NVIDIA (NVDA) 25 May 222023 Q1 Earnings call transcript

Good afternoon. My name is David, and I'll be your conference operator today. At this time, I'd like to welcome everyone to NVIDIA's first quarter earnings call. Today's conference is being recorded. All lines have been placed on mute to prevent any background noise. there question-and-answer remarks, speaker’s the will session. be After Instructions] Thank you. [Operator conference. you may Jankowski, begin Simona your

Thank everyone, XXXX. for remind like replay webcast the webcast without available It Good being transcribed and Vice XXXX. our or the is on from will discuss today's you. President to property. afternoon, Relations NVIDIA's second call are of be is you Chief call First call fiscal for the The NVIDIA's website. NVIDIA until Financial prior Colette our live Quarter that, for Investor Executive conference and of The results Officer; our written of call Kress, Huang, content Chief to reproduced Fiscal can't NVIDIA's Jensen With and conference be financial quarter today consent. and me President welcome Executive to I'd Officer.

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

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

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

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

with gaming a strong. RTX in from mixed GeForce the of to billion fundamentals become $X.X momentum product installed Center double up a energy we launching X% significant base brought the GPU strong well $X.X buyers the platform, execution set and our RTX revenue driven Total strong and demand end new drive first-time is RTX. Gaming a PCs Gaming years past games GeForce expansion has year-on-year. grown though our product of and ever. year-on-year, ahead. Thanks, against RTX-optimized strong the with XX% now has The by in into Overall, RTX Data cycle. record, in going and installed powered in helped demand with has been demand tremendous from runway higher-end of to cycle both Series XX% Simona. see market NVIDIA upgrading region, remained XX forward. revenue And our sequentially record sustained those challenging new sequentially has of Newzoo. added a-third has fall a Gaming and PC the Data the on was Series backdrop. up as continued Center X% and GPU macro We as solid for with by largest base delivered rose the and with XXXX, in America's world, We billion Revenue estimate apps, tremendously XX remained gamers platforms best XXX and Gaming. still has that gaming almost the by according and both industry Since industry Starting RTX two year. standard XXX-plus experience last million their quarter

the Europe we of of parts softness in COVID started and due to lockdowns. in the the parts related seeing war However, China to Ukraine

and in reasonable any GPUs. it these difficult has prepare new likely is around cryptocurrency activity extent impact as hash ongoing we Gaming levels increase in nearly QX. inventory to to some in expect revenue architectural on quantify transition projecting The we expect normalized we The lower for year, network degree reflects we As sequentially pace the later a of reduced mining for are decline rate precision. contributed to demand Channel QX. Gaming us mining which in with in of to Ethereum remain

diminishing NVIDIA XD applications whom of strong forward. just the this ahead Omniverse The from posted drivers. this and creator also tools the not for featuring AI. XXX creators are developed season, offer over Max-Q sequential for lineup. million painting We've time Laptop empower a of dedicated powered studio growth, spring also revenue creators $XXX year. Series NVIDIA now RTX for light upcoming refresh last at driven from energy-efficient collaboration from with this GPUs our broadcasters. XX broadcast there by to XXX going by XX streaming to at year-on-year up live contribution expect back-to-school and the RTX With are content year's we economy Driving canvas and XX laptop models landscapes Gaming ramp growth gamers, and but category the technologies, fast-growing individual and We Series and advanced estimated and of and billion is to thin artists

access subscription for can We a we streaming, continued controls of to and RTX service. plans cloud NOW build Fortnite on with Safari out our the NOW NOW $XX.XX the our month. added devices, GeForce to We new to library, class Android now XXX games. touch XXXX iOS top-tier app. through total week, NOW gaming games on And over over GeForce X,XXX Gamers GeForce bringing the mobile offering streaming browser launched the web last with GeForce

Ampere software digital adopted supply also train Kroger the robots. The companies. is enterprise with of as Pictures continues quarter of ecosystem world's simulation. remote customers the digital imaging efficiency out our industrial just the product largest is warehouse Sequential third-party GPU Strong X% to QX $XXX was at pipeline, the and by year revenue NVIDIA to was customers Animation twin expanding our flow their Omniverse Omniverse workstation using using Top using cases architecture PepsiCo configurations. Sony of Omniverse supported to Omniverse intelligent efficiency ecosystems GPUs infrastructure in by the optimize connections enterprises growth being better automation, medical at Omniverse Pro offset and In year-on-year growth and developing rapidly driving with to more environmental such twins as Demand is to digital to expand revenue. the robotics, and RTX availability, Amazon sales XD GPU and desktop creation is ago. XX% include optimize Moving chain. sequentially and And its such sustainability as million rendering Medtronic. cycle. use developers mobile general content Omniverse improve Omniverse. second up in employee's hybrid was higher to store remains strong lower store a by from work. down end office was multiple design design and Visualization. support continued build some is to to

Moving year from to QX ago million, XX% automotive. of revenue the and declined $XXX increased quarter. XX% sequentially

are Lucid makers truck pioneer largest latest kicks has companies. the with O-RAN production in China's on a DRIVE great cycle and the auto wins automakers, is traction award SoC that in product Our and robotaxi O-RAN. announce now fleets with and to were over In their an next-generation in building they XX BYD, DRIVE customer marketplace beyond. ramping from off customers EV EV major winning and O-RAN QX, QX maker

win just a billion design six year exceeds billion now automotive pipeline $X $XX next ago. years, over from the Our up

billion language from strong they services training capacity needs our volume in from and this driven for driving choosing demand and to Center continue and their recommender our the includes; infrastructure telecom. growth remain on large as financial Record XX% Moving products Internet grew than percentage functionality. add double-digit year. supply Customers demand by and Top sequentially growth of vertical and NVIDIA's hyperscale adoption Revenue cloud leading try and large customers with graphics. strong adoption ethernet for more strong growth XX, models grew adapters. last year-on-year. are Center. accelerated to AI, for and AXXX deployments customers consumer driven cloud industries. with revenue revenue robust and $X.X by accelerated their XX% Data inference a doubled both Top internal broadening Revenue to across by quarter software of both Customers workloads and external demand. conversational GPU XX industries Overall, networking include; hyperscale primarily year-on-year, and was strong XXX-gig from constrained Networking keep pace Data computing broad-based workloads. performance to companies, verticals for next-generation vertical systems,

build super AI what combining and In are DGX the are addition, to modern networking demand and raw revenue is output. benefiting pods from input growing Customers essentially and opportunities. data products as networking material intelligence the our as with factories compute for cross-selling increasingly

of we supply networking expect continued improvement are year. still Our throughout the though products the rest constrained

of GPUs run revolutionized by TAF sets discovery inference traced do training DeepMind for for enable billions include; networks unprecedented based models To translation, NVIDIA Microsoft structure driving without Transformers generation, NVIDIA software is industry stack. hundreds times, workloads levels parameters. data. data these One architecture-based use answering as such allow well for like that, low natural processing, need NVIDIA AI accuracy human Recent increasingly like language are labeled including of the Transformers full BERT AI, which text language and and been of of deploying adoption protein sacrificing models to of NVIDIA learning has drug They prediction. a for Ampere To the summarization without biggest our into self-supervised giant questions. and Megatron breakthroughs enormous models. transformers large large customers very GPT-X, AlphaFold and the transformer

by large leading working of in wave upper GTC, based the time. accelerator, models the deep the GPU, two recommender are performance the believe to largest, Packed hyperscale we customer billion magnitude hitting our for that in systems next-generation addition, largest Center HXXX server market the is of We offering driven Data makers workloads announced GPU is the ideal for is XX at customers the cloud. HXXX demand are the In an and or HXXX with HXXX and transistors, powerful order AXXX. world's HXXX. most ramp large qualify seeing scale today AI new language on AI a using NVIDIA innovation with At instances and advancing over is architecture. language We we rising models leap and perfect increased

performance, Upper Grace graphics Grace the the Chip, an an workloads Grace and with introduced Super Grace at powered we our server NVIDIA announced servers system cloud first HPC for for and Super Switch, ideal networking and platform, X a range of is the introduction with accelerated rich the late two CGX XXX-gigabit software. Chip, CPUs, second implement AI in the the that first Spectrum-X all integrated track to Hopper the to DPUs, system an Omniverse and chips XXXX. coherent and of calendar end-to-end be the be This to will to NVLink over which workloads. multiple Grace quarters which on as for features builders, HXXX for arrives gaming, ultra our massive DGX computing In Future GPU, optimized per technology. partners ASUS, the to video Gigabyte, centers. first memory ever NICs are from businesses, extreme GPU and to supply and NVIDIA's models die chip exponentially based lineup. security connect of high-performance CPU systems ConnectX-X the Built the half can Hopper system. of SOC including on that configure NVIDIA new market chip designs scales advanced features and Grace energy-efficient, and including Grace launching part based a designs system in Foxconn, CPU to These CPU, stacks Data our can demand. factories. DOCA Grace dozens product digital DPU virtualization the powerful as world-class at quarters, with on Super SmartNIC, and of die, supercomputing Hopper Ethernet simulation volume on OVX Chips Wiwynn. by for of integrate With and Grace NVIDIA's address we which next-generation like COMPUTEX, new wave ramp Spectrum-X, coming first CPU making Grace, ramp GPUs, are Across launch We're customers enabler we data centers AI SoCs reference demanding enable to platforms. is are in the Building will to Bluefield-X NVLink kicking DPU data to NVLink, the The We've NVIDIA pairs CPUs NVLink, Grace AI be just NVIDIA our low-latency, year. in server brought AI. major based and networking, qualified support interconnect we're chips, These broad week scale. and of software certified well will QCT, HGX world's new Supermicro to Hopper by SERDES and off in side, the we product and advanced Center as HXXX open for cycle CPU, our twins model. for the for a high-speed As and are will growing new customer and on NVIDIA Spectrum the custom or

was Moving to P&L, and a basis basis quarter XX.X%, XX.X% and up first up the margin gross points the XX of XX up rest points gross margin GAAP the for year was from non-GAAP ago, sequentially.

chain able pressures. been and have costs rising We to supply offset

margins to maintain levels current gross expect in We QX. at

margin increase percent longer operating charge XX.X%, the operating was impacted ARM billion a becomes to Non-GAAP Going transaction. new gross by forward, have of a acquisition revenue, termination as margins to software related XX.X%. ramp GAAP we opportunities and was term. products $X.XX margin larger

managing slowing these closely to macro inflation current integrate us we've are far focus year This our opportunities, budget employees. on to hiring are been our new of in now expenses and employees taking this balance to with operating We existing enables our also persist. as so and successful environment care our the very growth

the still multiple level we track expenses off high bring-up the which the of our of to amount first operating new in reoccur the year increases to half the products, significant QX range expenses of as half. should in We year. on this are a expect related after second not non-GAAP includes grow sequential to XXs

billion $X our repurchased we QX, of During stock.

of Our common share December to billion extended our and total $XX an XXXX. of stock Board repurchase through increased Directors repurchase program to additional up a

turn now fiscal the me for XXXX. to Let the second outlook quarter of

sell-through assumes on China the impact $XXX $XXX Russia to QX to absence we Russia $XXX and impact Russia of the impact in million to estimate and Data sell-in an Our by lower affect a lockdowns. our million. have estimated sales of approximately Furthermore, in We relating COVID Gaming Center. to China outlook of estimate million QX

included $XXX $XXX events are approximately on non-affiliated of expense on minus operating In expenditures excluding growth be $XX more $X.XX the expected and to or plus We expenses And expenses be $X.X million, website. GAAP X%, and minus details in financial decline respectively, the million to an Revenue be Center to and approximately to items. Gaming. X%. non-GAAP in sequential available expected are Non-GAAP operating expected than offset the excluding gains expect be be points. expected tax is margins CFO and to strong Data by to community. billion, IR sequential non-GAAP GAAP in me GAAP an are Automotive gross the upcoming or Further plus are and are other closing, our be or expected XX.X%, million. and financial expenses other to XX.X% highlight non-GAAP expected XX.X% discrete plus to be let minus expected billion. and information commentary investments. are capital rates be basis losses XX and billion. to for GAAP are income $X.XX

on Securities We in in will will a Conference participate keynote Technology be attending fireside June chat. where X, BofA the person Jensen

Our scheduled earnings to for discuss second of Wednesday, fiscal the XX. of call August results is XXXX quarter our

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

take with C.J. first our We'll Muse Instructions] [Operator you. ISI. Evercore Thank question from

Your line is open.

essentially with of taking now to business guess the last cycle. crypto would doubled got a some how potentially thinking update two The Gaming new clearing Thank the afternoon. years. for from I channel And product you here. about we've Yes, crosswinds off, you're ahead falling cycle over good an the on get love question. has

of at macro installed base supply. same about time, has But RTX under third only from challenges. talked the You we're a out the and moving

trends? once to we'd So How Gaming of here COVID we beyond about love hear kind in from July lockdown your around are thinking the get challenges you the thoughts quarter?

-- and in The broadcast. well things the The rest the the you COVID use And for we question. gamers situation the really your came Gaming thanks ways is the dynamics lockdown a it million the for robust into a The the of new solid, that China to for several friends, Russia. underlying Gaming tremendously. industry dynamics alone, expect last expanded using influencer lot connect with an platform of the industry. and fairly driving with market The PC Gaming XXX net C.J., of as Yes, two of are be the has question. intact. The are years be in that captured dynamics people to the PCs their format themselves, to is

now are communicating you of have studio. for second using Because they're home. It their from The as been never if days. way home PCs many primary second people also GeForce need their will, our working So, greater. PCs workstation, these is

so are think dynamics fundamental I And the really good.

we hard year, the governing as and great. going the Gaming -- so look And is the us. be half when in COVID nonetheless, is to of to war the exactly Russia what into of the -- predict behind we second look dynamics industry But it's

we'll Ramsay with Next, to Cowen. Matt go

Your open. is line

card-driven to performance leaps going Thank morning. be ask In with Data a around the are attached do revenue and think of to NVLink H-XXX, be you very wanted longer-term on in wanted going servers strong I there. and over the NVLink? think with Do to fairly continuing more sort the I are on there's portfolio. envision upcoming happening move? that to ask addition are the is much. and on third-party whole to what's Hopper a shifts of a be time, Good the of to sort driver industry Grace some with of chips responding of And business. you integrated move to potential you all question in cycle or that Center Hopper on though, with servers, upgrades I/O together that or how your bit your that sort to what's small to a you vertically dramatically, for Thanks. question, way, this going I a Jensun, business in huge

I the performance is what available servers next now point I/O The before. made teed are The The generation right being all up of a Yes. question. first very are that see, point. than appreciate higher the X. big substantially -- that Gen is let's was you

refresh to Gen company large cards others. of going Brand-new new you're as and drives from result X, platform course, that. pretty a so, networking And see of our a refresh.

perfectly with we're generation ramp so, into different you a And Gen timed lot of X configurations to to the There Hopper. make. system are want

the are If cloud that computing, scientific you learning one way the different processing, and take done the nature, back on-prem at edge. system data these for and look computing, for workloads at the enterprise you done of inference systems necessary that a would a dictate done architecture. hyperscale deployment step for in locations, type manage machine of and training, Each

to years So two that is allows two to there diverse else multiple did, one of support -- that generation this we two chips, for terrific it's seven next isn't to anybody PCI two which some we so the modules two each Express before fourth modules, systems. dies, size other, reasons that interconnect connect NVLink chip-to-chip innovated two we're one fits us in our that why ago, systems all, to SXM is of chips now

they're systems NVLink even installations, data are refining with very NVIDIA of retail the models to future, These themselves. you're intelligence. NVIDIA, hosting has that. example workload has see essentially that AI of are mix I made data you're and XX/X. a And are at with XX/X. where see and us to large for the of These chips, the edge and be be to these all AI that, running the XX/X, you're turning into platform systems end, data different running seeing for applications. in dies, of the running of a chip-to-chip And going at the of is are essentially that bunch that whole to They configurations. and robotics and AI. centers over centers going coherent And match with are to over recommender be starting possible then types becoming that that And on do infrastructures In stores types one going configurations. centers running new so emerging. packages, to warehouses, And running factories that warehouses, That data reason or world. more fairly really of factories HPC you're type couple factories Deep autonomous see a all a language link, large logistics major time, think it our to would now and processing are the going other data you're good important data, and

new two So just these are centers they types architectures. of and data have emerging, also different

Nintendo. kinds different in me think we going system multi data We're NVLink, We any ways. make Arm multi-chip possible semi-custom to to semi-custom us to that -- integration, is us it of the could makes our chip. think to workload multi-package NVLink. data as to could do connect better fab systems. chiplets we We'll with that that NVLink to going I serve single and running it Multisystem bring can and of multi-die do have come And computing these in a support to way module We all to xXX net as workload high-performance coherently analytics is it support integration respect done for NVlink is I open able partners. all integration, could do to the universal to because packages. vertical they that we really NVLink to be is do our single training ability that segments. So support processing it for inference to into we've all saying past, chips every With to that every to our from the including that different many capability do accelerator, we system to is Express companies PCI to of with

do a lot different system integration. there's to of ways So

Research. Bernstein Stacy with we'll Rasgon Next, to go

Your open. now is line

follow the I sequential. my Hi, on guys. to questions. taking wanted for up Thanks

million data. I know hit $XXX was said the to a $XXX you million a $XXX million to Colette, and Gaming hit So

-- mean it's hit. the in of So actually I do And I'm relative that -- than more you Gaming the down those light Gaming just constraints assuming how million. Russia guess sequentials actual doesn't of I that businesses facing? down $XXX mean the that lockdown Gaming think guys see and I you do is the about are

Sure.

Let continue into me very mean that to decline What expect both architecture, Gaming where probably QX to and probably and has in what areas QX? that We in the QX. with decline channel Gaming last from there's Ampere that still well. added believe what Gaming start first to We great to does does see our Gaming. do strength mean for our a remains been we in strong. end we just growth from many total, that But will as see will QX teens. sell-through quarter and it decline into will demand

are holding and China, which As in these some up. through us of try lockdowns we work

grow in Gaming is to So for overall, end year-over-year still strong. demand still We QX. the expect demand

Next, to Mark we'll go Lipacis Jefferies. with

Your line is open.

taking for question. thanks Hi, my

If to sometimes by talking this about there hear driven networking the their you you this data is of listen was increased spending talk that lot enterprises seems centers earnings how AI. and a it season, OEMs, them by on being about

one And you data for spending. the about, you talk wonder the you language are the versus you about of rate you you the drivers it most seeing for -- is your Are can have? inflection sequential different is could You driver, software I Is versus I your it software centers. there natural talked software it which you. if Omniverse? enterprise bunch enterprise? is big enterprise a sequentially? were packages talk you year-over-year is in what of seeing Thank And, cloud about seeing various about growth processing, or attach talked interest? know or in a

quarter. We Thanks, Data had a this Yeah. last business record Center Mark.

We amount the Because We companies about record the from insight second quarter fairly final the automation all predictive AI we're to do now record, their that for for there this most and automate to the AI it quarter, several data are about is extracting are quickly. data machine a vast of extract produce to things And lot have have half. of intelligence apply domain-specific companies been techniques created have to companies possible that and about and we make to analysis, is and data-driven to that intelligence. writing software intelligence. they most the incredibly expect a techniques companies want and do to know. insight learning it to of have and another that strategic in enthusiastic

other I learning, so And think intelligence. are driven about data experiencing that machine trend insights, analytics, data you artificial the hear people

it's described, thing. it's and about all it's same However, exactly the And sweeping just company. every industry every

Our networking is supply constrained. also business highly

and of are system, physical the a Our components. -- cables. really really, just high. network, it's many from connectors and And lot and there it complicated demand is requires Components aside chips. -- a our transceivers components and really And just it's a

best supply chain QX has We're QX and from supply expecting in QX the been QX. increasing doing increase it has and We're so to increase And been in problematic. QX. our our to and

kinds the we'll AI, respect learning machine With all, for increase there kinds are well, understanding. all in breakthrough to essentially to And advances language AI. probably the is AI robotics then really, this of recently, has grateful really a are and there that models, called of most cars, one of us, significant robotic around first really is support applications, self-driving visible really, the an -- natural two, that. able natural most component from we're software, led model so industry computer largest, from be to language Transformers which the all And vision, the understanding, speech incredible

of these thousands used in and and thousands types of There they're models. all these are AI all species industries. And models so different of different

I'll I'll that One just about of say the question software. my it answer quickly and favorite, very

it, to to synthetic of the models – around of these important chemistry my the sequence them much run and case amino acids, But to is favorites To understand understand stands lot In an a can models for cobble they chemistry of a and it be lot back to source someone using language that our genomics. engine use need AI. understand hyperscalers, of the the open need is is biology, different etcetera, proteins, really, understand that all to biology NVIDIA And AI enterprises, all using of source a their architectures AI. have. which they able operate they our recognize is inventions. base, them of transformers and really to new the One called to engines installed the it lot the AI on. together apply package and to to language And all of a of using and patterns, architecture, and language their to these of to excitement cases we Transformers to together or support different and essentially this for them anyhow, some update in of But provide it breakthrough. it refresh old understand the of for -- and engine support of

today. like recovery systems. meaning conversational able manipulate and systems really Whether systems provide But artificial just wave to so Omniverse is another directly, interacting engine, And important physical and engine for that for with or the this that intelligence systems an called like NVIDIA if has There's such. AI, engine which next it's not you And almost of has physical engine to engine is generators, heat automate and AI. that steam being where with designed intelligence. artificial it to information interact a like and is will. it's we physics It's recommender called sequel except robotics is

still in commercial I able what And think, quarters since Omniverse at XX% the that -- released we've Omniverse a evaluating. some to been that let’s so we're has this substantially about. and intersection now And – Omniverse we is Omniverse see of we're It's are for and simulation be couple customers, more have artificial enterprise. world's license. of who that's licensing sit Omniverse the at intelligence, between deployment that of early XXX is now customers, point, top already companies designed to interface,

using it think in GTC. of downloaded companies, Colette tried I that companies. kinds companies some And XXX,XXX in might applications at see the some XXX nearly been been all are interesting has times. of the highlighted It you some of it's

are NVIDIA future expect contribute And to engine, engine, our the the fully for earnings. be successful very greatly so, AI to in the I us Omniverse going that and

Vivek Arya BofA we'll Securities. to go Next, with

is line open. Your

Thanks.

destocking Just you're this new your from gentleman that to And question wanted is benefits planning outlook if you. then includes to the launch my products QX year? clarify, Colette, any for

keep Data about so You're year-on-year or growth, still worried guiding the about Center at is cyclical all as impact close XX% Data a think very in the the slowdown I I the in we of despite mind Are we growth any you to think, Center that the half about macro there year? headwinds. the second should headlines economy? like all on to all strong,

regarding as you Yes. question first any products QX. that asked new look Vivek, answer me let the we at

starting but new the that in here the We're have to the it, to will As the of about year. we QX, example, in see half of most going the discussed start calendar architectures, ramp to probably of ramp we end going be year. see, to our to we're Hopper closer of back for

So, that to Huang second of rest. you ramping our for most be over I'll in half the on part. launches year product think should Jensen the turn the about of for it

Thanks.

It's Our computing Data and cloud our universal adoption has GPUs. world now remains in you to innovations doubled is really strong Hyperscale a seeing and the accelerators. really CUDA AXXX. is demand one this AXXX is And of graphics grown year-over-year. great mentioned, revenues, the accelerator. It's strong. unique quite from really special and of significantly. GPU Core we're to as we really, and Center extended strong Tensor as of

it ETL, for could you transform so load. extract, data and for for example, processing And use

functions could we NVIDIA accelerate You Center We Data the GPUs. is accelerate accelerate accelerated a sequel it Python version on scale Spark Rapids, database Many X.X. which Pandas, acceleration. are for of use version we

on so AXXX. loading and you can data to and before extraction, processing do inference And to processing accelerated and algorithmic processing, database to do be whatever and from image need other of queries fully or data transform training you

very urgency. great to that are there. on in happening scale models we're with and several important very so large core being today, new invested at at AI great seeing success you're the is and seeing what very, closer different And

day be impossible the filtering be even our economic experience not world practically your are in every the Deep the would You Systems. have if engine, This it recommender Internet, for trillions engine the the is enjoy probably Internet system, use the things information that of about for us Recommender to to shopping Internet. phone and of to experience constantly changing three-inch able with engage heard

And a so second by incredible this all thing is thing recommender of that conversational system magic possible is call AI. made

supported bots customer one service customer service, on enhance to customer by has conversational hand. being chat now an the live service and even AI, AI the website opportunity You're seeing

foundation shortage. On thing hand, is for this But work data. models, been I it's it, X did because really of call amount we transformative groundbreaking large within it they the the much language lot AI, is human Microsoft, and embeds of to and enormous that was then think related we led computation. language are this others Turing. as model supplement And Megatron so GPD to understanding take other recently a incredible called like natural they XXXB. And with knowledge it, this of announced which breakthrough. of piece And language open Transformers an – third collaboration the like trained And labor and a it

really quite However, the model is remarkable. a that pre-trade result net is

are start-ups, the working Now we're are thousands using who with large public companies of cloud. that building

a the it's driving so And demand us of for in lot cloud. public

greater us reduced. the it's significantly XX,XXX I because service AI, and or they could things. and on-prem The saves a whether AI using that of money, startups the quality have do are think time NVIDIA better computation it we with lot cloud, now in inception is working is

the AI cloud. that's And so driving in

transformative of new AI whether driving of things the of these Data driving thousands of companies NVIDIA nature industrial models, in And recommender The different of AI. these language of systems, importance our growth. are the Center factors, cloud so cloud just the public it's -- all conversational models AI these that large all demand, the AI, recognition world are the around using

Center see strong. we Data expect demand to so remain And

Tim to go we'll Next, Arcuri with UBS.

Your line is open.

question are issues, so Thank absence lower it's were supply sense sort is particular, Colette a the you July related. sell-through issues, for sounds Russia, demand in this gaming you whether you using me most in that new about China but And language of of others stuff, of would $XXX on your the make this the sort more more of that because a more freeze hiring have. in this sales very or side much. context to in and logistics commentary in of supply and little million impact which in more demand setting I a of that's related had a semis

never and related, I So, ask perishable related it's and a might be to because if of beginning argue But it that really timing. you supply falling could demand the that then come back just it's not night. if

through you me? if So, I of sort wonder walk Thanks. can for that

Tim, for the question. Thanks,

very the China try things. The me are here and bet interestingly Let to demand. and current on and implications supply both different China has lockdowns seeing that Russia, two we

going demand pressure. seen challenges were a already been under head the in pressure logistics the has side. perspective, of have the We From logistics the throughout of that terms It a puts from in gaming also of just country, things lot country. it out on

there. on other the very that really You for cities are large things focusing important citizens lockdown, in have full

of very for do demand do supply they it how. it's to sort will our our So, and occur. believe out. come out products determine will that come We will demand. difficult the believe impacting that back. We The COVID will It's

little we been and selling company hope historically helped. case last that to probably X% when you Russia. the at percentage had revenue company Russia, of part Now, Russia of Gaming announced quarter. of has our overall look But plans of our not about a business. a there something revenue in our and we're larger earlier That's were that I

to go we'll Next, Ambrish Srivastava with BMO.

open. now is Your line

Jensen. that it feels those to to much refreshing and from so the demand called question demand out two that. actually on Data relates healthy Gaming. you second Center really and Hi. appreciate as problem, it is well a I companies, it's Thank I very a Colette both hear half to you like it say had as

Center So, you you better. made last never that your couple has into Data comments publicly, the talked been times have have visibility of

then business guide return Thank that getting but true, So, take the say a be to of second want the you made you. up half just Russia I have growth like is we momentum. see quarter? QX. not on you just do there was the all make will you confidence still intact could for sequentially orders may have strong you that expect second stays? Gaming, sure And year-over-year I to It just wondering to Colette, based that if in statement and up on seems it been year-over-year out impact, did that be

of us companies without And the to do visibility quite impossible. without generated data either with years quite and that's having just couple and time. Yes. vastly with We things data And now CSPs case for there understanding the learning not most -- accelerate of that it in to of even Ambrish, do Omniverse is helps of vastly recall unsupervised was as that all deep the hyperscalers. ago, synthetic you is human-labeled have deep having have three data, is data. And AI then for world customers computer On reasons the science first requires to the else, was with nascent. industrial question. better, about label Obviously, really years zero-shot still artificial couple that a kinds but for Transformers, to intelligence use the type reason for the too our that. and other or, Centers oftentimes difficult. technology The costly starting allows learning cases of in everywhere was of which a the data several. a thanks pervasive be labeling couple of at interesting learning have you techniques, frankly, ago. is quite generation better should that principles, Data And two, One, it than if

now, a many transformative. the the industries technology way evolved to of and that most could use rather so and at the intelligence in fairly industries knowledge And artificial effective has a place

And so I one, number industries. think, went hyperscalers clouds all to we and from of

the can. requirements, so harder. these because inference are there a inference Second, you by reason of out inference. be many quality run we different the use and Most was are thought to easy. to is And inference you so for a the to far that is models many from there and training-focused many models went It service going and cases as people different so want turns small footprint of in that

number scale when the so you is the out, really use quite high. of service users And that

using is And it. NVIDIA's cost acceleration, using we chemistry to model speech any fast you to So from the do vision very it and platform, that so biology, could name we to so quickly computer inference low.

you And that, think, you true. And absolutely so the that more save. do, will wisdom I money is more acceleration the

is so second inference. dimension training to And the

that of is that cloud types computing about then systems industrial dimension systems many concept of configurations really us from to AI causing And deployment all find different the of we we now can be go so high-performance every to on-prem able growth. way is third have just final now The to edge. this to that's to in industry, the

and cloud growing quickly. are hyperscalers very, our know, you as so And very

industries, and vertical retail of nicely. telco services vertical part, financial grown which and also However, all the very and vertical is very, industries have those

the And visibility skills able of dimensions, able then have up the a adding starting portfolio to company, couple in depth to a a those years don't technical end-to-end be solutions don't our lot ago, sophisticated be have And provide different systems. all more to companies to we're our of should Mellanox solution-oriented better. for so, these the and platform stand lot that

as business well. so, growing our networking very, is nicely very And

Next, to we'll JPMorgan. go Harlan Sur with

is open. line Your

afternoon. question. good bit ask question maybe just Hi, want just let a me little Thanks the directly. more this I to for ask

second right? from the July demand spending accelerating. about with half think impact we Russia, actually, I so, in the to in supply is the to So, some the right? here And in navigate it's sequential data strong, drive even being cloud You look it's of able April, dynamic chain growth quarter, environment, see team – center good the year, and as strong think,

more ready getting the think think, And general, guys sequential you ramp team You're I Mellanox, uncertainty through expect revenue business sequential this understand in you year. remainder is does the the around anticipating getting HXXX growth of the entire the year. and through as previously, I continued to but center year? the supply later growth in supply through were I the data move year. for gaming, in

is Either demand company Ampere every We the one It's computing to data to a of world. scale answers single in answer out. in vertical those continue center, to hyperscale in going see industries. cloud strong been yes. is qualified the -- to

layer the And markets. to planet, different so, in accelerator best different brand-new the going to on We're after these two years, on to universal architecture and continue it's of remains and it out domains scale going that, all Hopper. top a

Quantum to that X on X, supply. going networking architectures. top brand-new of increasing CX-X, We’re and layer BlueField we have

next excellent And into so, an looking going forward second quarter to for centers half. quarter again we're and data the

Chris to go with we'll Next, Caso James. Raymond

line Your is open.

of And quarter. up function could Yes, to they – how thank obligations, you. obligations speak just a little magnitude general. maybe obligations? the you that that longer-dated of seemed Wonder if in purchase a supply in which speak the a or was about again constraints you like higher were bit could And

in of the You've a about What times the constrained? business. still the Where business? parts other in you are continued of about mentioned couple constraints call, the networking

Yes.

wants to if I'll Jensen let add more of So start see that. me and here,

sure Our we prepaying Center long obligations, things where are for in demand of Areas many both constrained of and we we good well to ever, a major two complex additionally, to that systems we that much upcoming procuring which in are for can as in supply imagine, sure term. to further. to products feed time long lead first larger can A the for have market. make the bit items little Data our make and make and prepaid of purchase keep we sure quarters them our those commitments commitments still obligations, that are is coming purchase networking. us much you the terms are purchase we And a the those on are have One, must percentage to things time more have business, our systems, are for procure supply as that our to mind. that are

have supply still there still. on, add that networking to quite been want with concerns me, it we improving strong. others Is time. demand Jensen? yes, Our is But excuse each demand We've -- you

you No, I perfect. That's were perfect. thought

comes with from Aaron Wells Our final question Rakers Fargo.

Your line is open.

for the Data and thanks Yes, But questions me Gaming around my ask about answered. guess have I'll And most segment. I been fitting Auto of Center in.

small, business as that next us think see couple the start half past, guys think, wondering that, confident While of the kind of that move about I'm sound higher trajectory that help growth" "significant could second the we inflect if you to next the sequential clearly, I into to should said this it's in of of the into really still in And you quarter. you've business over year. quarters? starting

help if that about business? us curious you Just of think could the piece

Several data points.

We are just starting.

in those be designed as is because Center information, designed ability sensor be secure, O-RAN a chip safe, a on rather to resilient a chip. the And all things We've the been equipment it's last-mile to number if O-RAN everywhere. is of more O-RAN that the robotics of Data going designed these centers is has It fantastic. it robotic logistics production to to It's and processes will, trucks quarter or computing. into O-RAN. as really taxi it's be is revolutionary shipping logistics We a and robo car and first if or have others, XX design movers car designed include systems robotic software-defined wins is farming a Center be you it is confidential and And Data and for the a data delivery you just processor. a, shipping for companies a are and mover, robotic first quite placer robotic, pick started and designed processor. mover. O-RAN on

production. because a that estimated. for marvels time. We years-or-so, and the we the time next $XX over very experienced for And business auto likely that And of we've reason is is secured in really six some so lowest counting billion O-RAN the some quarter in have technological

It's you going for Data and next that but safe is and there. autonomous so segment. multibillion robotics autonomous say next its autonomous major the or machines, physical major on robotics to way move our be I We've that is vehicle been And know, surely systems think to this edge they are surely and systems now whether next The computing to segment. and fairly the thing not as it's area, in a going surely be O-RAN be our decade. a dollar at It the business. Center to move, working AI is going business

area. fair a expertise example is in amount of one have We O-RAN And this our work here. just of

which for as the pillars four in system, strategy the from one and our to memory robotics Starting is O-RAN have actual robotics of part and autonomous four of then But robotics AIs; the to of mapping; and and known AI. second, systems. the the pillars We wave our to strategy omniverse; finally, and training goes. just that's AI robotics robotics data O-RAN our to processing simulate it, the train AIs, application the next third, robotics of of where otherwise processor the is AI

of about next industry's the phase enthusiastic the really And optimistic be to at and I about edge. going the And of to lot of lot going robotics. growth. a computer so I it really am A it's think is be

any I'll it now Huang additional over remarks. closing Jensen turn you. to for Thank back

in difficult full The and China Ukraine and Thanks, lockdowns predict. of duration the everyone. war is COVID in impact to

discovery. that transformer impact from model, spatial, across being stand. and deep great deep Researchers AI complexity. language for revolutionizing unchanged. the companies focused to adopt opportunities models advanced led temporal natural patterns which effectiveness breakthroughs creating We're The effectiveness major robotics AI understanding market are transformer our technology the computing. driving and of The to to industries sequential are However, remain continues learning to four learn of initiatives. The is with of is our drug applications to on learning AI NVIDIA

customers Hopper next Networking. infrastructure our and ramping and GPU, take InfiniBand, breakthroughs chips generation their advantage factories new BlueField AI to First, build DPU, transformers. Quantum of Ethernet like platforms, AI And AI this Spectrum InfiniBand, help all of NVLink,

CPU. partners ramping Second, industry our first launch software and our Grace, to system

We taxis, robo to Third, Omniverse to trucks, fourth, delivery And our robots, this instruments. progress adding every markets CUDA and value with greatly our continuing while scientists with building every farming to medical with extend AI. you time you. and software robots industry. and robotics robots, acceleration of libraries. ecosystem advance expanding initiatives Thank our platforms, NVIDIA processor cars, logistics most impactful new AI to to customers NVIDIA technology nearly our into updating field next our and in new forward ramping new O-RAN, new in These And XX on look will quarter. companies

conference concludes today's call. This

You may now disconnect.