TRANSCRIPT

Q1 2025 — Follow-up call transcript

Q1 2025PDF · 11 pages85 KB
1
Meta Platforms, Inc. (META)
First Quarter 2025 Results Follow Up Call
April 30th, 2025
Operator: Good afternoon, my name is Krista and I will be your conference operator
today. At this time, I would like to welcome everyone to Meta’s First Quarter
Results Follow-up Q&A call. All lines have been placed on mute to prevent any
background noise.
After the speaker’s remarks, there will be a question and answer session. If you
would like to ask a question, please press star one on your telephone keypad.
To withdraw your question, again press star one. And we ask that you limit
yourself to one question and this call will be recorded. Thank you very much.
Kenneth Dorell, Meta’s Director of Investor Relations, you may begin.
Kenneth Dorell: Thank you. Good afternoon and welcome to the Follow-up Q&A call. With me
on today’s call are Susan Li, CFO, and Chad Heaton, VP of Finance.
Our remarks today will include forward looking statements, which are based on
assumptions as of today. Actual results may differ materially as a result of
various factors, including those set forth in today’s earnings press release, and
in our annual report on Form 10-K filed with the SEC. We undertake no
obligation to update any forward looking statement.
During this call, we will present both GAAP and certain non-GAAP financial
measures. A reconciliation of GAAP to non-GAAP measures is included in
today’s earnings press release. The earnings press release and an
accompanying investor presentation are available on our website at
investor.atmeta.com. And now, I’d like to turn the call back over to Krista for
the first question.
Operator: Your first question comes from the line of Stephen Ju with UBS. Please go
ahead.
Stephen Ju: Okay, great. Thank you so much. So, Susan, I guess sort of a clarification
question. I think I heard you on the main call say that the majority of the CapEx
you will be incurring this year will be going to the core.
And I think we’ve been talking about how some of those, you know, capacity
constraints that have been there throughout the course of last year should be
coming off this year. So are there any sort of emerging constraints that we
should be thinking about versus what might have been the case three months
ago, the increase in the annual CapEx guidance notwithstanding?
And secondarily, I think there was a lot of discussion about automation that
you’re bringing to your merchant and hopefully eventual advertiser base. So I
think it’s been some time since we’ve had a discussion on how many
advertisers there are on the platform relative to how many merchants are on
the platform.

-- 1 of 11 --

2
So I was wondering if you would care to give us an update on that. The
optimist in me wants to believe that there’s a lot more who have converted into
being advertisers. Thank you.
Susan Li: Thanks, Stephen. On the first question. So, yes, just emphasizing that the
majority of CapEx this year will be going towards the core. I also alluded to on
the call that the capacity planning process for us is pretty dynamic and that we
are both in an ongoing process of updating both what we’re able to get in terms
of capacity online, but also on updating what our teams can use in terms of
capacity, or put towards good use.
And so, those two supply and demand components both fluctuate a little bit
and that has extended a little bit the duration of capacity crunch for us. But
again, it’s something we’re working very actively to address. On your second
question, Chad, do you want to take that?
Chad Heaton: Sure. So, in terms of automation for merchants, I think what you may be
referring to is Advantage+ where we continue to invest and see really good
results from our advertisers. We’re continuing to see growth with a suite of AI
powered solutions. We’re helping more advertisers easily take advantage of
these tools.
We’ve been testing a new streamlined campaign creation flow that has
Advantage+ turned on from the beginning for the sales, app, and lead
campaigns. And that automates and optimizes several aspects of the
campaign setup process at once, including selecting the audience, the ad
creative, where to show the ad and how the campaign’s budget gets
distributed across ad sets to drive the most efficient outcomes. So, this is an
area that we continue to invest in. We’re seeing really good results from it.
I think you also mentioned something about the number of advertisers. We
haven’t updated that stat recently, but we continue to have a very large and
well-diversified advertiser base.
Operator: Your next question comes from the line of Ron Josey with Citi. Please go
ahead.
Ronald Josey: Great. Thanks for taking the question. And Susan, I want to follow up on the
five key areas and AI areas that Mark went through on the call. And a lot of
these areas have very tangible products today at scale.
And so, I wanted to understand if we have the capacity from an infrastructure
perspective to support call it the short, medium, long-term growth rates for
whatever is to come. In other words, it seems as if a lot of these top five
already have very good head starts to drive monetization the next one to three
years, but then I think your vision is probably the next 5 to 10. So, we’d love to
better understand how you manage those two.
And then maybe more specifically, when we think about impressions and
pricing growth, I think we saw once again, pricing growth grew faster than

-- 2 of 11 --

3
impressions 10% versus [inaudible] pricing grew 14%. Is this still the continued
mixture to objectives or do you think impressions might make a comeback
here? Thank you, Susan.
Susan Li: Thank you, Ron, for the questions. On your first question about CapEx and
whether we have the infrastructure to support the long-term roadmap, that is a
great question. It’s one that we think about all the time in our long range
planning process.
And obviously these are long lead time infrastructure projects. It’s part of why
the sort of infrastructure planning process is so dynamic. We are constantly
evaluating what we think is the runway that we need to support the broad
portfolio of core AI and gen AI efforts that we’re building for today and where
we think they might go over the next years.
So, again, we are leaning into what we believe is one of our competitive moats,
our ability to build world-class infrastructure. We think that’s going to enable
us to develop these technologies and services and host these services at scale
over the coming years and continue improving the way our core business
performs also by putting more compute against ranking and recommendations
work.
So, it’s a big part of why we’re investing aggressively today with flexibility in
mind. And we are thinking about this very actively when we think about future
training and inference needs.
Chad Heaton: In terms of impressions, so Q1 impression growth of 5% was primarily driven by
the Asia Pacific region, which benefited from continued Instagram and
Facebook user and engagement growth, particularly from video. I know your
question was more about the outlook going forward. We’re not sharing an
impression growth outlook for Q2 or 2025, but we do continue to see
opportunities to drive impression growth.
Operator: Your next question comes from the line of Benjamin Black with Deutsche Bank.
Please go ahead.
Benjamin Black: Great, thanks for taking the question. I’m just curious what you’re hearing from
some of your advertisers on the ground who may be subject to significantly
higher input costs, just given the rising trade related tensions? There just
seems to be a lot of trade related angst out there.
Do you get a sense that some of the advertisers are potentially pulling spend
forward at all and your benefit of that? It would appear from the top end of
your guide that, it’s typical normal seasonality. So just curious to what you’re
hearing from your advertisers there, really.
Susan Li: Yes. We shared what we have seen, which is we saw some of our Asia based
advertisers cut spend to the United States in April. Again, some of that spend
has been redirected into other geographies, but not fully. And we have not

-- 3 of 11 --

4
we’ve not heard advertisers tell us explicitly that they are pulling forward
spend.
We certainly think it is possible that some advertisers were spending more
opportunistically in Q1 ahead of events like the de minimis exception expiring in
May. But we’re really sharing what we have observed to date and our revenue
guidance range really reflects what we think is a pretty wide range of
outcomes, given the sort of broader macro environment.
Operator: Your next question comes from the line of Brent Thill with Jefferies. Please go
ahead.
Brent Thill: Hey, Susan, we’ve had a lot of questions around monetization of Llama, and I
know you’re not giving the exact timetable out. But when you think about the
timetable relative to Messenger and some of the other up and coming
solutions, do you feel like you can monetize at a quicker rate or just any
framework you can give us that’d be helpful. Thanks.
Susan Li: I’m going to take a crack at your question and you can let me know if I’ve
approached this in the way that you were intending or not. Direct monetization
of the Llama model is not really what we are focused on, per se.
Our focus is on over time, leveraging the technology that we’re building to build
into these five areas that Mark touched on on the call, whether it’s making the
current ads business better by making it much more personalized at scale.
Whether it’s growing core engagement by both improving the
recommendations of existing content types as well as fueling entirely new sort
of AI-powered content, whether it is accelerating the business messaging
activity that we see today into something that scales across more geographic
markets and becomes more of all-encompassing AI agent than what you see
today.
Whether it’s Meta AI and monetizing the way Meta AI eventually gets used, and
then also the AI devices as a form factor upon which to leverage those AI
capabilities. Depending on which of those five areas you’re talking about, the
timeline is somewhat different. So for example, Meta AI, we are still quite early
in building out the consumer experience. I expect that to be further out.
Conversely, in terms of accruing benefit to our existing ads ecosystem, for
example, that’s a place where we have a lot of active work happening already.
And we’ve shared some of the feature updates, how image expansion and
video expansion are being used by advertisers, and the degree to which
advertisers are quite rapidly adopting the gen AI tools that we’ve put forward in
the Advantage+ suite of tools.
So I would say, in the spectrum of those areas, the core ads businesses is where
we will see those benefits accrue first. And something like Meta AI and also AI
devices will play out on a longer timeframe.

-- 4 of 11 --

5
And I think content, improvements to ranking will happen before you see more
wide scale AI generated content. But I think each of those things is on a little bit
of a different timetable.
Operator: Your next question comes from the line of Rob Sanderson with Loop Capital
Markets. Please go ahead.
Rob Sanderson: Thank you for taking my question. I have two, please, if I could. Susan, you said
your guidance contemplates a wide range of scenarios, obviously China cross
border and de minimis exemption, all that, is obviously one of them.
Is there any way you can help us think about like qualitatively, what kind of
behavior you expect from that cohort, like in the upper end or the lower end of
your range? Are there any other specific areas of uncertainty that you’ve
contemplated in this range?
Is it just general economic activity over a healthy consumer and volatility in
supply chains? What is it that you’re thinking about in this wide range, if there’s
anything that you could provide specifically? And then I do have a follow-up.
Susan Li: There isn’t a lot more specifically that we would point to. We have reflected the
change in spend from those Asia-based e-commerce advertisers already into
the revenue guidance. Otherwise, there’s just really a lot of puts and takes in
the economic environment. So it’s pretty difficult to try to parse out very
specific assumptions and how they translate.
Rob Sanderson: Yeah, understood. Thank you. Now on product side, like the incremental
attribution feature you mentioned and talking about 46% lift in incremental
conversions, that’s very significant.
Can you talk a little bit more about this? How applicable is that feature across
your customer base? It seems like everybody would be interested in that kind
of incrementality.
Just to help me calibrate, maybe, pump the brakes a little bit on my
understanding of that, how meaningful that could be in terms of spending
growth and it sounds like a very big product breakthrough, at least on the
surface.
Susan Li: Yeah. So our incremental attributions tool is part of a portfolio of work we’re
doing across what we think of as value optimization updates for advertisers.
So ways to ensure the results that we drive are more customized to each
business’s individual objectives and the way that they measure value.
So incremental attributions is one of those. And we are seeing strong results
from the tests of that feature across regions and verticals. It is different from
the way that we have historically optimized primarily for absolute conversions.
And it is sort of, it’s modeling what are, we believe, conversions that would not
have occurred without the ad being shown. This is an opt-in setting, and it’s

-- 5 of 11 --

6
the only product on the market that optimizes for and reports on incremental
conversions in real time. So, we’re pretty pleased with the results there. As you
mentioned, the 46% lift, and we think that that’s going to be a very positive
feature for advertisers.
But it’s one of many things we’re doing in this space. We’ve also expanded
tests of our Conversion Value Rules feature to more advertisers. That’s a
feature that enables businesses to specify which audiences or conversions they
value more or less so we can better optimize their campaigns. And we’re seeing
good results in those tests too. We’re also working on a rollout to more
advertisers later in the year.
We are deepening our integration with third-party analytics integration. So,
we’ve been working on a rollout of a Google Analytics integration to
advertisers, as well as with other partners. So, I would say we’ve got a lot of
work that we’re doing here that we believe will help advertisers, again, measure
against the way that they specifically think about how they’re accruing value
from our ads.
Operator: Your next question comes from the line of Shweta Khajuria with Wolfe
Research. Please go ahead.
Shweta Khajuria: Thanks for taking my question. I have one on custom silicon. So, you started
deploying your custom MTIA silicon chips last year for ranking and
recommendation workloads. I guess my two questions are what percent of
inference workloads would they represent by year-end in your view? And then
over the next couple of years, as you scale your efforts on your custom chips to
support core AI workloads and gen AI use cases, is it fair to assume majority of
all workloads would leverage your custom chips? Thank you.
Chad Heaton: Yes, I can take that. So, broadly, I want to start saying we expect to continue to
purchase silicon from the industry’s leading providers and we remain
committed to our longstanding partnerships. However, as you note, we are
designing our own silicon for unique workloads where the off-the-shelf silicon
isn’t necessarily optimal.
Our in-house MTIA program is initially focused on supporting our core ranking
and recommendations inference workloads. We started adopting MTIA in the
first half of 2024 for core ranking and recommendations. We’ll continue
ramping adoption for those workloads on MTIA over the course of 2025 as we
use it for incremental capacity and replace some GPU-based servers when they
reach the end of their useful lives. As we look to next year, we expect to expand
MTIA to also support core AI training workloads and, over time, some gen AI
use cases.
Operator: Your next question comes from the line of Justin Patterson with KeyBanc
Capital Markets. Please go ahead.
Justin Patterson: Great, thank you. It sounds like the combination of building big models and
distilling some smaller, more specialized ones is creating some latency and

-- 6 of 11 --

7
efficiency advantages. Could you talk about how that’s changing the pace of
productivity at Meta just around how the timeline to bring new products to
market, and then when we might actually see that show up at the margin
benefit? Thank you.
Susan Li: Thanks, Justin. I wasn’t sure if you were asking about this with respect to the
gen AI releases we’ve been talking about or whether you were talking about
this more with respect to the ranking and recommendations models because I
realized we have talked about both at length on this call.
So, I’m going to answer very quickly more on the ranking and
recommendations side where we talked a little bit about the deployment of the
GEM, the generative ads recommendation model, on the ads ranking side that
happens basically after the ads retrieval stage, where last quarter we talked
about the launch of the Andromeda model, which is in the retrieval stage.
Basically, the big tactical breakthrough of GEM is that the newer architecture
here is twice as efficient as improving ad performance for a given amount of
data and compute versus prior recommendation models, which lets us basically
use their training clusters at more attractive ROI than we have in the past.
We also think that GEM’s efficiency at handling large volumes of data will also
be beneficial in handling the expected growth in ad creative in our system over
the coming years as gen AI tools are used to create a greater number of ads
and ad creatives that exist today in the more personalized ads world.
The GEM architecture also leverages event sequence learning, which we talked
about in the past. So, similar to how LLMs predict and reason about the next
word in a sentence, here the model can recognize that people who convert on
an ad for a hotel booking in Hawaii may often convert on ads for sunglasses or
swimsuits or snorkel gear, etc.
So, I think this, because of the better performance efficiency, will hopefully
help us meet the scaling needs in ads ranking improvements as the quantum of
data that goes into what is getting ranked increases over time. So, I don’t think
we have exact timelines on this, but I think this is intended. I think this
technological innovation sets us up well for what we think will be a more
personalized volume of ads in the future.
Operator: Your next question comes from the line of Michael Nathanson with
MoffettNathanson. Please go ahead.
Michael Nathanson: Thanks. I have two for you, Susan. One is Mark talks for a long time about
wanting to control your own destiny at Meta. I wonder how critical are glasses
to controlling your destiny and what would happen to things like Meta AI if
phones remained dominant and not glasses going forward? That’s one.
And then two is on your cost guide, I realize that we have definitely have higher
depreciation over the year. But is there anything else you want to pull out,

-- 7 of 11 --

8
maybe headcount growth that we should be aware of that scales over the year
that we have to think about? Thanks.
Susan Li: Yes, thank you, Michael. On your first question about the glasses use cases, we
are building for a world where mobile is dominant. But I would say that based
on the way we are both seeing the Ray-Ban Meta AI glasses sell, as well as the
way that users are using them, and I think we shared on the call that we have
over four times as many monthly active as a year ago, we’re seeing that there’s
a compelling demand for this use case, both in terms of consuming audio,
capturing content, calling on the Ray-Ban Meta AI glasses while you’re moving
about the world. Those are very popular features.
We’ve made upgrades to core features from the first generation of the product
like open air audio and hands-free photo and video capture. And we’re
continuing to see growing adoption and usage of Meta AI on RBM with many
people using it to answer questions, control their devices.
Generally, I would say more positive signals on our thesis that glasses are a
really ideal form factor for an AI assistant. The most recent feature rollouts like
live translation seems like a really intuitive use case, for example, for wearing
glasses. So, we certainly think that all the data we have gathered so far would
suggest that glasses are a very good form factor for this, but we’ll clearly
continue to learn more over time.
Chad Heaton: I can take the second part of your question about the expense guide. When we
look at year-over-year growth, so first of all, infrastructure costs. We expect
infrastructure will be the single largest driver of 2025 expense growth. That’s
driven both by higher operating costs as well as depreciation expense.
The higher operating costs are driven by increased spend on some cloud
services in 2025 to meet our capacity needs, as well as higher costs associated
with running our expanded data center fleet, which includes higher rent, energy
costs, and other operational expenses. The depreciation expense will also be a
contributor to infrastructure cost growth, though we expect growth to be
slower in 2025 than 2024, as we benefit from extending the useful lives of
servers.
After infrastructure, employee compensation is the next largest driver of
expense growth in 2025. The growth in employee compensation will be
primarily driven by headcount growth in our higher-cost technical talent to
support our primary areas of monetization, infrastructure, generative AI,
regulation, and compliance, and Reality Labs. We expect to constrain
headcount growth in other business functions.
Operator: Your next question comes from the line of Tom Champion with Piper Sandler.
Please go ahead.
Tom Champion: Hi. Good afternoon. Susan, I was wondering if you could talk a little bit about
auction dynamics and what’s the right way to think about the impact when a
large spender drops out? I guess specifically with China and the Chinese

-- 8 of 11 --

9
reseller cohort, it sounds like spend is being geographically redistributed. But
what’s the best way to think about how the auction responds in the U.S. when a
large customer or customer cohort drops out?
Then secondly, I’m curious if there were any trends in terms of revenue growth,
linearity in the quarter, whether it was relatively even through Q1 in April, or
whether there were any fluctuations in growth. Any comment there would be
helpful. Thank you.
Susan Li: Yes. We don’t really comment on the intra-quarter dynamics. But in terms of
impact on the auction, we, of course, lose some revenue if large advertisers
reduce spend. And that, of course, puts downward pressure on price, all things
equal. But we do have a broad and diverse business. So, if some advertisers
reduce their spend and prices fall, it creates an opportunity for other
advertisers to step in.
Now, that process will take time. So, typically, if you see large advertisers
reduce spend, you see more of an impact when it happens immediately. And
then advertisers observe and react to more favorable pricing, so then there’s a
longer-term fill-in effect in the auction.
The other thing I would just say is, I think we clearly have broad exposure to the
macroeconomic environment, we have this auction, which means advertisers
have the ability to adjust and pull their spending in real time. And online
commerce is our largest advertiser vertical, so our exposure to this vertical
makes it a bigger headwind if it’s impacted.
But I would also say, and I mentioned this on the call, advertising is a relative
performance game. The vast majority of our business is direct response
advertising. So, while you’d expect to see an impact there if consumer
spending reduces, advertisers are also more acutely focused on measurable
ROI and driving sales in that environment. And we think that benefits us on a
relative basis.
We also have a highly diversified basis by geography, so advertisers can use our
services to reach people around the world and target markets where there may
be incremental opportunity.
And the final thing I’d add is our ad inventory is generally vertical agnostic. So,
we saw a benefit from that during the early part of COVID when reduced spend
in certain verticals like travel was partly offset by increased demand in other
verticals that were performing well, like gaming. So, that also gives us a
measure of flexibility if some verticals are more uniquely impacted than others.
So, overall, I think we are in a strong position to navigate a wide variety of
economic outlooks. And again, what we are seeing today is that demand trends
are generally healthy, having called out the impact of Asia-based advertisers
spending less in the U.S.

-- 9 of 11 --

10
Operator: Your next question comes from the line of Jason Helfstein with Oppenheimer.
Please go ahead.
Jason Helfstein: Thanks for taking the question. Susan, can you perhaps, I don’t know if there’s
any additional color you can give on just the way you’re thinking longer-term
about a potential DMA resolution, what that would look like, or potential
timing. And then, secondly, we can guess where you end up with exiting
depreciation this year. Is it fair to say that depreciation most likely grows faster
than revenue for the foreseeable future, just given the capex spending? Any
color there would be helpful. Thank you.
Susan Li: Let me take your first question on the DMA issue. Again, since we are in the
active process of engaging with the European Commission, it’s really too early
to share more outside of what we have already shared, which is effectively that
the European Commission -- their decision relates to the consent model that
we introduced in Europe for personalized advertising in 2023 when we first
added the option for people to subscribe to Facebook and Instagram and see
no ads.
In November ‘24, based on regulator feedback, we rolled out these additional
updates, such as reducing the price of the no ads subscription and introducing
a less personalized ads option for people to continue using our services for free,
albeit with ads that are almost non-personalized. And we believe that the
current solution goes above and beyond what is required by law.
But the European Commission announced that our original consent model of
offering subscriptions for no ads was not compliant with the DMA. And while
they have not deemed our less personalized ads offering to be non-compliant
with the DMA, we expect that we’re going to be in this ongoing regulatory
dialogue and that we will need to make some modifications. So, we’re engaging
actively. We don’t have a lot more to share there right now.
Chad Heaton: Real quickly on the depreciation question. So, just first off, the depreciation
expenses in Q1 were $3.8 billion, which is up 15% year over year. That’s versus
41% in Q4. The key bit there is that’s a result of extending the useful lives of our
servers and network assets at the beginning of the year. Based on the servers
and network assets placed in service at the end of 2024, the financial impact of
that change in estimate included a reduction in depreciation expense of $826
million in Q1.
As I mentioned before, the depreciation expenses will be a contributor to
infrastructure cost growth this year. We’re not comparing it to revenue, but we
did mention that we expect growth to be slower in 2025 than in 2024, again,
based on that benefit that we’re getting from the extension of the server useful
lives.
Kenneth Dorell: Krista, we have time for one last question.
Operator: Your last question comes from the line of Dan Salmon with New Street
Research. Please go ahead.

-- 10 of 11 --

11
Dan Salmon: Good evening, everyone. Two questions, maybe, Susan. First, are you seeing
any benefit from the, I guess, state of purgatory that we find TikTok USA in
currently? Interested to hear if that is helping your ad business right now.
And then second, Mark’s line during the public call about how he anticipates
some of those ad creative tools I think you were referencing earlier, increasing
productivity and making advertising a much bigger portion of GDP today. I
think everybody on here has a good idea about GDP numbers, but I’d love to
hear more on what the house view is on how big the advertising market is today
and where you think it could go in terms of a percentage of GDP eventually.
Thanks.
Susan Li: Thanks, Dan. On your first question, we did see increases in engagement,
particularly in video time on Instagram and Facebook during the brief period of
time that TikTok was not available in the United States. But outside of that,
there’s not a lot more to say there. And we’re basically just focused on making
sure that we are building the best product we can and focusing our efforts on
the product work that we’ve talked about, the OG Facebook work, the work
we’re doing to attract creators, et cetera.
And then your second question about advertising as a larger share of GDP
going forward. Next time Mark takes a chief economist role, we will have him
quantify his estimates. But we’ve seen a general transition from offline
commerce to digital commerce. As part of that, there’s more opportunity, in
particular, in online advertising.
And I think there are really two dimensions. One is we continue to make
advertising, performance-driven advertising, in particular, more performant
relative to other forms of advertising. That has continued to sort of move share
towards us, but also grow the advertising pie broadly.
And then I think the second thing is then if advertisements become a lot more
personalized at scale because of some of the tools that we are building, that
also grows the ability for these ads to become a lot more performant.
And we also have a big platform for the distribution of those ads. So, I don’t
think there is a specific quantification here, but I think there are a lot of factors
that we believe have positive runway for us and for this general concept of
continuing to grow the advertising pie over time.
Kenneth Dorell: Great. Thank you, everyone, for joining us today. We look forward to speaking
with you again soon.
Operator: Ladies and gentlemen, this concludes today’s conference call. Thank you for
your participation, and you may now disconnect.

-- 11 of 11 --