NVDA - Earnings call transcript Q3 2020

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

Operator

Good afternoon. My name is Christina, and I'm your conference operator 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.

Simona Jankowski

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

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

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

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

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

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

Colette Kress

was of Results by billion and year-on-year XX% sequentially. gaming. sequentially. up expectations $X.XX driven down X% Starting up and QX gaming both Simona. strength X% $X.XX our business. was desktop our Thanks, Revenue billion, revenue year-on-year notebook exceeded with in XX% down and

tracing and price GPU advanced most uniquely the every features lineup graphics. RTX offers ray hardware-based cinematic GeForce point for Our for

have launch dozen Tomb it Duty: the franchise major than strong an multiple PC in game as Wolfenstein ago, two little implementation Warfare, for had shipped with came the Call the of performance. trace week's tracing are last Of publishers, Red their revenue. heels Dogs, Raider, Dead of X X/X is tracing a Reviews the growing on all ray top record-breaking launched reflects a Call including for by of all-star Ray a a CONTROL, RTX as drive praised than of Modern on features. our that Duty, our gaming Watch titles GPUs way. action more both holiday lineup desktop Redemption and With note, this ray with supported release Minecraft, season, Doom, late tracing and now Cyberpunk. While have or GPU October year Battlefield, such gaming adventure game-play of excitement. and more business ray titles

Gaming for our laptops laptop bringing Max-Q laptops more technology, were laptops triple last a This Max-Q models factors thin market. standout, This the season, form growth. year. than driving gaming enabled number by year-on-year high-performance growing addressing are the and demand and students prosumers strong gamers, many to and sequential includes NVIDIA-powered for of light and studio by XXX partners holiday our

of XXXX refresh Super GeForce it GPUs late called GTX best-selling XXXXp PC performance, faster Super $XXX. more can The new buy the and our with best Pascal-based we at you priced all gaming for the the XXXX, XX, which began memory more mainstream just delivers GPU than the announced Super, XXXX desktop In October, October on World It prior-generation GPU shipping time. gaming. and features. our XX% performance XXXX

for streaming with and models, October AI TV XX. the best streaming widely TV the media as the ability player neural recognized Shield first real to Shield video upscale next launched the new streamer deep time the the Shield of has using announced high time market X from TV to definition XK These market. NVIDIA-trained with on to been networks. on Pro, which also generation bring We our

global Two made game-streaming service. we Rostelecom building NOW partners Finally, progress GeForce cloud GFN.ru Taiwan our SoftBank business. in Mobile as Korea's our with for and and Russia's LG service gaming joined providers, out

kick proof-of-concept gaming will a Additionally, Telefónica in off cloud Spain.

data to million, Moving center. Revenue XX% and sequentially. was year-on-year down $XXX X% up

GPUs both grew driven being by is activity engage providing breakthrough, visibility in language and networks fronts. we neural and understanding. on of BERT driving turn, the our for the Google's its natural is computers hyperscale Hyperscale language in is dialogue, improving. way sequentially capturing superhuman of year-on-year, Our believe driving conversational introduction the That, of for model, levels responses. with ability revenue for intelligent demand context understanding, human-like and a is to our AI, X

to have XXXx, some massive image-based are parameters XX First, and complex. models cases than highly more in these XXx, They models.

these As in past. what VXXX-based from infrastructure to the expected is a Model is complexity needed of requires models beyond in significantly compute orders that, training grow result, magnitude, here.

AI understanding, neural from real-time conversational and low latency requires succession to and recognition, language text-to-speech running networks Second, encoding. speech quick very multiple de-noising in voice

chain, in first less the requires is While tasks, a where milliseconds. the XX handle driver conventional at inference. can approaches AI AI GPU-accelerated fail than inference for NVIDIA's these inference acceleration. entire Conversational This is GPUs application major

inference volume percentage more internal public and GPUs, continued Amazon this revenue. announced year Inference general doubled availability this the AWS earlier to and of center with solid In higher continue total We than our to the training of clouds. of Google revenue on Cloud GPU last platform quarter from year. data records. type in the both of TX globally, shipped a were rollout gain a VXXX September, TX hyperscale GPU double-digit addition TX TX activity, In adoption following in

the We week, AI Last inference, of the first announced. industry for benchmark results inference, MLPerf won. were

the the center among that applications, data completed demonstrating X commercially deployment NVIDIA and data for a both demonstrates were the center performance to and all only key differentiator computing across which for us. In best MLPerf is is wide-scale benchmarks. programmability in This our workloads, accelerators edge ones diverse and available performance critical AI solutions of for addition platform

software-defined Los the markets, Mobile edge. with XG platform and RAN more well VR Congress, GPUs telco Several new gaming product This GPUs. wireless by our market At solution accelerated a AI the extend helped to collaboration be enterprise the into up Ericsson. World in wireless announcements this AR, new announced applications we quarter NVIDIA edge. computing as AI to enables to opens RAN It as Angeles, accessible

announced of an Gamble, technology NTT Electronics, We and Vegas. of more than Platform. NVIDIA With adopters Computing Samsung & EGX ecosystem and worldwide, XXX cities Walmart, the the Las Procter Intelligent early companies San Francisco Edge include BMW, East

help with a and manufacturing industries of urban announced Microsoft edge data This facilities insights Additionally, manage stores, created we the better computing. growing warehouses, will from retail infrastructure. intelligent gain and by on collaboration flood

delivered Washington, we out of registered accuracy. pieces NVIDIA, U.S. Postal Service, sold the our delivery the with CIOs service week, technology last announced with from annually, package almost D.C., XXX Conference AI XXx adopting with the that than and is X,XXX world's was more largest held Finally, and GPU faster higher which employees. processing developers, At federal in enabling billion we mail data Technology event, of

Moving thin sequentially a workstations Revenue ProVis. primarily and to design mobile RTX the up to light workstations. driven by reached new for from year a that powerful technology NVIDIA up million, $XXX XX% X% wave graphic yet mobile carry. prior and enabled have applications enough of are enough Max-Q record and

become to a this category opportunities. exciting expect major with growth We new

with being GPUs. design creative applications RTX accelerated are top XX Over

Adobe the tens driving artists now to of RTX available apps. Creative last week, designers, apps RTX accelerated demand and to at Just our RTX-accelerated Conference, capabilities of millions Max Adobe are were GPUs. X for added

also continue higher science, Strong health and applications. customers. this to and care from growing manufacturing, AI deployment sector, see data quarter We customer education VR of came demand public

was $XXX was down Revenue sequential million, decline year automotive. to recognized XX% nonreoccurring QX. turning services down contract Finally, onetime from X% by development ago in driven a and sequentially. The a

Additionally, of we saw infotainment revenue legacy a general industry roll-off and weakness.

as deploy their driven of fleet their AI-based the grew continued Mercedes-Benz Daimler business cockpit of they across infotainment the AI Our vehicles. ramp by systems

autonomous York given Optimus settings be August, autonomous first by Ride pose challenges powered objects NVIDIA real Urban In the of program in that launched time. comprehended and unique need for of New driving pilot number to City's vehicles perceived density DRIVE.

and about software opportunity the last-mile allows We Our stack auto. to long-term and services. effectively in and these remain DRIVE shuttles transit excited provide computer first- safely

is the leading well we robo-taxis. in-car test develop, -- to well computing operate shuttles for platforms development vehicles, with all GPU enables safely AI AV end-to-end offering in and to platform that Our are and as and ranging and from servers believe autonomous We customers consists trucks positioned industry of as simulation. cars

from operations selling were Moving down rest a Non-GAAP sequentially, was up operating XX% million, a million, P&L. gross the were margins GAAP XX% reflecting $X.X EPS from record expenses from a respectively. lower $XXX GAAP was operating $X.XX, average of expenses down $XXX and Cash costs. EPS and previously to year-on-year, year ago. the flow component sales prices was of X% non-GAAP higher QX GPUs and XX.X%, GeForce X% up a earlier. year XX.X%, from $X.XX, was billion. was written-off non-GAAP inventory, GAAP benefit and

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

plus to This commentary or GAAP expected $XXX both notebook expected expected to non-GAAP billion, to by XX.X%, million and excluding GAAP a details be income are strong $XXX are revenue and rates CFO the sequential and million. revenue. and basis expected Capital approximately be are and available offset minus non-GAAP IR expectations or X%. Further and million, million. be included tax expenses and margins data respectively, expect non-GAAP reflects approximately operating gross and our non-GAAP minus to or GAAP XX.X% be We discrete in in to website. expenditures center, are on decline X%, $X.XX are for be in respectively. growth both other gaming Switch-related information plus items. GPUs financial approximately seasonal $XX for $X.XX $XXX be XX to OI&E X%, minus billion to GAAP are expected plus of points.

let the the me closing, for events community. In highlight financial upcoming

and Technology the December XX. Annual Auto at We be December and Telecommunications Crédit Suisse on XX X, Global Barclays Tech Media Bank's on December will Deutsche Conference on Technology, Conference Conference

please questions? Operator, We will open poll now you for call for questions. the would

Operator

first Merrill line Vivek from Instructions]. of [Operator question your Bank the comes with of And America Arya Lynch.

Vivek Arya

that. you you that that sequential more and Is you data into more the what's more mentioned color I if on seeing center. wondering I give were going for center some of think us on Colette. then in think drive it a how Just enterprise, just will one, you the hyperscale, markets was driving strong And term growth follow-up HPC? some on and Jensen, first into you sustainability longer near data about year and next my have color data center going do what QX. growth For could

Jensen Huang

Yes. had QX centers. We Thanks a in a strong hyperscale Vivek. lot, data

number shipped than shipped the we most we time, earlier, mentioned are first very driven the by and of Colette inference. of more for a record And And VXXXs As TXs TXs. VXXX. TXs

doubled it several It developing learning on most and some And double-digit, solid a center people of models. really vision. is have all recognition that been now time, is driven working image -- deep deep been -- started data doesn't really year-over-year. But computer take factors. with by we've business as that for the and inference our The fact, In -- know, you capacity. much learning

Over important the couple deep a for last couple One of development years, happened. learning developments have in using very recommendation is systems. a breakthrough of

or based systems Internet. on backbone personalized your you possible all there looking you putting is of just data your patterns right all just or Internet personal interest the the As your out of data on entire the made that search, watching is your web know, of recommendation movies, your all is you're shopping, news, because or about the you connections, do vitally at whenever recommendation important. profile experience front so systems in there the social much doing of is pages, by of Whenever use

seeing very the deep system For we're throughout recommendation based the world. time, first learning on

out. backbone increasingly, learning. going you're And so deep this Internet roll see And is to based people be going the of on to now the

in Conversational encoder. AI at language to beam And together text-to-speech then system, goes The a of done in very, to could be not then accelerators, understanding, recognition, possible that gets to has to has a noise forming. pieces; and takes learning in amount now gets then Then do done time. it's connected speech And real it some processing Whereas able connected AI. coming speech go into second be acceleration speech which real part been is recommendation do that and and very to seconds natural first, which recognition. time. off-line, to deep or we're quickly. be conversational Then which images done to accelerator requires it really without real you without in of on handful has conversation all time. the all It of process models, an do and to NVIDIA's it

that in works And approach. is as stack understanding language recommenders, really, breakthrough so learning-based And natural various encoded working transformer, really what the combination breakthroughs from a from natural well come bidirectional incredibly is significant. have understanding called of speech deep since is derivative these language as well. quite the then, And really that breakthrough

the And just with very so is do complicated. what work work is of hyperscalers area everybody, larger. getting bunch There's of are about they're we're models to whole across a this very, the and -- together, we seeing work people models world, that The has really large.

so is one that's which the And large category, hyperscalers.

to sensors you way, do of want cloud, costs in there's vibration taking many types the continuous sensors, The that industries the the various kinds there action. at this for the equipment, applications, screen a it's money. whether situations. temperature used where the of And sensors And AI to all introduced second, about to that have point we of out edge. monitor This into to which sensor, are is the data, the you video which health all of of there's don't barometric monitor is lot back based are the a other or which reason of kinds because sensor, processing quarter, of really conditions is on

or want the whatnot. You to Maybe or you're matters. gates drones point the action robots of latency vehicles take action or controlling at or because

you to company And you afford sovereignty. put the the the Maybe you at lastly, of issue then own that do major is processing edge, that to that and cloud. data can't all and, therefore, one processing into your have are data doesn't

the you a smart so logistics, is for And takes places it's go to to manage afford And this cloud them. around Walmart We we're can't which departments basically our at is faraway there BMW logistics, logistics Postal visual the adoption. just And called cities, announced we've last advantage we retail, States platform. EGX, -- much it using the really enthusiasm And this managed industries: these for announced manufacturing, & for seeing IT because the And and full these services secure, we the could in in some Electronics Gamble seen inspection. out using probably Service. quarter. then operation week, Samsung there great so Kubernetes largest manufacturing, native, warehouse, world, Procter is of last remotely, United various with a so deploy that. built single stack NVIDIA's every model. edge completely platform of

clear say And so intelligent that largest would those -- in be the will are the edge industry likely I AI reasons. world for rather

it's if logistics, just you size made kind nearly of bit could stores save a If retail, $XX little lot could trillion. money: convenient, transportation, industry it And the retail of of warehouses, more estimated be the a farming.

mass. farms think in the million world, X/X the I like land covers world's X.X of there's

last started where with this quarter And I that make grand AI so think there's could of And edge a places at could of this NVIDIA adventure to put lot EGX. we the be announcement be big is the difference. and going a the

Vivek Arya

you you're to how at in I PC growth in forward Jensen, very lot quick And more that some gaming, but October? guiding a on Right. I quarter looking a probably decline. had of imagine are that follow-up, console going is you to seasonal due in think good January, declines, as

from come are into are Just X going as out? looking competition supposed you that then -- also get growth and next into you October at year January gaming how consoles PC to new

Jensen Huang

Yes.

selling in QX for because vendors they During a season QX the we of that declines and QX, have And that reason in builds, we see seasonal up seasonal console meet notebook to their so manufacturing of QX. could the hot also decline is notebook normal normal all builds. line the and see

frame declines And doing in seeing gaming we -- is the normal just -- QX PC what see these for QX time and so are Overall, we're fantastic. seasonal of systems. RTX and

that home the Let RTX so me future would this it's tell point, is think ray run. and clear important. that a I fairly tracing why that at it's is you say I

major -- consoles effects anymore. not look the The had stutter on signed their ray compelling, to every go has photorealistic in step consoles. the ray possible tracing developer about just tracing. so really and to Just is Even next-generation it's next-generation back game include to

clear world And gamers the so million it, of several is hundred tracing ray there PC that and have now I the think benefits are forward that future. them. that it's RTX the I'm looking upgrading fairly And in don't to

is this until this really created is gaming. the my the now and And Notebook is really we're didn't platform laptops PC came this and beginning this Second, brand-new second-generation gaming likely just XXX turbocharged platform, that last combination Over that. are our RTX a PC sense gaming really of season, of game available platform, new be along. for going exist And Max-Q, Max-Q notebook And gaming. we innings gaming PC that to Max-Q, emerges. largest segment. in a

upgrading it's tracing gamers PC And PC and drive of so growth the of going called the to and really exciting, notebook base is and gaming quite gaming time. new combination this continued our segment entire some to installed RTX for ray

I'm so excited And that. about

Operator

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

Aaron Rakers

I have I a well. if follow-up as can

Just margin had also in the some selling margin the that through here, of solid written-off quarter, components. thinking the you benefit trajectory about you gross of noted gross upside

in I January do here even in first we is was what So quick I gross quarter. trend? recent again, And question impact about this have reported gross that margin trajectory think beyond terms of about be that we And What should of guess most the quarter? follow-up. how a the margin thinking

Colette Kress

our with gross components, the point the benefit, overall margin. our current we the primarily inventory to to of percentage release as the quarter, X was In question. net Thanks as associated Sure. refer for provisions net our about

time. is of forward, over mix largest you gross our still the going As margin driver know,

we'll term. to and long benefit gross improvement expect to the the continue the see, long margins we received, that term, gross margin outside do of we improve, for Over

Jensen Huang

Yes.

become you company. As just to that, know, really as software you know, NVIDIA's a add to

care take do, of -- a whether cloud put at it's GPU, they gaming, course, look having start of But scenarios almost is we of of work. almost or platforms graphics, in you just If that these can't almost rich starting health or point. platforms: all in or all the really cars, you rendering, the stack chip or self-driving from our GPU all the artificial these analytics world's products, software. a of And a best these and robotics intelligence data everything

most businesses software-rich, And then we now our that of we're address so highly And platform they focus are on. company. and a verticals secondarily,

so great our a a from would be time. And is company all available platform cloud as the deal intensity, over OEMs it's platform the providers. of software and higher that has natural that margins And

Aaron Rakers

year-over-year Yes. Very in you a print. -- as well your mentioned that basis business as your seen grow then you've center helpful. both in prepared remarks And this within on hyperscale hyperscale data last quarter-over-quarter

You you are seeing improved out mixed also visibility? improving. feels mentioned Can like you your Or there's your data us of because What just hyperscale what is there? some understand that help it seeing in what your that points underpins visibility you're the exactly piece guys business? in

Jensen Huang

this high. software very the is That deep it's Yes. is time, inference from is to see going We're mode of reason industry this strong not computer practical foundation We understandably learning is of because that inference. And one -- is so AI, intensity stronger that the accelerators. much the computer QX. be And really a computation without the deep And of to first is opportunities. QX. largest going learning had for for graphics, the a aside

to Those recommendation mentioned about deep I learning movement really so are hard earlier the large-scale to And train. models systems. really,

real-time about Because not AI. really processing, practical. requires is seconds I mentioned several earlier conversational conversation

of we it it to milliseconds, have all possible. And complicated milliseconds. MLPerf tests more than makes submitted if as reason supercomputer that reason great sits quite on chip our accelerator this we a the you that the them. What chip the it the really inference far that performant, the understandably for although makes benchmark, that chip. top top software, it's large. one and body now. The And many of we challenge. science really, of X, both solved, compilers do tens -- -- accurate last only complicated people one because so it because so years -- This on we've won are as And those it's it, sits all and great the And And really and of challenges. talk it's software the X just make really so results of you in problems hasn't and working have why about been demonstrated the it's is and software for of is wrote And to X a at is been X that a for well stack complicated. the computer hard

the And opportunity. been so for about great talking this time some inference is now. really We've

now production. workloads the into Finally, diverse of moving set large very are a and workloads

I'm so a -- inference and opportunity hoping visibility that And be trends the I'm the us. should and enthusiastic seeing progress for large about the market

Operator

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

Christopher Muse

I on to last guess follow I'd question. that love on,

commentary, your to you, on than first much is here heard benchmark. as particularly this relates clearly, bullish think, So it more I I've inference, before Jensen,

well? XX of out would as I -- guess about grow enterprise, opportunity can see see you kind well mix center could looking And as you time data considering, you months the inference next as training XX, could a a there inference as talk cloud be over think, how large within versus bit so over I versus into

Jensen Huang

that's good. C.J., really Yes.

it break me Let down.

we So are three inference when and public there about cloud. think hyperscale, training, parts:

of data to whole work be now of now might the X as ago and did with And work, recent that really This event that the natural natural only lot also as The a of natural computation scale AI X watershed able model doubling open Google where you for was kind and large the monitoring of much models the to at models. of These understanding. years computation growing necessary language amount solved. measuring amount has is because large Training, AlexNet the wave the computer reason is in BERT understanding necessary, problem. every therefore, breakthroughs harder well. now And was work The AI and And been have for language a for now train much, they've a done are to these larger. watershed getting a months. recently of a just transformer-based is, of problems vision-oriented language are event sudden, all that of is course, understanding. seen

scale the training so And the tremendously. of grown has

we're year I to is fair a number do think what to see systems very of this going this sizable training. installations GPU of thing, very

The and never second us, our really it's untapped large. because is to we've us, The to bandwidth capabilities They're memory large of why market for require amounts programmability. of inference going The very large really a until very is going an validate spoken I and about been for this reason amount really that models to be market opportunity intuition able be market now of inference. and is memory is are part untapped haven't large significant large diverse. computation, very They and large complicated. that of amount it

one And so never validate -- hard. them so also able but hard. this have really, they just the results validated of benchmark finally, it, results is I've our Inference the submitted this business course, resented -- some before, frankly, then our sweeping intuition. attempted, with really and is I've And although only the to benchmarks talked and And been it. many and, about MLPerf of only that

We're of systems, language And now really a interesting conversational so recommendation across with really, CSPs understanding, are AI, global. natural our working problems. whole now bunch engagement just

reason is because the the is world is of third our Now CSPs, And the of in so number well growing half the the the so ones many incredibly. public of the and particularly reason almost growing the why cloud represents that still cloud with start-ups piece. cloud, for AI about is

around tracking close of health more in Fintech, extraordinary. Internet, to companies care, I transportation, in the just number retail, AI In think start-ups XX,XXX AI and in there consumer in is out the world. we're something

$XX they X billion, have into billion to or don't fairly these over invested last because X it's use $XX so some providers X, infrastructure their And in been cloud have think course, the start-ups, I that invest own of start-ups. complicated. service years,

so there. of growth a seeing lot And we're

the into, so intelligent inference areas us And platform and this X growth ventured just And AI primarily And time. cloud we've training, and EGX been that's edge, Then there's public growth: public called hyperscalers cloud. points secure. start-ups. building incredibly which of X hyperscalers. for -- we the the native. cloud The some is of recently it's give It's

it saw You announced been It's early some largest And is in afar. the we we've world, and performant. quarter, of Walmart this can last stack And -- working adopters. manage some -- & from company them: some with the the Procter and we and Gamble complicated. USPS. It's BMW logistics

of I platform, And will new long that the this serves. the so think, be term, that it because is the And industries likely largest opportunity. reason for

Operator

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

Harlan Sur

lot somewhat level. But less that and, changes, also has shown slowdown. fact, a at sensitive to the tensions, these insensitive of price trade gamers concerns enthusiast tend in are There be China economic macro history around to least at trends to

of gaming the the demand gaming segment, discuss is of big part out So such trends just can this you given that a China geography?

Jensen Huang

the get enjoy. This it's couldn't category Gaming is also because powerful China, before and our is fastest technology never we GeForce as to both as own to new notebooks existed adopter gaming really well in category. brand the solid in that so delightful of gaming a it is RTX there or notebooks. notebooks This

and been in success GeForce great notebooks notebooks and RTX RTX really saw we so has adoption with fast. And China,

Your -- buy could because after which forever; want esports, and people are you play mobile, comments games most sense gear, free-to-play. that you you you it the are games the best is largely free-to-play enjoy make much of gear, pretty these days. which primary The but the

enjoy in can for after outfits. personal own that, invest of you time. some long a You I it your think And quite

gear the really so important. is And

One work, of social. done the particularly great where with has really areas China, we've in to do

your platform content We Broadcast called feature a And that, of new it extension basically Engine. And Experience. to to have this as GeForCe broadcasting RTX applies an called there's it. share AI

make could movies. You

turn screen yourself You your game scenes one networks the We background cut And the art, with it overlay of green AI. what irrespective a out share social coolest to noisy and of is you could and the got. and capture of behind that applying use into of all you've could you. background stitch features top you the without it you AI out, to favorite on basically,

as platforms videos of like user-generated kinds back and social there really content social all all that. things community you really to kinds so hyper of and has short -- they a know, cool And and China and share super communities have games

And one really it so that that GeForce makes feature has additional successful.

Operator

of question line next from Goldman Sachs. with your And Toshiya Hari comes the

Toshiya Hari

I on about your prepared the wanted talked you down in automotive. infotainment quarter. Colette, business to ask remarks, legacy being in your

customers, what that do you years at of collaborative legacy us haven't versus percentage you're ask newer quite solutions? many, trajectory more were your doing I inflection automotive expecting the the a seen be we've announcements, infotainment next it revenue is maybe importantly, AI/ADAS that automotive in would of of growth Jensen, we of that a ago. over announcements, customer And can to curious, if this And couple point many hockey-stick with Just X, like yet X.X years, speak the the appreciated. question because heard work sort some feels

curious when of how our kind just going So expectations -- we set should forward. we should

Colette Kress

our of our Yes. Toshiya, representing business. question It automotive our first infotainment or half is maybe more systems about revenue regarding business. overall the me still in automotive the legacy for address let

including as our grow well, continuing grow may well as well development solutions we to be and sequentially quite as both cockpit AI vehicle We that our autonomous as year-over-year, services. doing, have

Jensen Huang

the passenger-owned in -- And I say And car about you that's amount out of think XXXX The I passenger-owned be have them a first road, the as Volvo. in we're couple it are slipped I -- first we've know, late we're them there's of is and to so. The a so is the vehicle to XXXX, see. Most expecting time past. expecting XXXX, lot already early And developments the years think or going going we've -- would industry, this is probably before, the XXXX of the still that of then pressure, some quite talked so. about it I'm spoken And well. under on most Let's AV the something generations. and be frame.

consists areas. strategy passenger-owned Our vehicles. focus, One our several of is area, of course,

taxis. part is second The robot

delighted progress. going in every company carry a in they with China. of major about with their large see increasingly And third and to know states. goods. hear that don't of part developments has number trucks, taxi And vehicles we people, do they're have the we're They're in then news the shuttles here of. We that robot to They're carry when them, you just And Europe.

Volvo. And different so That have a Volvo major as and with we Volvo know, companies. you was are X development Cars Trucks. Volvo Trucks,

heritage Cars. going Geely, major to automate we is Volvo of to the Trucks One them with program belongs them the of delivery Volvo Volvo. a have And goods.

You delivery also during mention companies grocery or delivery delivery. on within warehouse we're I'll us various with see goods or GTCs, product a that working

eventually very the is technology -- develop I certain. started like things point, And moves think, -- believe logistics vehicles see that to continue that You're autonomous because bunch we the be at capability a I everything that to going fully whole for down it's to have has is, similar, will propagate passenger-owned this fairly autonomous. starting and or into vehicles. technology development the to of that,

computing it's and the as well is Now the software-defined, as systems. and simulation developing it's both our system, strategy AI scalable, in in-car development AV

when large and and a And And they are, start of a great they're AV data, so using working ramping when most somebody's miles for it AI, on market opportunity us. up they're of us. opportunity very and becomes there's them collecting for

single anxious and will is conclusion. every I'm AV. be. so a This foregone car progressive to they And be. And They as will developing be aggressive company in see

Operator

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

Stacy Rasgon

I have two data center questions for Colette.

question, first want center I growth your The data in for sequential QX. kind to to strong of return outlook

business Now QX. out XX% strong growth didn't sequentially we And into call this quarter. as going you grew in actually the were

seen you QX? how think in in to growth call You calling are like that as to QX? we've in expect Or than suggest stronger that the me it already calling strong you to saw it QX it that you for what QX and relation QX what about QX be growth? well wording as in define you do out for in Like sequential didn't out of Does what expect and QX in out sequential we being would you're QX. given you

Colette Kress

our in would gaming QX had how both we Sure, and QX indicated the expected Stacy. growth than had and in data in guidance And that When we from our results. center also completed quarter stronger finished from had provided data guidance gaming both and we we that. QX, stem from center. We

QX to totality a decrease is sequential versus in QX. Moving QX,

to data What what expecting as strongest we reminded as our decline associated to we in QX. do that terms our we if QX our seem We with seasonality have therefore, wanted notebooks with did with of being the it our growth sequentially. about in are primarily was, seasonal center likely, and consoles quarters strong with also associated have that, the teams move we we and that emphasize have totality overall as in to overall downtick to sometimes a well overall gaming in QX be want

Stacy Rasgon

the So I guess you being would as growth to strong ask saw well? again, as what QX you in define question

Colette Kress

believe XX% Thanks, data higher how we in terms our get we QX. Again, QX, but would expecting I of when was ends quarter we'll expected of growth. strong we into see Stacy. the are growth center, to than

Stacy Rasgon

hyperscale And for Okay. said you second was up question, year-over-year. my

we after, doubled like I wasn't down from that to that? of in very about Day, was suggests it expect we've me coming enterprise almost drove and you few lot year-over-year. this back How think And entirely going know the this year-over-year. opportunity context over the off And at me quarters -- suggests to your seen Inference there? what But that strongly should forward? year, commentary the what's going growth enterprise the -- we which about the And enterprise. on the Analyst Now in was do just going last growth, where What down a to it peak. that's was of was said that year-over-year, down it year-over-year. last

Colette Kress

Sure.

year has business throughout. ago of traction keep DGX. also associated had year announced at very, quarter, a base. enterprise Our a continued But in things been in strong a beginning our deal ago we to mind in small systems a lot very, with and We've see very we've that ramp the from there our very great very large to with over

look quarter-over-quarter from just a quarter, can So period X when we looking bit we of at or have lumpiness. a little

in just impact extremely deal year-over-year due to QX. really that an is year the prior So large

Operator

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

Mitchell Steves

one for just noise, any have I I for question, Jensen. background apologize but just

us year on give the was utilization today? I'm the last What for rough how two. -- advanced of couple or the application? update what XXXX, in you So wondering can how a GPU it is learning deep just over that's

Jensen Huang

year, see related used see. started sold we would was growth record GPUs of for now say this to VXXXs it sold than more we now training. where quarter, the I have inference last point to inference them TX Let's were both of to all more and nearly XXXX, sold And -- we for -- training, the we this we highs. that's

that both DGX we to hardly They're TX existed great. you just than doing Colette large record ago, and system year it's comparing more But a of have And sale had And year. are is at if made, highs. TX so levels. them didn't the we that VXXXs, year-over-year, that, selling a a now year excluded comment the ago the this and didn't VXXX record

for gives And feeling so a it. you that kind of

major grow my really I is it's kicking that difference that quite that's is really nicely. the to sense to gear, inference going and think continue into

Operator

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

Joseph Moore

And But how if you you And of lot announced more talked is little could environment. compute is and the that AI guess address talk wonder aspect in with I that opportunity a C-RAN opportunity? you kind you the GPU? C-RAN a big core to a about XG bit -- services at about Mobile World. can I the IoT there

Jensen Huang

Yes.

building ASICs you In where data it's like look is are their for for the the data the are networks, that to in world center XG. And to heads radio reason that mobile units. to at support untapped too If would level radio today, able baseband into players basically not there software just of of are market. is the the because and rigid move DRAMs necessary CPU the people really performance center. that's And an the to for a be the put able BBU,

VMware. along cloud. that data -- announced center-ready part telcos, us support core we that And with another goes that partnership the open the centers needs stack all a of entire we top data center, it's VM realize to center that so with -- a can and VMware. a becoming whether partner of they're partnership are a from our we and is stack momentum They data data a our had And recognize with GPUs in announced would GPUs. recently, adapt like to in, solution of like seeing the quarter, Hat. programmable the the they increasingly, OpenShift their that We Red They environment richness on during software

stack to now that telco all the Kubernetes. machines ranges And have VMware, we're enterprise virtual to with with world-class from and way a so Hat, all Red going the hypervisors way

world course, And -- centers vRAN our C-RAN, create software-defined to radio. stuff and strategy goal of And is things include data computationally-intensive to center really firewalls. the of in But data centralized networking XG like the is our will, is as networking. well so as new the really this networking software-defined the

able going whole whole stack customers stack of of as need And software center. as Aerial learning essentially to to is is going create to deep and deliver performance, that all the us put -- me, data to basically and so this we tracing. be essentially software allow for is is a it layers for the what run XG the way and for -- XG deep a learning. MIMO And incredible the many software in the to done software optics scale for to for software it we've same we're XG stack Cuda-NN call a excuse in ray And to flexibility creating run stack what exactly and and the highest Aerial.

power putting know, data as is and it The you of into center, flexibility fungibility.

baseband traffic helps traffic where the the of X reduce units center. put can is you center could data in hub. on capability to latency XG, the And center, you in the computation going, regional And but cost. depending could do you another that shift from the center can't in data low that do you the the that somewhere a them data cell in data something towers, in basebands, With

and that they're is second to it's whether retail would world. center's be for telcos or is And applications The whatever data a a and all geographically provider the edge service are benefit distributed love smart the it over things at because edge. located the cities warehouses the or like stores is computation

to with able AI in their data exciting to so use them. is be to to use IoT also combination really able to be And center

and that that's really going lot XG these so the more think we're and -- run to and run the data data AI will at And the as edge future really networking, that IoT applications center, centers I is mobile network providers a this -- have including the And software IoT see the edge. service as to well

Operator

Harsh Piper of And your from last Jaffray. line question comes the with Kumar

Harsh Kumar

give background us of And console of wondering, for you the could was you and down. the an noise. versus Colette, guidance, I maybe gaming. it's might us idea business laptop of apologize impact us be the what give But I the driver give an bigger idea the the In could maybe there?

Colette Kress

be in to for what expected seasonally case In Nintendo the are terms wait say, QX, our for of them of they we need. consoles, to would do I both of the down. assist

will on both ranged, the decline. you. businesses see And cases, have how quarter them million Thank of in see ends have a to So about in $XXX that. of these we both the maybe But in we'll totality X, sequentially of quarter. totality,

Operator

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

Jensen Huang

a The our had great is initiatives. and growth gaming impact moving to demand. Thanks, big graphics of that. and computer strides in strong We're making driven good world X tracing, hyperscale by quarter business ray reflects everyone. We

are Minecraft. and RTX holiday this of upgrading We're look of and graphics season PC rapidly ray of Call reinventing applications including Modern adopting tracing. RTX-enabled, Duty: games the millions beyond blockbuster artists game and hundreds all-time, are forward of biggest to and the by of of Some millions gamers creators RTX. Design the best-selling computer Warfare; used to

catalyst continues will inference. we AI, a strong quarter, on our demand And lastly, and race both for conversational The which this visibility was for and to us be Hyperscale is training powerful in improve. our IoT, our platform you world's XG, largest the and industries. revolutionize February. will beyond progress on computing cloud extended the edge, the where updating forward GPU-accelerated AI look We have in to to

Operator

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

now You may disconnect.

NVIDIA (NVDA) 14 Nov 192020 Q3 Earnings call transcript