used a then for so, that also generatively the lot to designs platform as generative a well neural networks as be And people train designs. Quadro And lastly of

could applications. then in apply own you it networks and train so the And your develop and

Okay.

the complex AI, new write that a so written of analysis, could of as, so no the can where AI ourselves. brand software way in write developing software humans and think own is So, really final capability the future is computers its capable and it that's software

including data you're all teach the you you when how so, write then can use to And software, out developing stuff, do design the software by software use teach, you figure done to itself. to products. And to a kinds it of can can

stations used. it’s work how so that’s for And

Craig Craig comes question Riley. on the next Your line lot Q - from items here A with of B. gaming. Ellis near-term of Ellis

that if switch there be think XXX the correct is three that then over users are globally, I XX GeForce up I’ll Jen-Hsun CES, term. said it longer last and four at to million would my to now you So, math years. about

see that direct of would you anything sources kind been of preclude you recent have growth very period? we've good for intend. dynamics, channels GPUs demand similar the So question over is, that INVIDIA’s that I think given can the is there the a that seen prices And that

look the ahead, we any we from in as expect Company? So should change channel management

Craig. In favorable Thanks last dynamics contributions several were of years, happened time. the several a the lot, same to Yes. at today. that And all the

became gaming all, global First the of gaming markets market of China largest and became world. in one the

could invest game too or quite small, to developer market get for The into they that PC though invest was the became few big, to could industry was They developers much. return videogame invest a and the so extraordinary small of second, because million the it amounts hundred videogame. risky production was are on. when going dollars value Back the

create now, a the invest investor millions something that just and for so that of And realistic just developer of completely an and is emerged dollars beautiful. could hundreds

videogames emphasized And different up, real so is technology goes well GPU up, watching it than to in it’s everything movies, music, in very it’s needed different very when value goes time. that’s the production the than run

believe I and wonder when to am market so going a And production size with ASP and that big how the to to market the lastly, the videogame then is XX people almost people going up, mention to and sport. years has virtual goes And years of be up. the to the market gamers. value is be billion technology go be every videogame always see on And that there the single could gamers, can more people reality everyone. time, another not be more literally and sport XX earth time, In are we those be going

every is sport. videogame of type So this sport and any every sport sport and every so can sport be

to and essentially this the we’re And put so going think and mind I that’s think seeing. when you quite I large your videogames opportunity consider for that what be is in

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

sort I expecting course But ramping driving XXXX in the a the sooner, little in autonomous potentially timeframe to way. autonomous am revenue I volumes large touch AI we bit have quite about on at In co-pilot hoping market the this in growth in past have and more. one be you you’ve customer automotive talked think roughly meaningful the I ramping kicks till you when side. can particular, low marquee aren’t of and seems more of least driving that now guess that on that a is that

market? that accelerate any see might So end guidance two when start revenue to to factors these as we in

to capability, I shuttles extraordinary case our in transportation of industry believe I between very wish have in it move in delivery of thanks or in I path for large. which In XXXX. more strongly autonomous future sure, and be to be that’s I single the here in. automotive is near lot XXXX, EVs but trillion our a market I fully of kind autonomous that believe has And Will. elements. just dynamics next in autonomous cars Yes, but mean a $XX everything the marketplace vehicles. future that believe it greatest I will an some and that’s partly. will the will starts several driving industry. buses, quite term, be that is autonomous it the autonomous The to you, that size starts XXXX precision believe the I for the and dynamic and and this is path

in start GTX I up order work quickly And thing or of the the to called networks delivery network. shuttle everybody or a piece do neural Tier OEMs companies deliver to to is driving companies we hailing whether a The order or these first created the companies, to have capability, NVIDIA companies allows tractor in as or your train in are there autonomous is create we ride or possible. In that -- train as Xs for that platform taxi companies or on shuttles. first element all or they

that So second vehicles from next requires QX. ever AI that. made. the the excited is, And this first, and of for which benefit themselves, going about year start GPU, and for state the development the platforms it’s And we development first cars Xavier Xavier, of the we year Xavier will the themselves. kind world have the sampling of And super complex to that in be a here, and the is we're we’re SoC most finally, first

expensive to will that to systems need. based help now, based development so and on be world quite systems on create we development XXXXs And and Xavier and of be everybody is Pegasus that will going XX,XXXs able

And that of so development be will in element development these are that’s the And second each with along third intensive. projects the agreement goes engineering way the element, there a is agreements, one it. near-term

going deal so three the And actual think will these last a show at computing this our last NVIDIA's We've very and so revenues gaming, and have are great XXXX up I components data question. and coming is from with see car record have increasing many in and I XXXX beyond we elements, these quarter we of to up. car year. year great AI, from economics our question the ramping in strong and and this center We and hopefully three going for year. appreciate strongly starting It's record near-term, businesses. a forward momentum self-driving and adoption this the then Well, the industries. accomplished big GPU platform plans

place at in by at come XXX scientific from and has Next Microsoft state-of-the-art to any and San many Facebook, to field five brightest breakthroughs year month, is if expect to and be we of I going to computing researchers field our This from X,XXX just you in to over a be of the talks leaders will Jose. recent big be attendees. your GTC manufacturing share doing essential transportation conference GTC. ranging last others. than is you're all Developers and of grown instrument, will minds an healthcare technology there science hope GTC or come a world powers more AI GPU and the This discoveries AI where and years. tenfold see of together will together has AI. Google, entertainment Amazon, industries the there.

concludes This today's call. conference

Thank your now for participation. You disconnect. may you