NVIDIA (NVDA) 16 Nov 182018 Q3 Earnings call transcript

Good afternoon. My name is Kelsey, and I'm your conference operator for today. Welcome to NVIDIA's financial results conference call. [Operator Instructions]. I'll now turn the call over to Simona Jankowski, Vice President of Investor Relations, to begin your conference.

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

fourth hear The conference available the be until transcribed the of webcast call consent. results written a discuss quarter for is by financial It prior can't until telephone The today's replay can be XXXX. will You November reproduced XXXX. our to content or XX, fiscal call our property. replay without for of NVIDIA's

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

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

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

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

center, X growing all up record QX automotive a revenue earlier, from double professional of with all and market Simona. $X.XX hit our billion, platforms levels. visualization XX% Thanks, reached digits. Data year

elevated short crypto to expected the took normalize. expectations than to following of than were sharp card Gaming channel However, expected inventory prices, falloff, crypto longer post which was took gaming sell through. longer as

QX, normal build we gaming chips consoles Pascal allow of sales inventory also midrange In Tegra for to [segment] little expect shipment for reflects channel seasonal outlook the to cycle. QX to [ph] normalize. minimal Our very game in the due

near-term change does not channel situation long-term a presents While inventory fundamentals. headwind, it our

chips position announcement. growth toward Our we market software year-on-year high-performance gross excited and as have about We recent with of from million. XX $XXX strong self-driving expanded in million higher-value for addressable but grew Tegra a XXX reporting included points. crypto a basis and remain the down reflects and was processor sharp computing, points shift prior $X.XX exceeded Turing ever, GPU ray-traced a gaming, revenue GAAP non-GAAP From ago to non-GAAP $X our year basis cars. grew the competitive components and our gross XX% our and platforms opportunities AI also mining GAAP continued and falloff quarter. fourth $XX consecutive to demand. revenue architecture margins following rose X% perspective, income rendering, the billion. Both margins This million charge is as segment for net

graphics. of continue and Pascal, In technology growth ever the Revenue our driven Turing series, jumps by delivered strong notebook declines. gaming to our new and architecture. mid-September, was more X% sales, down generational first with predecessor, gaming offset of billion sales our GPUs sequentially. gaming console biggest GPUs. $X.XX we as are Year-on-year year-on-year of initial up Let's we GPUs delivers based performance well Turing was gaming its Turing-based more than as These RTX business. which XX% the have Xx GeForce shipping on RTX Xx ray-traced the began up in for

the performance frames is by HDR the technology XK per this architecture. advanced a two release for lifelike to as With baseline anticipated first milestone game-changing reaching and GeForce The RTX followed reflections This XXXX way were the pipeline quickly of the displays AAA XX to gaming XXXX, supporting with RTX the on delivering upcoming the even first gaming to on GeForce These titles, week rollout is features, on the RTX points. highly two ray-tracing, new XXD. V brings shelves RTX RTX high-end never NVIDIA well of affordable GPUs. cards are Battlefield gamers. games the second cards most launched major establishing and games as price before. were XK life hit The NVIDIA now RTX quickly enabling Ti becoming its of like itself a GeForce a

to cryptocurrency largely wave the high-end of have normalize. ahead the RTX. through Although ended, cards has taken expected sold Pascal than longer has channel

both Pascal choice However, of and as in holidays, levels Pascal cards, to work channel the for the down is the expect X. expected. than higher on we remained positioned over next midrange the midrange inventories GPU inventory gaming quarter channel and prices or well

remains Tesla strong including the products, We to the business both with another during strong TX revenue Cloud X% and GPU million, from Moving up architecture and year-on-year had quarter. VXXX to DGX center. systems, our Demand the inference of quarter data launch up Turing continued $XXX for benefiting sequentially. XX% and grow, of the

is after X and TX Cloud adoption including Platform, TX the already learning learning has cloud on graphics. and workloads, delivers availability. of months inference high-performance its launch, received deep fastest learning, machine integrated performance It world deep XX designs GPU. The record and accelerates data any for the training inference, computing, server Just and diverse the into analytics is first server its cloud it Google

NVIDIA deployment scale in an can announced Hyperscale Platform. of NVIDIA centers. capabilities updated software TensorRT the and This stack two data includes at hyperscale also We platform software that new important inference TensorRT drive critically

at frameworks center and time. data on the significant multiple same drive can to savings. higher First, multiple the This models directly to enables same run translating it utilization, GPU

XXx the is is poised data With delivers every server NVIDIA the faster hyperscale launch, platform targeting the Second, TensorRT TX Completing CPUs. it peak take inference data And market, our to performance new integrates inference with delivers ideally layer GPU it orchestration TX XXx in centers. for predecessor. center performance with data centers. our software platform, of for in data this Kubernetes, centers. inference hyperscale stack, told, hyperscale node the the leading suited the All its inference

its Another opens RTX product, incorporates important Server images and computer-generated workload. such designs, that GPUs to of the was Turing look new RTX photoreal industries, this which Turing-based instrumental capabilities, a prior not Turing NVIDIA reference retail, XXXX. the architecture, With able to entertainment, up ray-tracing to for creation market launch the real. quarter architecture. address or to large X media product to Rendering is this rendering as and manufacturing were and GPOs, Yet

point -- RTX-accelerated An the been compared render NVIDIA's and Adobe, major of Dassault is many done on such platform equivalent this CPU has space performance ANSYS, cost, others. has X/XX power. farm Autodesk, with at the as So X/XX most rendering and CPUs. including the developers garnered X/X industry farm from support, up an RTX to key render

a Oracle, Enterprise, with can market, as Q-Flow SAP, NVIDIA learning, for use for as one largest data have performance up amounts analytics and Lastly, is NVIDIA the GPU-accelerated machine an Microsoft been platform With soon, a now, to Azure make Packard data acceleration Up machine faster been Hewlett behavior. as Dell of the accelerate done accurate and, announced business billion predicting enables forecasting with address predictions architecture learning including big inventory not GPU $XX data fraud, XXx used understanding and speed. has high-performance RAPIDS, computing learning. extract source science IBM, well NVIDIA open and including machine analytics a deep opens The libraries Google, to unprecedented from data learning computing, analyze received market at to machine learning, and machine GPUs, and source redefined it's retail EMC, visualization. up machine accelerated. and all data to for we ecosystem, open massive unified and learning GPUs for GPUs high-performance which workloads than server community. CPUs. called support, suite have can and credit industry learning scientific applications as of deep launch now enterprise companies RAPIDS, analytics RAPIDS from the data analytics, such to use wide computing card cases, customer learning Virtually range of machine With broad insight buying data learning, RAPIDS until

set in hyperscale deployments and accelerate broad ranging GPUs stack software out supercomputers centers. a diverse workloads, from of scale-up and Our data scale software to in

last Just number the year, week, and at on from this display the of including systems was world. in supercomputer list where earlier TOPXXX the this in #X capability systems GPUs #X Supercomputing XX% jumped Dallas, the using NVIDIA on Conference

manufacturing in sequentially. as the and industries, major the a on At Revenue up up and products this Moving a data XXXX, reached construction. developers, mobile shipping XX AI under key series scenes $XXX August, and in extended and artists priced Quadro has customers prior by RTX RTX, and record XXXX And to The positive. the photorealistic the already well Turing on the public Strength the GPU designers and game professional revolutionize of XX% earlier year including to most range architecture. the engineering with very GPUs started XXXX sector, versus architecture, These engaged million pro several the work leveraging and X% them the RTX will in reaction render of high-end recognized desktop center. in across and $X,XXX. million, as the and we enabling our XXXX, studios workflows. Quadro server-grade, announced QX, We movie with including based been we Quadro introduced GPUs RTX SIGGRAPH in advanced real-time visualization. week, their

AI-based AGX early million, monitoring Finally, sequentially. vehicle XXX-degree will slated Cars. up assisted Cars turning first in release production Europe, in X+ announced NVIDIA's Volvo driving growth next-generation, a integrating for X we that $XXX infotainment the for year This of features, DRIVE perception mass-market reached design to reflects cockpit to -- and system. the driver and systems. from Automotive autonomous Level The This a and win. next-generation our Level deliver in Xavier surround next-generation selected engagement addition initial development ramp up our Volvo GTC car automotive. is XXXXs XX% production sales QX At ago X%

global suppliers, In for automotive Continental selected that addition NVIDIA to have driving DRIVE AGX Veoneer, and announced they autonomous systems. Volvo, their

including Lastly, and data is autonomous first for it autonomous With visualization. into AGX this platform, started collection, customers in-vehicle their quarter. perception, solutions XXX-degree this software kit developed our then and driving test production. advanced the DRIVE driving, NVIDIA autonomous platform, world's in development driver move This our run and surround can monitoring have DRIVE driving easily shipping and Xavier

beyond. We are next and opportunity we excited about as year the look AV into

expenses operating to our prior for and charge outlook architecture sheet. $XXX were QX falloff demand. XX.X% and sharp the expenses P&L of were following XX%, was rest year-on-year. was $XXX and $XX due components the non-GAAP in the million, up the crypto million margins the GAAP gross Moving operating and million below non-GAAP balance XX% to chips

tax of We revenue was net U.S. continue driving growth expansion and and earlier. discrete was XX% key reform well $X.XX up income EPS transition was center lower gaming, million benefited GAAP and tax the Accounts reduction $X.XX, GAAP net as billion, year amount $X.XX a from $X.XX gross other was margin Non-GAAP growth, to our our expense. the was receivable in well items. reflecting income $XXX tax invest as a long-term tax and as EPS platforms income as year income billion from up in Inventory billion ago, $X.XX, from net latter part XX% $X.XX production including Turing fiscal of quarter data $X.XX automotive. the $X.XX we reflecting the the $X.XX Turing's was operations This quarterly quarter began billion our $XXX the in of year, $X.XX quarter, the end through end billion Capital were shareholders in QX. expenditures returned the at billion in XXXX. as Quarterly dividend of in from million. compared of announced QX effective products. a $X.XX prior flow to of to to billion have RTX $XXX compared quarter. cash We've the to prior increase in million. in fiscal the was ramp shipments

$X additional by for $X pleased that, to billion and an XXXX. let turn repurchase to With of also the the are share intend authorization to to we return increase that fourth billion our fiscal fiscal of an the quarter shareholders to of end We me XXXX. announce outlook

outlook is expected revenue As Pascal of gaming channel. our noted card inventory the the work-down by midrange in impacted earlier,

the our seasonal our revenue is a typically expect mind in build business. about we console addition, that patterns. portfolio X/X gaming in In given Keep midrange desktop of gaming decline

approach weeks inventory increases that compared channel that assumes and and outlook QX Our QX. levels QX with demand exiting in normal gaming

expected or will to community respectively. Now be available commentary community. and X, operating and Technology, on gross We'll Barclays be to financial be closing, for some our income XX. plus GAAP minus non-GAAP the in at tax we will are questions? we in expect $XXX quarter the call like X%, Global of total, respectively, XXXX Consumer and call XX or events be for results for the are the other on discrete at expected next to Operator, and Media and the to Further will $X.X be information questions. GAAP million IR now both plus to and expenditures in place GAAP $XXX will poll January Conference for open financial expected GAAP be XX.X% Telecommunications are X%, we approximately Electronics on minus that, XX.X%, Capital $XXX to discuss OI&E revenue Show financial In basis million, billion, included and Vegas $XXX Las rates fourth And through plus XX. our fiscal both our million. $XX non-GAAP meeting non-GAAP website. to be earnings December are highlight please take and of and million. be the February financial I'd margins you With items. approximately expected CFO to presenting minus X%. with to expenses the from upcoming expected are non-GAAP excluding or details points. are million X

[Operator Instructions].

is Harlan question first with Your from JPMorgan. Sur

spending. the to near the slowdown for your anticipating is customers growth But from the actually Tesla-based the your concern Within your cloud seeing flip sequential hearing continued team the data January some on products. guidance your demand, for is side, that near-term quarter, business? some team There of pretty in around we're strong center seems NVIDIA be in China cloud term,

wanted So January dynamics into on just views trajectory quarter. and to the I get the cloud your data center

And the well do and continues demand of grow more if at expect of Yes, centers. and more that to happening, The Law ended. we while to continue data what's for you Moore's computing -- the getting data increase background we in only look know each has more one way to much computational which CPUs intensive, is is of CPUs computers, buy computational center resource more algorithms, those the is faster. running computational learning machine aren't really and demand or buy because to to

as alternative go we the adoption is so growing are of so is accelerated reasons And that I that we provide a to we is data offer quarter for and The done the beyond because center allows CapEx fast, path computing this that of is really would which forward the one Law. result NVIDIA's the up. approach that, think a platform why the Moore's have alternative, fantastic. There have several things is last of that

to whether it's although of the platform expertise, fair platform. where community lot quite deep financial artificial of you or is over today, That know machine entire starts health as areas we algorithms source that very brand-new as for new train that top of excited worked a never open all learning using is have the there's new well care, to amount a very really computing now a of that of data-intensive. and industry the platform, first -- a method successful. services have the that a That's intelligence machine parallel data an learning accelerated the accelerated years has And excitement learning, in platform, The logistics, that. and been of accelerated RAPIDS. of model computing with run incredible. traditional expertise, deep vast pull majority and introduction Dask a they amount from that the high. is several are then and an our whether the around about super is called it's the of Apache engine, quite in of together is CUDA The before. in there's fuse domain to retail stack where on domain all last And approach distributed Arrow, our And the quite We there's course or together learning

on much, third, And allows allowed. Cloud is have Turing-based film do thing affordable to just of rendering along engine, stack the it, than first recently, accelerator, Law the Turing would is GPU. together TRT then put that TX second excitement with way the created has announced we've second-generation Moore's And AI Kubernetes, all architecture much at software our our top that Docker, lot center. of The Tensor we Core, us the inference of data which the of all a more in a

-- data to well. to our business do quite to center be continue expecting I'm So

Great. your team And seems think it lineup for down early well, continue to term, Turing-based products the is these you kind strong. October, I AAA-rated be growth quite games. motivator the be part pretty rolling is, midrange eSports Pascal January big in what that to a be started products? actually strong working of drive sequential then of near the cards, high-end on just families And just demand for as your obviously do team appears demand out the know just gamers. to the of enthusiast-class Turing to quarter but I in pull product anticipate strong RTX

all is it V the content in -- great than GPU first the coming. support world's ray-tracing. it Battlefield the it's to yesterday, application I it Yes, or real-time It end price It's the generational the happened every the include GPU Turing first real-time performance to biggest launch towards single we've quarter, the GPU it's artificial At with was, leap the the are intelligence. is ever ray-tracing. I and Today, and gaming -- performance substantially think real-time ray-tracing, world. last point to released had. the introduced And it think of higher highest serves, the was

Turing really So I -- do we well. to expect

Turing go we'll mainstream. bring on, the deeper we surely, As into

to bring as have but don't want announce we to usual, many as gamers brand-new as anything architecture a to so possible. we today, And

Hari question Goldman next is Your from Toshiya Sachs. with

your inventory. question of console the had $XXX October. pretty I in the much to coming effectively to to quarter? the business, did the I you relates How January? about that it in as a gaming down with gaming. And game in January outlook to is X/X Again, October into that estimating million and on And you close typically, midrange is think are around assuming related Colette, quarter? it channel X mentioned $XXX million that, hard I'm you something was

Yes.

of because have terms changed launches traditionally of of commented about quarter-to-quarter we the large in mind, keep the the time of what I products So period in fairly a overall midrange seen over that. size have

in has let's over, again -- X/X it time. and So of look QX, we And that been as of good and seeing consistently the when we're about the we console at QX into ahead think to an What that's period quarters. the in build terms is of hundreds more seen for number move use about our slows guidance as in we say, they X-quarter overall holidays. From in perspective, down we QX. you an about a sometimes have the in normal that than holidays a of prior where we're middle just

follow-up, saw segment? about you've XXXX And opportunity nice Jen-Hsun, as that in then the quarter. you the acceleration Great. and your business extensively a are talked into your think expectations for there. growth going What RTX this I long-term in ProVis

appears has been way. deal great most show be every of to rendering until rendering do The there's created not grow. in now accelerated rendered are photographs expect an now, video, ever in digitally I in more movie, photographs it's that software. created. digital to more the And platform, every TV every to a it. what is content possible catalog, ProVis ProVis -- -- being Almost And every

great. the that fundamental world's accelerated change. the is earlier, workstation a -- expect GPU, been market content to that really, from industry that mentioning the had and surely was the is architecture RTX really I has and close ray-tracing creation decade I digital as So it's first enthusiasm platform a not

first RTX And so the is one.

really expecting well. So I'm do to ProVis

with question Lynch. Your is from Arya Merrill Vivek next

given be inventory? it people hasn't. out within cleared needs Is that what the the were and maybe quarters should that could -- Jen-Hsun, of quarter, down this pause? or down Do then part you that kind Turing this April curious, we Because in come maybe work of part how Is some inventory? question, else? waiting B of to quarter was work midrange I'm the happen it something think that you seasonality could created think to about it it Pascal Colette, take October to X but pricing? thinking mentioned And as X to for this

QX hangover. into the post Yes, channel well, came with inventory we excess crypto

the with for declining, pricing grow, the and to We we to it -- expecting while and elastic pricing volume demand sales grow volume decline to we decline were expected be be but expected and marketplace slower to was in than -- to for pricing.

it its At and expected, has I pricing the levels. and longer the the most took point, slightly come than took increase think we volume longer -- expected. just of down we longer -- this prelaunch than took to pricing its than below

is is stabilized, so And not when and get through in upcoming the to for now channel, the hoping our back freezes has the #X sell people at probably we now and decided return card opportunity market pricing inventory we that the stabilize, for channel waiting it to it's is and selling took frankly. longer to the the channel customers it -- the I'm expected that graphics that that to I'm pricing the buy. will in guess right I hopeful world, that to an levels, than quarter But anymore prices that into normal. that some volatile come expectation will has. XXXX the sell

our come it, keep that but to back our by will normal expectation eyes on is of the end levels the we'll inventory quarter. so And

we're going inventory going what channel as terms see end to end probably we the from Vivek, will terms QX QX. your to QX, also terms we the do into question at we normalizes where of regarding in in QX, believe the be Okay. expectation, answer in of of up of

QX. go be normal up that seasonality So and we We we QX expect into between do as to follow QX. won't

question a Evercore. have C.J. from with You Muse

question like channel partners the quarter here. terms as roughly down of follow-up into it's current architecture what ahead side. with the are of drawdown in of drawing looks then hearing from a part the of does that that, one. are inventory launch a sound excess? that on Turing $XXX what's two, does kind inventory million in curious, you And Number Number channel right? And you It guess, just mainstream? fit I your

The last yes. I think -- question, the I'm understand. not to answer first The is sure your question I

You The what framed it It's -- is nicely. last question answer was again? the yes.

regarding midrange. taken The Is Turing products last question were was our there any statement that about account? future into

announced Yes, serves but it Turing that would product, future create to parts a be GPU marketplace. of us mainstream expected we haven't our Turing for the the

that. And but it so conventional would be anything, us do of we're not to announcing

was guess, I quarter. levels drawing in the current below normalized perhaps inventory question were down you the whether

the QX, to the by end shipping get so channel channel segment an normalized really we're inventory work. would see then not midrange has. to we Pascal sell back through like get the we into the get doing of opportunity of and No, to that our that product we'll And

Your next question Research. is from Stacy with Bernstein Rasgon

first is I explicit. looking pieces wanted My QX? and for non-gaming, in Colette. of the about If aggregate split your into more into for one grow think to little are you a sequentially non-gaming I to be gaming just sort business

Stacy, yes.

rest we grow believe I of aggregate, do to that In the yes, sequentially. the think will the yes. business answer is

down gaming that the kind Got it. fits I the XX%-plus. you're of implies with that talking guess X/X because

your magnitude point. thinking this a of that about at So what that you're is kind

correct. is That

it. Got

the the Or yet to last maybe felt could realized what calls the on you for handle yet several like it tone quarters, question, little the you go needed and you you wasn't really getting what you knowing the several a guys seems over For channel this the worry. the better a give feeling, idea few happened? good you. quarters, changed the happened last from necessarily into maybe full been that kind that wanted like just my I we've Like us quarter the that quarter you corrected? the channel, high on exactly that case. earnings you've not and to were inventories that is And call, Can this saw something saying, -- in heard feeling last be did a Because late new a had for like quarter? that like what it be this is a Was it? when in the different and of from second

Let's down, as when else. it crypto I we're down, than did but to started to surprised and did. surprised, down we more mean, than were and obviously. expected hangover more The longer see. and -- we it much by -- We it, lasted expected Prices quickly drift it demand come much quickly we anybody than expected. come went to up

demand And its so uncertain I uncertain think and crypto, was the People the be channel about wanted were protect to price. about when price will stabilized.

down to a little we expected. now And levels. normal below Pricing so all the prelaunch longer of that uncertainty, froze is market than I think,

come hopeful am that we're the through the seeing I we're And happen see that. sell-through back to so And will going demand and holidays.

the and it. didn't just quarter. other we And Was question? is we magnitude What that of I another we so realize there the until was the the of it that that's that, -- the expect either -- question? one was end first it towards didn't think

question next is Moore Your Joe with from Stanley. Morgan

those do Turing to Like announced of a games year-end. I bit stand will a already is have regards you many ramp, how that With where support guess, expectations? support? that your how DLSS supporting going by And now. like with DLSS availability you've know of that I better relative technology? seems are number It is to quite

is in the Real-time seen Yes, a most new the had. had And It's enjoying seen for never this about has This going everybody Battlefield before. are V, time. time. think the great, the is very something ever long decade. very, substantial And technology first real-time is people today images we've that with I that ray-tracing for is ramp beautiful. and computer the biggest dreamed leap generational graphics ray-tracing are a

So great. the ramp is

demand ramped on course, than different more out. are a Tis GPU and quarter are much of the XXXX the any the towards the nonetheless, fantastic. Turing Of end -- the situation past. of largely sold into into products The high-end But

I everywhere. think sold out still it's

that And so I'm demand it is expecting think a new just the I fantastic generation. great. be to

FANTASY you In terms first of working FINAL one. And them. XX of saw that. out. is We're the on have also a hard the content, pipeline, we about

they these which RTX But point graphics when is same thing, will is price enabled. when planet. card so at I RTX be on And last the content released, will performing games one get say than higher aside, any the --

price every point, at so it the card. single every -- single And graphics graphics performing is highest

performance and unambiguously in is of world and coming. then course, new So the all will be GPU great features of the it these highest --

New Your with from is next Pierre Street Research. question Ferragu

still to ship channel if than your like drawdown my weeks inventory because more of is around than quarter, put XX the more cards I'm a midrange quarter, means trying a the today full inventory for don't of magnitude head this that you sales.

be, the first I sales bringing I have my you would these thinking would that? inventories? be business quarter of business XXXX Turing QX. Are down And are more expected full like, my inventory, in of excluding the flattish the you card, so ramping inventories, question the rest in Ti XXXX would sequentially. the GPU looks -- the in high-end to while available right of a see, gaming am And the question like then like, the is and rest second And it

what trying figure out the first the I'm question of part to was.

to midrange had take it's the of if than to more weeks down. XX going whether bring a of inventory, was Pascal question it than more quarter first The

channel of between the XX other the than weeks and inventory us more think has brand. I

is is that brands to One estimate the later. X much inventory of is the things have. our other And quarter how hard month

of whatever ability also the to it take, The see we just enough. is us. month estimate but other we our just their is it, brands. the inventory try much obviously, is And action whatever quarter. we of end not so after brand's amount best our in the channel estimate inventory X the to other well harder. We And didn't see It's

there's And -- question to XX so to answer through I this about about weeks of point. is sell the yes, from your perspective, think at inventory our our

with Mark from is question Jefferies. next Your Lipacis

is And had impacting the product the wondering you that or mean, main else delta as the hangover? product leap, has Turing impact do I I'm and the an was transition Pascal. to crypto think? transition, transition contrast I hoping larger is the the you you've well. described something generational -- as could only the greatest difference there you Turing this if Is to cycle transition to to

Turing performance at single highest is GPU price point. every the

role so we made about as It's played all And no transition. it in hangover. this the This its experience is transition. new crypto

great. back-end If ramped -- a loaded, great of launch. quarter, a we you basis had at as fantastic, Everybody Turing the great. excitement expected. the the performance on It Turing, expected, We the is of was great at end we the is as just did ramp was look And and it and the it on job.

think not Turing's this the we sell we're as just midrange we success. And back we channel is the towards segment of And get ramp thing inventory through the Pascals. And the see then a best business. we'll into this came there's ship choppy, was end let out big to quarter, and this the amount about. fair of a marketplace talking to And so do -- really as any into more the coming a didn't was quarter's to products of I It's that looked the Turing underneath time, because conclusion until quarter, quarter. to that

of the and I final had unanticipated unexpected, knew we it we pricing seen think after it a I than made brand's different. expected. can't prices for we little lot marketplace, the analysis other of the would worse took inventory a and compounded came kind have expected done in kick a decline wish earlier, But this it what the to than I And surely And the those in sure factors even setback, so and exactly up. between longer in -- slow the volume is channel down, we

You have from with Rakers Fargo. a Wells question Aaron

ask piece billion look level we're billion. about XXXX? Maybe question. bit gaming going at of then the as guess, burn? do gaming billion, can revenue to the build you've a fiscal bring I think several back Do level? And on say growth business. is, to inventory to a the that back I return How year-over-year $X.X $X.X like a $X.X since If October look over were the you quarter. look roughly the past running quick let's do that, half We the at the $X.X differently little 'XX quarters, I Prior billion now question been start business question that off into to back of of we in second of we that level. the I we

guide XX% the how business this I'm curious, sequential is the seasonal just had center the your that factoring data current growth QX, in growth of sequentially. strong really in quarter? the Over you've past business, in into few years, piece about sequential in

Okay.

Let one take first. me second the

market initiatives doing XX world. XX dynamics a of energy-efficient We're path we top the of is systems systems, Take mean, announced business -- forward. think, at We're world, the in has spot a it Europe, is center into with three part of then Law out intelligence was, year-over-year. this computing I And know, coming broader which is computing and I on. is. growth to to system the accelerated Japan. which been year as nearly it's XX% States this and high-performance an look new of is great. a expand most accelerated now And the TOPXXX, data end, learning, today, of Moore's the in Our the the part #X largest machine of that's we computers XXX in the us the in the going number quarter, fundamental in not artificial with United on

second the is we an rendering do enthusiasm days. the And their is rendering then to time the XX the third literally TX, it around of first to that production photorealistic went that just was adoption. in put we It's incredible. And fulfillment, for call a the it cloud ability brand-new the that The is GPU, time speed incredible from to cloud Google just time.

And TX so to do I expect well. quite

center dynamics are quite great. I really data So our business think

think your forecast, In gaming. question turns just the it are But how dynamics terms we'll I fundamental to Back about out. see great. of

back setback -- maybe bumped year where to years, up look we stage, And to channel that here business, expect? where in were will you've come $X.X statement the at gaming a the inventory, out if through we billion. about out overall type the you get come have billion At when $X.X came what you this the you regarding about So overall after somewhere X gaming of in mid the of could of that? growth you

Yes, reflect price, done had that fundamentally market all to But factor job a I'm experiencing. going to know. to would is. better nothing just we're have expected. to Cryptocurrency come took those. just the came down, that let the we than the just take is but different hoped we we thought extraordinary me do it cryptocurrency down we about we And that prices what we you And it managing demand let longer dynamics. internalize start the the modeling, guys an declining There's and when started that's the gaming

In take if terms the of look by which affected the XX% at dynamics, marketplace, gaming, some our notebook you gaming grew China. in a not is of crypto, year-over-year

And gaming to going quite graphics. seems computer unquestionably robust. market so is gaming RTX the redefine

the that good. think I are dynamics so And

inventories. work to the through have We channel

to shipping build channel holiday well as And course, any they we console had a decision out before products the as the quarter seasonal into of more This plans. season. the -- of quarter, normal tend build seasonal simultaneous not midrange

X effects. the or nothing business But these fundamentally gaming So marketplace that simultaneous see you different. there's the that's have we about gaming

terms several talked that kind add the is about gaming continuing. To of to Yes. our pieces that of we in business that, tremendous also in strength think about

of terms success console terms notebooks our growing also our business overall our are gaming extremely is extremely healthy Turing, and strong, as for well. in of In

of So but at of overall, the as quite we can the you to inventory gaming well. have growing also all components, think midrange, a different piece about channel just see, is

Raymond James. question next Your is from Chris with Caso

from what for don't know And after valid regard Volta what there forward. expect launches to of QX Is going the launches. and perhaps can to a in both how happened expect and ramp the to from we might in compare the new what we I for comparison terms magnitude it's inferencing goes? after inferencing there Quadro A question with training if that any comparison

it people years announced quickly order X on saw And of compiler inference is called cloud hand, that build. is in our that that to of complicated. top. -- TensorRT, is that, so This to multiple their on an Google one of X, to customers The on ramp top talked The the optimizing the get model things what in Google's of I've of server fifth-generation many people models inference about. the scale optimizing that to requires ASICs being cloud. simple compiler in past completely TRT top one as took is inference appears to the be we vast as platform. the porting think really to recently just on The us then complexity there about the a of allows server, because built the talked model to actually super is That build majority something inference to TX run Kubernetes complicated related of of write. also software And that top in on piece are in

And engaged -- on of companies TX. world we're so most the around heavy their now together. port together workload with top the to top pieces come models of technology or Internet that we're applications the And putting have on

of that. decision comes that their we're it so down how when would that about buy, to tells And happens, working like our they And and us hard to on adoption many rate.

an And most high-level step is we unquestionably at the their It I high-level model It processor than in this center. And can image really the we that this: watts, a at we the as come any it look and data center look -- also inferencing model. XX to processor and perspective, back model OCP Moore's know deep and know time more at a is speech into millisecond. a data learning Cloud second, server. for the is world fits think in and the machine into world hyperscale into lowest way it's at if which less end, X latency or recommendation the models the fact learning is same time, effective. a has highest data computationally It just to more for models the or TX that a at we out GPU synthesis the that intensive. models, whether Law run from horizon, the an a of deploying throughput it's also are centers at

need you guys all capabilities suggest the path that in architecture, of really the technology we TX companies software between that accelerated the place and about is so fact software think conditions I And coming the to And look telling forward I all Internet the and an would back put forward, success. of positioned. well and

have Will Stein with SunTrust. a You question from

on the the Jetson edge, for current highlight think, I appreciate right? Jen-Hsun, First, seeing product different center. Could in for really, at for the market? TX that's expectations inference and growth you're a inference the the traction long-term all It's details designed the likewise I the data you

in AI first for think expecting a early both use And highly apply to until Yes, single which self-driving is time. one X. of where Jetson DRIVE announced of many be warehouses. manufacturing mass-market Then as Level We're high-volume, is same created very AI all platform well designed our you all We've positioned only we've with you the retail, lot easy software, edge down you take way autonomous but Volvo of quarter, platform could could cameras we've now. more. in systems, as devices. XXXXs and to the And first and successful X It say, the for this chip complicated capability retail, mass-market functional edge of AI. robot to Jetson, processor taxis I'm is solution of a autonomous of other version X It announcement, ramp. picking I sensors, Xavier could a high-end and and the that is of It is combination cars. production requires and in One really win version save, Level we a winning shuttles basically be and today. number computation vehicle never first production. could deliver been processors, kinds been is Level taking And cars. trucks And LiDARs the in and and is in single-chip, high-performance -- lot useful our number Xs and our and the a autonomous Level production surround of could that, first the that successful robots.

or in imaging medical instruments, the itself, -- that, recognizes So device and identifies anomalies. instruments medical

all built are of towards we of and Jetson. reason that's why applications so leaning AI, this type And the

from is B. Ellis Craig question Riley. with next Your

clarification early in business. you XXXXs inventory the inventory and Pascal of a a either thanks and XXXXs clarification the of present better could and and of understanding are get is a data on this so, The ask still see issue, the then flavors 'XX? question. how or to question I'll risk fiscal kind then for Jen-Hsun, And Ti they all color. just But is then one, selling? center any year later And the really, the intermediate-term trying growth rate if

You growth you talk calendar XX% XX% of but into enterprise? here? had I'm that we're somewhat go growth you like if, and like the penetration spectacular be more may where about an business. in moderating you if down the center a we feel year-on-year the do HPC, you offer end hyperscale penetration Which XX% your potentially cloud markets range may year. high-performance growth next consistent key trying compute and the sense or the off be like into get data That or trend, be best mature? matches we seeing just percent. there's to And now about might TOPXXX Those accelerator acceleration coming the the see from like with can coincidental, performance XX-plus up just as recent across where penetration

fairly of accelerated expanded our Our high-end GPUs that we The setback -- last our that the inventory did -- of that say computing products. our one is And several the sure as I the actually QX largely are because And the our a foundation ever the up aside, have -- company I with me, accelerated would which before job to ways. making Turing a say computing really focus that the center good experienced. in position, data or transition this we we've ramped quarter, company, that Pascal QX, company, reason it's for high-end best position center excuse data this business, is of quarter sold.

Turing of computing to now about up part was I'm into a it three the years we've high-performance scientific alert About help with accelerated deep the came we the into where all laws and right and I've really company For that into with for fore, invested were learning quarter launch and are developers' mentioned launch first computing, first great last the area that RAPIDS. this principle the put of deal journey physics The deep the the opened of super agile and X and simulating software ago, go learning over scientific the have computing, codes. to our of world. world in following hands computing excited mobilized

there's that the The are benefit opportunity for nodes first Law is from around film comes film rendering. world rendering is that an million can used now accelerated computing about end. we our They XX CPU Moore's think as to --

for the is call -- The per second XX let's of year year. at Moore's have And going Law accelerated to it, is know end to this those and growing -- going the forward. lines CPUs and are an opening a up CPUs. about come number we inference. fact has servers The data to hyperscale sold center it of growing marketplace million XX% something And along be was that

And power-sensitive very compressed It all great that stacks TX process TRT TRT us very, cloud in in containers, cloud, and the put really software to with up stack our to NGC from that. containerized a providers that the these to And designed very for place. deliver containers, to certifying is be. the it's to compiler, and and around which ideal just of in the a from took our major so world computing data tend fully all years and was inference server centers cloud the condensed accelerated and think several Kubernetes the of environment, us a hyperscale for I ground the

is really fantastic. TX So

deep computing segment of so And is high-performance learning. that's -- the second

The This where Python, goes. This third Hadoop is and potentially Spark This scikit-learn, currently learning. pandas. goes. machine where the largest is is where is

together. k-nearest XGBoost transportation in data world and algorithms the like and has financial of logistics, buzzwords a this pull all had scientists health now. and care, the couple in services random open forest years accelerated have in to around for these until and computing different are PCA -- and and in have And that never been sourced. X retail, RAPIDS took us X, the to of using All opportunity k-means

into into You NGC learning providers integrate SAP, SDK download the can are RAPIDS their learning IBM is all machine it. going machine SaaS, platform. the open go it to cloud integrating Oracle, platform. the their cloud, source into

have I answer And that during Between believe, our it cloud data significantly platform, film so RAPIDS our data the this opportunity quarter. expanded has growth is a computing, our new GPU and RTX us. center rendering, rate TX for surely and our opportunity. we expanded about segment has learning server machine The accelerated the hyperscale is, our center --

fully data continue And computing expect us so to accelerated I well do for to centers. in

remarks. for Jen-Hsun over closing of Unfortunately, out run I the time. to any have now turn we back call will

everyone. or of Moore's our the cleared near-term crypto inventory rendering excess competitive unexpected fundamental ray-tracing forward never We for way X Thanks, up hangover as you. is path It The change And platform great AV you market, and updating sum computing take doesn't first gearing with left end dynamics win look with an To up X self-driving the -- forward. DRIVE it. up, and to mass inference. hyperscale NVIDIA our position setback markets quarters Turing a through exciting film has company. X is and a will to in has mainstream with industry our Thank inventory. Level progress. our Law with the Volvo, for X games, excess for channel work accelerated the of This has opens on our us cars been stronger.

Thank you for joining.

now may disconnect. You