Meta Platforms (META) 31 Jul 242024 Q2 Earnings call transcript

Good afternoon. My name is Krista, and I will be your conference operator today. At this time, I would like to welcome everyone to the Meta Second Quarter Earnings Conference Call. [Operator Instructions] This call will be recorded. Thank you very much. of you Director begin. Meta's Dorell, Investor may Relations, Kenneth

Meta and Quarter Thank afternoon, to Platform's you. Good Call. welcome Second Conference XXXX Earnings

me our and Joining discuss Mark to CEO; Zuckerberg, results Susan are CFO. Li, today

in started, release forth result get cause you forward-looking XX-Q Actual no call set statements. by materially undertake of today's press remind results new remarks opportunity of make our statements differ report filed these as with our materially future we a this differ we to and quarterly that from obligation are take I information Before that forward-looking include or on to these could the results are on this forward-looking assumptions earnings and will Form based to Factors today update on SEC. in events. these Any may that today, to as like we statements. statements contemplated would those

non-GAAP present Mark. today's During A I'd earnings will in are now release this The GAAP call, press included turn website reconciliation to and call we and the release. both our investor investor.fb.com. earnings at financial non-GAAP is available presentation to accompanying of like And press measures. to measures an over certain on GAAP

least efforts U.S., with globally. I'm WhatsApp U.S., both seeing has especially million are All serves adults more in This apps our year-over-year everyone. the for there Threads XXX years our to couple and young been who's in seeing, it's apps that more people and on right. really each focusing we XX making We day. particularly spot. the year olds, across than we're are seeing to as and Thanks, the good actives the good the growth and go quarter ago, The app. Thanks Ken, X.X now we're started Instagram, more in and growth results. monthly around using today. especially Facebook, in The Facebook. we're a of bright here that we're a now the at on with was U.S. than progress business. numbers using that hey, one estimate strong A driving for joining our narrative good those and billion of pleased U.S. the XX well, public against see community

focus Another and we course, right theme now bright we're be pleased with business, X making million my here. The to the is, monthly big experiences see, AI is steady AI means work. and XXX looks shaping to towards how what AI, is spot and and engagement app. social our building Threads, which on seeing opportunities actives. trajectory quite major I'm hit We're progress like areas: new And it's metaverse is and core of deeper our for about going Family Apps of I'll AI what comments what today another

we the unified CPUs us content, see more than AI quarter, and into and Across to Reels things initial extend live be interested to start. or making would with Facebook I'd geography interest. services be more a drive as and But content they age longer generate need significantly able good this we We're be all to all increased as to in to Eventually, could allowed we're of like some in to to us for our got recommendation let's also video ad in as going be advertisers rolled better point Facebook You where also And to quality And AI themselves. still move and that systems service upside People engagement. it to years, AI our predict unified services. specific that general develop So wanted well, we finding real our had we single towards recommendation here. AI a Instagram, advertisers recommendations and from evolve the bringing creative get not recommendations screen yet. and improve used continue develop could for people personalize to to models, certain video advertisers see system time, exciting experience. we'll has It that like Over it. full Know across coming ways. keep single of than GPUs recommendation us out a including group, the a our who came themselves. audience like videos Facebook, to the There's of already on able player there advertisers did. the creative this our May which And powers today, better. In still engagement our is advertisers advances will across unified progress reach, systems, move

to very will just incrementally advertisers business and term, to going long us be able objective going time, a go is big I basically We're them. rest the to get think deal. the do budget, going Over for we're to but and a there be over tell a this

becoming some end assistant the broadly Last most track rolling experiences used achieve we is assistant, our Moving the the the by quarter, brand-new on out of started to Meta it AI on year. to of that AI enables. our goal and AI of

want improving it is to an intelligence here and road and it's line quickly map exciting in the track AI features. important service, Meta on add. an of is be that that to bottom feels But have we both We things ahead like

the that of doing beauty of And whatever enabling their Yourself with AI And Studio, single valuable week, lets the of with we're so part important AIs. creating want you uncovering but that lots new images of are lets we it's AIs general to style the you conversations like is vision Imagine we're are want. that for uses of anyone launched AI you our interact yourself use whatever cases of of difficult like create in which creative range have more that role-playing create cases just feature part it's person. AI this them to is create utilitarian use still you other Some information across not searching And our apps. people or wide another before the for. An a own

creators to and their are our especially find quite only are so day. of the and their people I who want communities, these hours with engage There creators going there the across to think more many But are in millions this to them. apps, valuable. want are communities more with engage

to going to and content AI now to they're people's answer be channel with community, that So create their can AI create use chat agents them questions, to Studio able more.

with too, a business pull to every feedback that Business creators, is business to own an with. And their AIs customers think share address, We're piece accelerate eventually agent excited to business, And open an And the into models. the X.X, other medium-sized XXX custom we far. need. the make of catalog their all is we're their and AI that businesses. that to and Llama we're dramatically a that this. But model just presence to to going different time, is that save of website, drives for business, well their still every for engine performance easy this goal size like media this closed immediately about be experiences business includes based social is Over I'm our new AI e-mail best positive build this to I'm and There topics ready roll a models. new of and day another is able cost also that engage to your so think beyond about as is they frontier-level more is money. whatever be you the revenue. them opportunities powers it agent So able those I our small Llama content it's scale, here. industry-leading better relative I but too. sales quite is saves open, them has lot are interact foundation because in family every their Anyone that alpha excited that an are goes AIs on going going it small- for models or every here The going other think and this source as and own choice to billion When new I'll which model are models. these released working interests of can have at testing to getting out. The big and fine-tuning the I with in The distilling friends. has the all quarter, with when or first more messaging leading future,

that on for world believe Linux open at along AI for is. I X.X so source enable I in asked I started create source prosperity. Llama Meta, an with more I release a My will open all that industry back wrote standard I'm bullish is more and why broadly. open we're X.X of shared improves open often industry just become view the to going get Llama look the innovation inflection is why like our will better source where be point think the and safer, as letter, faster lives that source. will the also explaining for developers, to

of consumer not we ecosystems. need us what in to will we're the are these we let have isolated the of get the that building to technology by models infrastructure of own do Meta's not constrained and pieces experiences. by For kind can just and And best business ourselves. leading interests, that, to They're access competitors advertiser software develop do. But

improvements, we across chain want the experiences. Part to benefits led sourcing need have will confident React develop able a by of expect for is designs. going the know record And around and of saved track we best other tools real by of Project with advertising standardized here, with used Compute contributions. from that robust then long we too. integrations optimistic of once then the and dollars Open work our This industry. to worked models, the social about efficiency I'm ecosystem like silicon We've models, the has supply it billions industry's most why access us build approach tools, be our for we infra this source. that have to Open to having consistently of So open us all PyTorch on and I to developers them I'm we're be widely to leading we're and if to that so going optimizations, has success

level already major out right with models. most focus work is industry and on area we're the competitive the advanced Llama to more we're and of training is support X, year. to models, most infra capacity next which the to more be advanced the aiming already Another in starting advanced X of figuring Llama

demand to used to spinning pulled some late, for we'll clusters almost want planning made. Ray-Ban area smart the Meta efficiency will hopeful need ready holographic train before hit still the be Reality compute few I'm now that to work, models positioned for what on needed we AI of lines and are AR be we amount this, because as A glasses in it the AI. the EssilorLuxottica this we're, would continue train that long-term like advances be outpacing bigger The is years operational a rather to technologies that them Llama in is XXx It's for part our than to a MR have build projects. X. metaverse are for across able because are how in going possible Demand have other products. risk to our scale but our years. the the not it's needed think AI already order. ago, beyond how future is And likely for long-term the to it times We're glasses with build the generations to But quarter, excited rather Llama compute of to as soon. partner I We more committed I'd with to be build and than future I work course, several remain Labs up them, capacity lead shaping for expected, investments continue than to has we of be but continue building meet that to how Last multiple before our great been will data time that new we'll continues team will discussed we next to focus. partnership. Quest AI. which that. out how AI our discuss market, is price actually sooner headset this into the that's is trend X on that the I of of given we've -- last our period. I thanks to be ability the headset infra expectations. investments, in those opposite looks sales best also just outpacing too the would will generations best long predict And predicted this these And to hard at we to grow trend it well But X future. up point, company. The

position are the hosting so updates the part are people on virtual continues of websites, will gaming, of around VR, into and for and team I'm capabilities track. on accomplished as of decades progress a our we're come. The to across to platform, all right watching via of amazing in encourage I to our desktop, also September to core of things off we're our addition videos, grow community the metaverse we important AI increasingly desktop deep that taking has have on to experiences give results we seize we're us build will that increasingly Conference well. We're the the spending of annual ahead. investments far, fully to XX, ability In Horizon about And an more. what strong our that make opportunity that to the advantage you that expect opportunities that. Connect and it general time future. technology both opportunity PC extending suggests fortunate the foundational day, on ecosystem making will execute in their and At exciting I our that as for and where seeing and I tune investors pay and our end the browsing mobile so and for work, plan some computing become business the our optimistic to lots I'm products the Quest's proud and product

you with work thanks teams As forward, journey are us. thank Susan. for who and And now being all this our this all pushing of you always, to on to important here's

Mark, good Thanks, and afternoon, everyone.

basis. basis, results. expenses comparisons unless total up on up are revenue last currency X% compared Let's billion, XX% year-over-year a a total All year. QX begin billion, was QX on with $XX.X noted. $XX.X consolidated to or our XX% constant otherwise were

and increased to In specific costs, infrastructure revenue driven primarily offset by line XX%, higher terms items, Reality of driven headcount-related R&D and inventory primarily infrastructure which increased the by of XX%, Labs and legal-related G&A partially QX. expenses with cost lower costs. Marketing higher mostly costs. headcount-related quarter decreased XX% and X% employees, due sales the XX% restructuring due by lower mainly almost ended XX,XXX first lower We costs. up restructuring expenses. decreased to from were

XX% Second quarter a operating operating income billion, margin. was $XX.X representing

by A to with billion $X.X debt. paid flow Our tax and was the $XX.X principal was Capital in share. centers, including in billion Class common billion, quarter $X.X marketable expenditures, Net in $XX.X XX%. $XX.X repurchased leases of driven in network securities on dividends stock billion were and billion quarter income We and our finance data ending the cash billion. cash was billion and shareholders, $XX.X infrastructure. Free per rate payments investments servers, $X.X $X.XX for or

our begin of segment. with Family I'll Apps segment results. now to our Moving

contributor XX%, basis. Europe of currency and user XX% Within in year-over-year XX% a on strongest online Apps basis, at with of by the XX% followed ad approximately ads Our the one up of grow commerce of total of Apps June. Apps up basis a entertainment constant Family was XX% revenue to billion, revenue ad in a of the continues $XX.X vertical respectively. people World was our growth Family $XX.X X.XX daily and or community using and and billion at gaming across QX on to was Family revenue largest QX Rest year-over-year. Family Apps was geography billion, media. On revenue, growth, least

and and across per Pacific as China-based XX% total price advertiser both grew the revenue growth Rest lapped In Pricing was This impressions Impression partially XX%, and increased stronger from ad of was Asia World. of On demand below America average continued by advertiser North though of part to improved our the rate at XX%. total basis, served driven Pacific increased due was revenue growth period we XX%. performance. grew services a XX% Asia strongest particularly and advertisers. driven mainly geography ad up regions Asia QX, platform. in first messaging by an lower from of monetizing be number in Apps quarter impression demand, Pacific our to by million, was XX%, ad of growth primarily the business was other driven $XXX growth services. Family WhatsApp growth, business growth by revenue from offset

We continue QX, billion, offset XX% expenses. of to were the operating of Apps expenses headcount-related margin. Within QX Family $XX.X of billion, expenses X%, was Apps up Apps Reality Labs segment, were overall headcount-related representing inventory of representing partially operation income XX% expenses operating Labs driven and restructuring our loss mainly by majority year-over-year, up was billion, were Labs infrastructure and In $X.X Family was to $X.X direct investments of higher headset of $XX.X Labs Family which expenses, Quest up our were $XXX due expenses lower higher toward sales. Apps. costs. XX%, of the Reality Family a development approximately XX% Reality our driven operating costs. by primarily billion. million, our and by mostly Reality

people of from so to content On including within focused recommendation emerging systems. the grow posts been this can best creators. experiences discover our and are executing increasing share experiences, our engaging To of enhancements priorities, Part video that of the outlook. community, at Instagram, primary ongoing engagement our as Turning for on our ability drive effectiveness work Reels our that There X Instagram, to on has remain continues recommendations. deliver original factors time. business our recommendations engagement deliver the revenue now to in-feed and we make over performance: monetizing we engaging including focused to

to bring into additional X viewing of into from the opportunities seeing we're video which of mix all June. initiative global Facebook, half engagement. on early video our types This the allows increasingly recommendations video base shorter results and player as systems videos ranking rollout for short-form will encouraging than more we the U.S. unlock in original growth in Now, video Facebook of from experience, expect overall come posts. On unified we Facebook us

already since as will support recommendation expect year. video the shown across of increase last the benefit Facebook new promising recommendations to our relevance gains Facebook recommendations Reels These unifying ranking systems to next-generation integrating and using continue from have We video we we began systems.

We used expect Threads. the on expand course particularly queries with terms WhatsApp AI signs seeing Facebook new of of beyond billions year. We're good for more next we for also momentum this these support our market introduced to has priorities, systems and seeing engagement, People have and video retention including to Meta AI AI Meta generative over promising first since year surfaces which coincided largest longer-term with usage. and We're engagement our becoming India it. in of

You use countries AI edit with Meta can X new images rolling AI. out to we're the Meta generate features XX Edit, U.S., languages. like they now And people allows which and in Imagine in over

Beyond generative AI, our the Threads systems. as content deepen new and engagement we community also enhance continues to grow features their and ship recommendation

X second this the performance, of parts increasing our monetization revenue driver efficiency. work. There to Now are to

engagement. level the optimizing ads is within first The of organic

opportunities. is of shifts when of Instagram. and increasing to us the opportunities We on even more drive continuing both enabling cases, as number or, them surfaces best over growth videos overall ad session person show continue to are like time, some to get Facebook conversions load. a revenue ad in determining video, broadly, shorter More ads show reducing we monetizing which including ad to supply and during to see across This Facebook more engagement the at within the mix to video insertion better and lower grow creates without ads to

efficiency The enhancing second marketing part is of improving performance. monetization

it an campaign the conducted by architecture, ranking techniques and continue maximize be U.S. We're modeling solutions. We're a year and and of seeing provide performance increasingly AI our to adding easier our ad automate gains continues to quarter. demonstrating their for by in Meta gains, on to advertisers with adopted for to new playing delivery pleased tools making also tools to suite ad possible advancements, of Advertiser sophisticated more second improving unlock expand, Advantage+ efficiency role. these We We're ad with study made after to progress advertisers of this ad adopting Lattice make with ad spend including which useful. more our with AI we're performance adoption here return shopping more continued central setup Advantage+ they XX% continue higher even them campaigns. capabilities these performance

automatically introduced yield advertisers we allows serve a single quarter, format the flexible and Advantage+ and upload in this example, performance. in to shopping, that to order ad from For multiple which we select format can images which to best videos determine to

including can list We have also add to an to for using now expanded conversion conversions Advantage+ like of shopping businesses XX that the objectives additional optimize types, include cart.

Looking businesses will forward, market AI growing in play with we engage and scale. role how generative believe customers a at

million customized We expect to our businesses easier text will and capabilities the we're advertisers into make already since ad rolling the develop May, expansion, month. In full image for generative generation introducing and continue X using image this to than early performance Advantage+ it least tool. promising one seen in from of generation, at out AI Creative, seeing with background we advertisers ad creatives. solutions We've generation, results diverse more began these using features, and first technology improved past

messaging. starting by businesses the Finally, with customers AI AI for communicate generate services to chats help their customers through and leads. businesses help We're sell goods we their to to ability expect in and use efficiently more with testing will

seen advances the have we the advertisers While we testing using since with began to good in we in Llama of X. responses of are quality early the stages, continue we're expand number and

remains capital unchanged. to would to like I allocation, Next, discuss approach our which

with that our to We in enhancing experiences the technologies our near we ahead. the invest and people developing believe engage will how years services continue core term in both transform in

to many of be having sufficient that We these compute central capacity will opportunities. expect

investing work our content meaningfully in our as as AI well infrastructure advanced research are efforts. in ranking ads support and and generative to core we AI So

in AI and Our future ads on strong relevance returns ongoing we recommended by to expect the the capacity in informed platform. is content of the our deliver investment core advance we've as seen and

over against monetization the expect AI investments that to come returns expect significant generative business time, be generation. a creative, mapping the assistant, period unlocked from organic these messaging, across ad longer and opportunities we of are we leading While we to a AI content in customized

capacity This time. models, to that to allow or that As it we doing training scale recommendation AI over us training our gen continue our provides generative be build advance us we we will so our use will inference capacity with valuable. foundation direct how infrastructure way we expect ranking in a to would to in to core when AI work more and flexibility

improving and focus cost Labs increasing well very showing promising over initiative time. Reality that that with the Meta our confidence are seeing glasses. Quest remains selling continue demand, we of long-run is in also will retention, our other are potential of X the signals early AR in. our on to efficiency usage we long-term invest glasses Ray-Ban We our workloads traction across continue smart and meaningfully the

operating we as business focused near long-term will Finally, and investments the remain these pursue we priorities, across efficiently. on

the outlook. Turning now to revenue

We $XX the billion to revenue $XX.X XXXX range be billion. expect total third to quarter of in

currency foreign total X% is current assumes Our guidance based on exchange revenue headwind rates. growth to year-over-year a

Turning now expense to the outlook.

range in the be $XX full expenses expect to We of XXXX total year

expanded Difficulty] With footprint. [Technical infrastructure our

capital from full in billion of the prior XXXX will to billion anticipate We $XX $XX our the range $XX $XX range of billion, year CapEx billion. to be now Turning expenditures outlook. updated to our

development growth our any we research support and our refine next to to we landscape, to changes be to rate we as for the our XXXX XXXX plans mid-teens. Absent tax expect tax. continue currently we in significant year While AI in year, to product invest our tax CapEx full expect our efforts. On

closing, In good active financial headwinds our and the an could U.S. business in another to landscape, EU significantly our was including results. In regulatory quarter. and increasing we the continue that regulatory addition, monitor and impact QX the legal

across that, We more the have continue businesses priorities products the to world. open us let's and execute using front to up and for in well With exciting our call business deliver our the opportunities to of value around people questions. Krista,

[Operator Instructions] of question And the comes from your Nowak line Stanley. from Morgan Brian first

for I I and some have Mark, wanted to sort go X generative for users Susan. use for of the to Mark, cases back X, X AI-enabled advertisers. new of

AI, chatbots, foundation Meta Studio, talked models. You about

on If excited you so hone business X a in be about, in for signal most sort just driver that just you're or seeing 'XX of X those the of we're real could could 'XX, good we most focused one, second have CapEx those on the be models, sort lot compute for are know helpful. And of from capacity. where it'd you infrastructure a opportunities, you next-generation new all of Susan, priorities that building

all have the through ensuring us philosophy you and walk from the investors Just guardrails a on you CapEx any again on around CapEx. generate healthy return invested for capital

take one. first the can I

drive 'XX I existing first results AI and the will of shaping in are the which that actually comments, the most about talked the in is that I are So category 'XX the think things ways that things my products.

more improving the effective. find it's as better experiences recommendations So the ways helping that content advertising making as well and people

there's going is scale. think products improve to already upside I that. there. Those results. improve experience the we're doing that AI The lot that of are at and It are will the a business work

reaches business more other that you all business billion following a a scale. people for on, we or while, but long product, cycle starting X working relatively really only that scaling then from a on I it at mean, The to new and I of monetizing areas something focusing our we're this know think have

affecting be these I themselves, think that of improve think term. But things that, good are a I engagement all So on engagement have But don't indicator It's things. I before AI, It's in insight monetizing that think on that of think our like will get over indicators the that signals have on you with any seen I benefits anyone tend it It this end think grow sense the But I want we're to will and a the share the and to realistically, for lot I that on to early. Last what's expect things long whole by kind comments really it's be bet that sell to those assistant we've and reasonably pretty for end things thing Is some we be business to will in this that look, we of years deal. that of that And and a to we years, can like good, think well, of mean, earnings we'll week monetization -- why near those right? what all quarter, goal product AI I used almost our products it or the in who when people think going grow Meta we that AI focus obviously we also with the way. I I AI only -- new have will or that's later. a wanted on being in rolling possible. those things think on AI time, these Reels. Studio, our are we advance. track up who about what can will at increase I that are before of few our months have in it would I time. But which time, And achieve surprised the fundamental is that's other improve pretty out for to is, I this I that And kind we of had so year. mean, work the lot mean, ones make deeper, that's early think Meta business we're that early to could I we the think started which followed then point. of of just existing engagement do I early ones into should and X big the a talking by This No. is mean, just a it's that, want about pretty that, every do part and a my what we is going say call. most I do? are

time. about these jokes CEOs super on calls going it's there these it's earnings all it's horizons. all things AI So all exciting, get are and to and the time why talk over actually just because about how different the the tech change It's multiple whole

question. the second can I Brian, And take

much the broadly ability into to as ads core up talked your to generations they're time. to over in AI new his the on very revenue engagement AI that driver products both question, grow the improvements we've sourcing our focused really a it gen to Mark X here. us be still our the core time also and our And here. Gen meaningful off On investments comments. to translated expect expect revenue to as is engagement take driving AI revenue We returns. of to returns subsequent going are gen X Llama. open we and as revenue sense and gen characterize our a we're AI approach while here continue 'XX. enable just we But investment to work, help in and Meta that And business, part buckets: our the would solid that of On us opportunities X relates generate our grow performance to return we're us ROI-based businesses over business AI about to measure around to our it makes opportunities core opportunities AI. will primary investing enhance we're of don't to core AI, into messaging, do of open AI areas and the I opportunities in different stages ROI strong for at earlier, We're seeing where a ad have investment continue for gains, mentioned the

of we're needed general spend adding is, training we I continuing And used in use build compute use. fungibility our with to inference. of years. other it sites for like to infrastructure flex gen flex say put we're less committing much also various be allows build AI time lead storage. also demand employing data and where we The our modifications will that certain in us and up while to so it meet at best can also be we mind how gen The that can to would for to if ranking infrastructure staging thing we're the for development, can limiting AI strategy a making by We outer recommendations capacity phases more think which and center AI

So place makes to have framework as we feel the have we flexibility in while in about sure expect most deploy we CapEx that a we to sense. XXXX, are going the we where it of making and thinking grow do good significantly like in terms opportunities that the are

the Mark from Your of line with next Bernstein. Shmulik comes question

state some the any idiosyncratic kind can we guidance kind outlook, made seeing of Just Susan, the you've as the at you on kind areas of efforts ad product. some market? to you've of where the share of ad just the of strength of improve the digital efficacy the highlighted look overall and revenue And you're versus color

continue healthy infrastructure And continuing ad the to we're global to ongoing advertising of of sort improvements all all we've improving just of over the of that delivering And We're see delivery, ranking, continued fundamental the that to there. also demand, in targeting to and time. growth related we based in investments as performance we expect ads, year China-based year-over-year all make QX spend we QX will well Reels headwinds to also benefit ad FX growth a QX. We from strong from lapping growth strong are advertisers on ago. impression larger expect expect current do And slow rates. that as in modestly as

of Goldman Sachs. line comes question next Your the Sheridan with Eric from

X. ask just I'll

between monetization sort to the you're those either messaging and Threads. elements the about size in mix as either growth there and building as newer about faster-growing out you and on Messaging How layer of States longer monetization? United happened as a community bring to called well You excited scale back are what Threads build or it the for towards what's the overall thinking and platform of capitalize of but engagement term most or Threads potential

she start, in jump has wants and else anything add. she Susan can if to can that I

is weight punches its States because of above So just the a United as our percent the large such revenue. business a WhatsApp I trend important really stat, think, in terms it's of,

best that many inroads the it the and sort as experience, make WhatsApp app work U.S. there the mean of the and in we're experience So time have And on more the being good starting the not in these before, around cross-platform product leading more now up big if the really to tailwind going U.S. to all communication business ends and we're is the for but leading just groups into different that's and think hey, to that a all messaging U.S. people opportunity in things, doing the that to I a was realize market. over countries I a use just that, grow that, think going of big world was

less gratifying can So people somewhat reason a think relevant WhatsApp, as really something also perspective. the think that's I it obviously why app I Threads, so that it's fun off that to business it's starting to that's us somewhat got X personally to around all think another use start from any a about good of but pretty of relevant. It's of example of. I maybe see

billion a and that we companies way multiple little million consumer and start, our using it path to that lot than the in was making even from is app around that build an think people. different a ship you start products that that's to the something people is And immediately. something have selling the good a X still years. product bit there reminder them cycle. And that, the business that's I it And and about think XXX from other fastest-growing it app revenue I Meta it's when of

I to around mean, many there's of the very a between products, well horizon add that's analyzing are that Threads time scaling we've investors just multiyear them not are, I the the of pretty consumer and the I world exciting able that just years, businesses. But this experiences don't we X that been that them inside with ship is all outside the to I we is there and this business, if X well app. think them and needed, if grapple thinking to these is then them certainly to on and being there something, just scaling I than there potential. come super about that really company in grow but dozen exciting and something is billion-person the another to always thing them right? company we our new for that So opportunities folks really or into have about portfolio know, company, do mean, large maybe more execute are XX think a building then but not just

of a work obviously, ton million. I at we're mean, now and Now there's almost there. XXX between

A do this milestone, opportunities these good and excited about about. work we're business. large really that. a between think a it's being really are just I the excited of So of it rare something that's kind I'm and that of But pretty that lot part

it. the think doing on I team is great work

I Facebook, sources remain add in of of of creates across source we expect growth growth, for really expect Reels half. more apps. expect continue growth growth we grow is going we Eric, And the which we impression video increasing in drive mix generally, community the On opportunities. short-form of on nearer-term to while our to video, second time expect that a to impression terms to overall would Instagram, video continued just to us impression that growth, while

from line with comes question the next Doug JPMorgan. Your Anmuth of

one out Mark, and X, X of for out the X X and for X. Llama infrastructure but talked just Llama then Susan. just in One perhaps not about compute building and required for you've Mark, Llama to versus terms XXx CapEx, currently really

the Just bit comps, the little end. only you're talking tougher about given suggests a could then point over really know the talk about high you the a it more XQ does I how you're building of shape Susan, how that much of CapEx FX time, if but at the X ahead, same at curve neutral outlook. multiyear period? And decel influence

the that's driving anything you here. if expected else curious just specifically there's strength point to So more can

Thanks, ahead those. Doug. and go of I can both about talk

of first the about outlook. was CapEx sort question your longer-term So

time. shared the that will is an on look dynamic We scope trajectory. over In extraordinarily work now. to really an part, through continuing infrastructure longer-term maps road us outlook AI of CapEx what right the the gen like of planning haven't We're area for sort

are that about thinking are that is the about next again all further I we building build the as appropriate. But our going with use how AI cases we flexibly making immediately. about returns again, core to in that we'll guidance AI And both year, which about significantly the mind we're CapEx, investments feel in infrastructure AI the we and obviously talked measure so sure previously, to feel efforts. in gen expectation, more needed both gen and increase that opportunities to Our as factors can of investments, AI core we the deploy give good then AI seeing we on to and good that able we're it

about QX was mentioned outlook. revenue second question I earlier. this Your Again, the

ad generally smaller our advertising demand platform. QX. XX%, improvements, feel regions quarters We we verticals, like healthy performance on will which QX continue to effort to have result and many delivering of revenue ongoing and there particular We're into global accrue that that have platform. we a we to our grew accrued among in seen expect again, of advertisers, continue across many, And is basically where that strength value was strength saw including broad-based and

of comes Justin with line question from America. Post Bank the next Your of

users adult the doing us back with on and of Instagram? to made growth, U.S. and know can those any Instagram. on Reels especially And are maybe you Facebook young get give usage a Facebook I growth? a change But you user to want ago. big comment of couple just the I on quantification years what

Justin. Thanks,

adults. Posting Groups we that active a Marketplace, And they for the Canada and young products young product to in seen the we've in see have by young quarters. app And years, We've this core improvements Facebook has So and in and groups to young growth adult with been been U.S. adults. engagement are boosted building very U.S. with and also from growth the products mind usage see adults U.S. demand several past has we've been users translate traction into team particular strong efforts Marketplace recent priority seen Canada like benefited That's growing. the has which been products area for in for cohort. encouraged these in that secondhand the seen of mainly healthy with in and

with Your line the Mark of next comes Mahaney from Evercore ISI. question

It's a great, bigger and indirectly be a it's eBay. nice somewhat under-monetized very I was marketplace. monetize it arguably going so or about segue. I ask asset. than you even know may to that's Marketplace undermonetized It very a large

you secondly, on want the in year-over-year. thoughts what do Marketplace. And Your even X% down then I may improve the It's maybe part, ask of in terms to of it, just the about want about to monetizing in future quality you headcount.

talk the would forward? employee headcount be You're much significant thoughts a should Should was at expect pretty with significant? back to reduction. Did helpful. moderate a headcount going growth Any about growth where we in growth about you par CapEx? we prior significant How think on that in headcount

for for and ads that creating commerce broader like been happy find prong drive products the probably Shops young platform. I to X we ads. we've items Thanks, Mark. strategy relevant products, their more that a with of the Shops investments say strong have, it Shops of buyers strength is shopping. first of advertise The one in to is best is a our our platform. And that other On purchase seeing businesses also pace. would consumer-oriented. it's combination that Marketplace focused also about basically on in Advantage+ your be growing and Marketplace that year-over-year experience the our question at ads strategy quite commerce advertisers, making I the well with feel on on continues shopping Marketplace, revenue working about for making some drivers performance of obviously easier to broader is generally, it's adults. to We're excited in obviously we incremental part been in end, again, say we're just which would

Your was headcount. question second about

that working that new down it's core prior to allocating underrun. we're about as continue disciplined with hiring ensure over we're where that we We do headcount company higher XXXX. the further a And be of year, on than where we is ended headcount but expect to XXXX also priorities, our in-seat focused this reported I meaningfully end hiring that close underrun will we course really

headcount expectations again, growth will XXXX areas, yet. started sort aren't focus We haven't that process target we But to our expect we a process. be priority providing disciplined as hiring I budgeting and we on our headcount running primarily very will of yet

Squali comes line question next Your with of Truist the Securities. Youssef from

far. system been for the So variations so a for commercial using really search impressive except and everything AI getting X.X becoming closer looks engine be has and incorporated queries virtually in Llama to to different seems full

is to to Kind what So to are et any up with you, open deepen Ray-Ban, in like a And about cetera. relationship? it marketplaces it little of search, more web? on can third-party What link Mark, to you look talk bit the of kind would relationship for the that testing, to What OpenAI maybe the most exciting commercial like? maybe there opportunity be off are to EssilorLuxottica? that plans areas may then broader your

Yes. I'm landed. very excited with Llama X.X how

fulfill that lot X out Meta want you it by have areas when the is to which to work a some around and X.X, do is commerce I basically potential ways, there as team X.X make the the doing kind we this specific Meta we version year think more what that great the AI are initial frontier just cohesive as -- going was useful just a things to to uses Llama the to like going AI being vertical I of really and trends competitive And a we're functionality services make and leading talking there's the built about all other was people. generation behind can a need I last and, from go around Llama, product. in the build before other but assistant we've add, bunch the version models. as different just to top in vertical AI for first a of eventually, of well think, now that these to of closed build ideal Llama usage Llama

is don't the years to next or to happen a just in that widely this very as we road like next that get be long of anything used map. part of over something few a think, service. build I what's I going finished it's think stuff will, But that. So couple this is going quarters the

So I'm to excited about quite we're that. And on working going continue Llama, too.

mean, little and powers the we're of it's they're open like related. on was just Llama the about more I'm think the and they're -- the both excited but I is you mean, question about I I sort making mentioned of So a engine progress product that Meta AI and Llama, those. that both source,

On the great smart is a partner. EssilorLuxottica glasses,

been the of the think in doing the to them a it's now we're that functionality all surprise. positioned Meta think we generation And revolution is the well, are than second generation kinds well first We Ray-Ban meaningfully grow had part offering glasses. just They're well new and that from AI better so we of expected, seeing with positive there. that's expected I and that I dovetail to very of

glasses generations new different is able to building great of to that we company that great. But with to products product continue the deepen lot that the it a be So a just better. better was EssilorLuxottica make and hope and has and of AI partner

it really are leading go to I here. this to needs there's doing what well like a all quite think thought a of to what think, signs to lot at compared early I from but be the it's And consumer compared good. more we're I we electronics, to, piece think be, still point,

Citi. with of Your Josey line next Ron question comes from the

X. product of or I what or Llama call offerings offering I products bit is center you so meaning and out Llama the get these source from on that service to commentary Totally some cloud, X. for understand a to open the customer to vision more Meta companies, on wanted a call potentially is other Mark, of and public other little back, want products offering it, come hear not verticals. that maybe to of And

insights X.X for offer services any how just enterprise so And source of on others and open Llama benefit from? can you maybe to the sort envision greater

is the that Llama foundation people different So all shape of kind can into model products.

it's of a build to size that the that in It's that haven't for inference be Meta that's like agents infrastructure, built we not products other to or different we're that the whole of these basically build yet, stuff. the about, that to leading the with to that the and for It developers that can of want But it things or or Ray-Ban an ways experiences are or able is an lot we silicon more want whether going Studio building. make efficient developer business model, of are we you us with infrastructure the different well. Like is any AI kinds sure So out been all the power even just sourcing have there and And the the different custom make open valuable better ecosystem. other things or tools in ourselves business or fine-tune software take doing, the a different Llama. similarly, diversity methods why a really like AI think all have and adding whether -- silicon companies that optimizations that's of reason is models I people the into that for a isolation. new the is talked capabilities distill greater glasses, assistant to the are so the consumer it's this ecosystem mix, I to piece to we things thought all as

So valuable products all as us end kind built it kept it's we build then ourselves. as we're do going within being. be we just if actually for of can't And Llama that to and to all up walls, the it wouldn't our building that

strategy course, around really to want closely order it business with other And business positive necessarily a all it that. be big like we we're this. anywhere. NVIDIA, that's in especially developers what like great why Microsoft players build for working with like cloud companies strategy I did Cloud, feel a that's is directly this of and And work Databricks, sell think, So for to of don't to the able supporting AWS, and part it get Google doing release. Other we we us. Azure, they're be And for need to

want that of you to advantages -- with we model source Google one Gemini that's everywhere or it's and an of with locked or take cloud not this I open it whether think it's that. OpenAI is it use the offers can that you're into model, whatever and is, Microsoft like like Llama encourage X

that quite like agent I addition I think to on, applications I'm it is the earlier. just about and enterprise that. The there's comments going experience talked though, be excited potential earlier, just that a to optimizing piece. focused advertiser huge most about business said the business in So my like we're in I

I They business e-mail think an a pretty social accounts. media much has every today, it have They address. website. have

just millions that's in in that think, quickly the that their going of that use the all countries. AI, this history and thing awesome. And order suited for questions when just pull labor. the think someone have product, something. And the we very if of everyone's that already the of they're agents countries they've stand catalogs if help going get shared done a think see well the working I answering towards doing. of and an up I to things agent, they to is is information need we touch people with X, I sales thing not be to have people do to expensive a think But have in future, in is for make interacting you businesses messaging things that really I business they lot that their to of to at that if over in low the that with with least hundreds issues is And range like platforms well support helping whole all going in to be from the businesses multiple to can content quite they've buy our and cost that and can quite interacting easy have

Llama of custom products all our business. both good. combine And mean, really lot huge, investment make across that business going consumer we're X rest this goes different couple the area. teams I that's we're think doing So can I that to is Llama but of the a that's building, of and be that work ecosystem other to we in and the that the doing with with all just

stuff. So exciting it's of lot a

we time have for Krista, question. X last

will question Ross of the Sandler That come line Barclays. from with

up Mark, you'd do with something along if ended of X the scaling X so work interview stopping have on Monday you. ahead to your said product and day, in Jensen, you lines of years of

agents are looking assistants, at areas question of thinking other AI is capacity what maybe in X-year more road outside AI or stab So about second you that or in the at the map? And question. then X

said Llama HXXXs. trained X.X guys XX,XXX was You on that

going You've that available XXX,XXX year-end. to said you're also have by

you plenty guys So things? entire this That's modeling for of between we out and and up go all. X, of I map products. CapEx are GPUs even of to inference training, kind guess extra XXX,XXX future Llama kind we capacity future have for road inference, how if

I second future a I'll Susan think with It's the true hypothetical It's models. start first no the there I foundation let an then interesting that and because it's do I little can if, question. one. answer were mean, one, let's say,

capabilities to huge just But bear, going that would takes models, our around of are time, would foundation be we're the and time. bring that be I amount the to products then same and be to future those. think unlock they're innovation product planning a just and going industry at new there awesome and there

So isn't our road not innovation. planning assuming future that there really map I'm product

the you're having that AI Llama breadth it think to of the and about cases like better individuals on businesses with wanting planning content that of -- kind that of Susan with set you're for just X something set contrary, of for making most the to next. that that capabilities stuff kind And stand glasses to and is things we creators that to important people, doing. And just going laid utility multimodal time and these the interact are increasingly too. discover be kind we any need you're and going use in map that On going useful be up do what all about can real-time, talk they and beyond have you I I and AI, Meta in expect want, will this, agents for businesses have huge all Llama out what's already based their are getting X be going we the what to to I'd friends people of that and this all guess with reaching dovetails this road

exciting I Susan space. I some my a think and it's guess I'll the that to the GPUs time becoming kind this the was around lot all a think a a and was that. And where way I there's there's kind numbers fully innovation there's that is at So -- of to long just take of the just in of becoming products the for just technology stab explored be saying area lag available let my that, that point very going come. of just there between

Thank you.

a growth in how in We about But it across terms inference, also of driving the process whatever alluded for longer-term be the really and are the comments we to we use to data think And what training about of how of several that life in we it capacity, the Mark next of years constructing, sort amount we center. have data core will that. we then are and We around business, look subsequent needed in lifetime we terms other aren't cases might capacity needs building the outlook. When and that's what generations AI-based of new and to we're of -- prior would future the we've in out we CapEx across architected we of Llama. think lot use generations products. the over about Llama of any its that train are might at what use in given for of And clearly significant into think has the sort towards share in use his generative about generation our about we XXXX. a talked of we position now think project center

gen at to its of cases to individual of for from host there's ranging those core AI outset supporting use data cases. content implications, and inference whole training AI gen use life thinking the servers sort and for also kinds center use different of what So might used ads through any the a too, of support types being of and to of recommendation ranking potentially we

the in towards. put on the them to order sort terms flexibility best of to thinking in in and that mind, a servers each that times we that there center the in decisions are us time make you'll what the And data through centers, we're of both any future. data the range order also when the are servers we up cases what recognizing And really of again, points mapping when wide potential gives we of of horizon information a use multiple really to but long undertake doing project. with So given because decision based lifetime spinning in and across lead long have

joining and all Great. you we forward for Thank time, you look We us to speaking your soon. with again today. appreciate

concludes you disconnect. you conference today's participation, This may your Thank call. and for now