TRANSCRIPT

Q1 2024 — Follow-up call transcript

Q1 2024PDF · 13 pages171 KB
1
Meta Platforms, Inc. (META)
First Quarter 2024 Results Follow Up Call
April 24th, 2024
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 the Meta First
Quarter Results Follow-up Q&A Call.
All lines have been placed on mute to prevent any background noise. After
the speakers’ 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. We ask that you limit
yourself to one question. This call will be recorded. Thank you very much.
Ken Dorell, Meta’s Director of Investor Relations, you may begin.
Kenneth Dorell: Thank you, Krista. 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.
Before we get started, I would like to take this opportunity to remind you
that our remarks today will include forward-looking statements. Actual
results may differ materially from those contemplated by these forward-
looking statements. Factors that could cause these results to differ
materially are set forth in today’s earnings press release and in our annual
report on Form 10-K filed with the SEC.
Any forward-looking statements that we make on this call are based on
assumptions as of today. And we undertake no obligation to update these
statements as a result of new information or future events. 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 press
release. The earnings press release and an accompanying investor
presentation are available on our website at investor.fb.com. And now, I’d
like to turn the call back over to Krista for the first question.
Operator: Thank you. We will now open the lines for a question-and-answer session.
To ask a question, please press star, one on your touchtone phone. To
withdraw your question, again, press star, one. We ask that you limit
yourself to one question. Please pick up your handset before asking your

-- 1 of 13 --

2
question to ensure clarity. If you are streaming today’s call, please mute
your computer speakers. And your first question comes from the line of
Mark Mahaney from Evercore ISI. Please go ahead.
Jian Li: Great, thanks. This is Jian Li for Mark Mahaney. Thanks for taking a
question. So Susan, as maybe when we think about the expense grows as
you lean into these AI initiatives over the next couple of years, could this lead
to any lumpiness or temporary deleverage of margins?
How should we think about the consistency in expense grows in relation to
revenue grows on a company level? And just to clarify, on the total expense
increase, beyond the Q1 legal expense, is there incremental legal fee baked
into the guide for the rest of the year? Like is this more of a one-time
accrual, should we think about it more as a recurring expense? Thanks a lot.
Susan Li: Hi, thank you. So I’ll answer your second question first and then come back
to the first one. So the 2024 total expense guide, we brought the bottom up
for two reasons primarily. One is the reflection of the higher legal accruals in
Q1, and obviously this is an area that is just very difficult to sort of forecast
with precision over the course of the year.
The second area is deprecation, just as we are pulling in CapEx, you see the
corresponding depreciation flow through the P&L. So those are the two
drivers of bringing up the bottom end of the total expense guide.
On your second -- sorry, on your first question about AI investments in the
next few years and lumpiness in the margins, it’s really hard to say. Part of
this is because the question has two parts, one is what is the revenue outlook
for the next few years, which we don’t share a long-term revenue outlook
and there are lots of factors that can drive that.
What I will say is sort of consistent with what Mark mentioned today in his
remarks, we have tremendous ambitions in the AI space, and we believe that
there’s going to be a lot of opportunity for us to build out technologies here
that both will accrue value to our core business, as well as open up new
business opportunities for us to that.

-- 2 of 13 --

3
Doing that means we are going to want to have the infrastructure capacity to
be training the cutting edge model, so the successors of Llama 3, and it’s
really a long-term road map as you are thinking about that. So we are really
planning ahead as we think about this investment cycle and what we think
we need to do to enable the foundation models in each successive
generation to help drive the levels of innovation that we’re looking for.
Susan Li: Throughout that, I will say, just on the overarching sort of financial outlook,
the year of efficiency is not lost on us. That is something that as we kind of
put together the overall financial road map we’re very focused on sort of
discipline throughout the company, even as we are investing to support the
AI and Reality Labs initiatives.
Operator: Your next question comes from the line of Colin Sebastian from Baird. Please
go ahead.
Colin Sebastian: Oh, thanks guys, good afternoon. Maybe I’ll try it this way, Susan. I think it’d
be helpful maybe to get a little more color on the linearity of the CapEx ramp
through the year. I mean do we get a step function higher in Q2 or is it more
of a ramp towards the end of the year that then forms the run rate bases as
well into 2025?
And I guess, if I can sneak in one, I mean the commentary on WhatsApp
momentum I think was pretty clear. I’m just curious, why not give us an
update on the revenue run rate there, especially since messaging with the
revenue opportunity in AI seems like a pretty big focus. Thank you.
Susan Li: Yes. So there are a number of factors that can influence CapEx landing in a
given quarter -- when you can take delivery of things. So it’s a little bit hard
to perfectly nail that down in advance and that can result in some lumpiness.
But we generally expect that CapEx is going to grow throughout the rest of
2024 due to higher server and data center spend. On the server side we are
seeing some relaxing in the supply of GPUs. So there’s still a range of
outcomes embedded in the outlook and obviously supply availability is one
of the factors that will influence where we land.

-- 3 of 13 --

4
And on the data center side you’ll start to see sort of meaningful build
activity from the new data center architecture that we had announced
earlier. But again, when we think about the long run CapEx picture, there are
just a lot of factors at play. The pace of advancement in the underlying
models, the adoption of our gen AI products, how efficient can we make our
models, what performance improvements will come from future generations
of hardware.
So there’s a lot of uncertainty, I think, on all of those dimensions going into
the future and we aim to be as flexible as we can in planning around all of
those variables and the degree to which they might change over time.
Chad Heaton: Yes, so I guess on your WhatsApp question, there’s a couple different ways
that we monetize WhatsApp. One is through click to messaging ads and then
the second is on paid messaging. On click to message, that continues to grow
at a healthy pace. We’ve seen click to WhatsApp growth be very strong in
particular, and we’re seeing a lot of opportunity to scale that line.
As we look to paid messaging, that has now surpassed a $1 billion annual
revenue run rate and is growing at a really strong pace. We’re seeing that
growth driven both by broadening business adoption as well as increases in
the volume of conversations per business.
And in terms of the type of message, we’re seeing really strong growth from
marketing messages. So these are both lines, again, on the WhatsApp side
that we’re seeing a lot of growth from.
We think we’ve got a lot of headroom, a lot of opportunity, that we’re going
to continue to invest against.
Operator: Your next question comes from the line of Justin Patterson from KeyBanc
Capital Markets. Please go ahead. Justin, please go ahead.
Kenneth Dorell: Krista, why don’t we move to the next person and then we can come back to
Justin.

-- 4 of 13 --

5
Operator: Certainly. Your next question comes from the line of Rohit Kulkarni from
Roth MKM. Please go ahead.
Rohit Kulkarni: Hey, thank you. I guess just on the CapEx, following up on what Colin asked. I
guess to the extent you can give any part or rationale behind the uptick in
CapEx? Now it’s five quarters it’s been around $27 billion, $28 billion, $29
billion annual run rate and suddenly no matter which way we cut it it’s
jumping to close to $39 billion, $40 billion run rate.
Maybe if you could provide anymore color on what’s changed that this uptick
is necessary here and now. And then just a historical question in terms of
the curious trend in how pricing and ad load has trended over the last four
quarters. Any periods in the past that you think are most comparable when
you thought about pricing starting to pick up and then you had control over
ad load, other impressions? Does that pricing rise even faster? When do you
think in the past few periods are most comparable to what you are seeing
right now?
Susan Li: Thanks, Rohit. So on the first question about CapEx, what you're seeing in
the increase in our outlook and also in Mark’s remarks were really
accelerating our infrastructure capacity build so that we can be in a position
to have what we believe are the training clusters needed to support the next
generations of foundational models.
And I mentioned I think also in maybe my first answer, and we are thinking
ahead. So we’re not just thinking about what is the CapEx we need to train
or the infrastructure rather, sorry, we need to train Llama 4, but we’re
thinking about, when we want Llama 7 to come out, what capacity do we
need? And we are putting place the groundwork for infrastructure to enable
us to do that.
Now, the two things that I would add on top of that are, one, again, we’re
going to learn a lot more about the ongoing cost dynamics, the pace of
advancement, all of those things. So as we learn more over the next several
years, I think we’ll get a lot more information that could kind of influence the
shape of the curve.

-- 5 of 13 --

6
The second thing is that, again, we’re building this out in a way where we can
repurpose the capacity for other things. So if it turns out that we do not
need as much training capacity as we expect, we will have the flexibility to
deploy it towards inference capacity, for GenAI usage, or if it turns out that
we don't need as much of that as we expect, we will be able to put it towards
core AI ranking and recommendation use cases, which right now we are still
at a place where capacity that we can put towards those use cases are quite
positive ROI for us.
So I recognize that the CapEx outlook has certainly gone up, but we also think
that we have a lot of flexibility to maneuver in the years ahead.
Chad Heaton: Yes, and to your question about pricing and ad load. So first, the global
average price per ad increased 6% year-over-year, which was up from 2%
year-over-year back in Q4. All regions saw improved pricing growth due to
stronger ad demand and ad performance improvements. So feeling good
about the value that we’re delivering, and obviously advertisers are seeing
that through increased conversions, and then they deploy more of their
budgets to us.
In terms of ad load specifically, I would talk about last year with the ramp of
Reels, we had an opportunity -- originally the Reels ad load was quite low. So
we ramped ad load pretty aggressively last year as we were catching up and
working on improving our monetization efficiency of Reels relative to Stories
and Feed.
There’s still some opportunity there. As Susan has mentioned, the relative
ad load on Reels is still lower that Feed and Stories. And so, we’ll look for
opportunities to increase that, but to do it in a way that minimizes the
negative impact on engagement.
But in terms of the big opportunity to ramp ad load, particularly in Reels, that
was mostly captured in 2023. So I don't think you should expect to see ad
load increases like that continue in 2024.
More broadly, I’d say ad load is generally not as good of a metric as it once
was as we’re just working more on optimizing ads where we look at this on a

-- 6 of 13 --

7
more personalized basis. We try to figure who is more interested in seeing
ads, especially at a given point in time where they may be in a particular
point of consideration and then we can serve them more ads for that type of
product that they’re considering. So, it’s just much more nuanced as
opposed to, maybe in years past, we could talk about very broad-based ad
load increases. But we’re just in a point now where you look at with a lot
more sophistication.
Operator: Your next question comes from the line of Justin Patterson from KeyBanc at
Capital Markets. Please go ahead.
Justin Patterson: All right, hopefully the second time’s the charm here. I just wanted to unpack
the Q2 revenue guide a little bit more. I know there have been a lot of
moving parts with respect to macro, events in the Middle East, FX has moved
a little bit. And then there’s also just the calendar factors, Easter shift, Leap
Year. So, could you just talk about some of those factors influence the Q2
guide? Thank you.
Susan Li: Hi, Justin. So again, going into Q2, we continue to see healthy advertising
demand, strong engagement trends, ad performance improvements, all of
those we expect will continue to contribute to revenue growth.
However, the outlook reflects really the lapping dynamics, stronger demand,
in particular, from China-based advertisers who were, at the time,
reaccelerating from pandemic headwinds, as well as stronger impression
growth last year. Last year benefited from Reels ad load optimizations, kind
of working down that Reels tailwind and the user -- the Reels headwind,
sorry, previously, and then, the strong user engagement tailwinds.
You mentioned, obviously, we expect currency to be a headwind this quarter,
Q1 benefited from having an extra day which Q2 does not have. But really,
our long-term outlook continues to be really grounded in healthy drivers
across both engagement and sort of ongoing ads performance.
Operator: Your next question comes from line of John Blackledge from TD Cowen.
Please go ahead.

-- 7 of 13 --

8
John Blackledge: Great, thanks. With conversions growing faster than impressions, do you see
an immediate price uptick or is there like a lag on pricing upside for Meta in
that scenario? And would you expect this dynamic of conversions growing
faster than impressions to continue into 2Q and the rest of the year? And
then just quickly on the buyback, $15 billion repurchased in 1Q, it was above
the free cash flow that Meta generated. How should we think about kind of
cadence of buybacks going forward? Thank you.
Susan Li: First of all, thank you, John, for giving me an opportunity to clarify a
comment that I made on the call that I didn’t make as clearly as I should
have.
So, what I meant to say, in fact, conversion growth accelerated even as
impression growth slowed modestly in Q1. So, that, similar to what I said on
the call, reflects improved conversion rates. Now, what the impact is on
CPMs I think is a little bit hard to say because not all conversions are created
equal.
So, different types of conversions can have quite different values, whether
it’s sort of an app install for example or some kind of web purchase, those
will likely have quite different value. So, there is some question about the
mix and the mix composition of conversions will have a big impact on pricing.
But broadly speaking, driving growth in conversions and, in particular, at a
faster rate than the underlying supply gives us confidence that we are
continuing to improve ad performance over time and deliver more value to
advertisers.
On -- your second question is on buybacks. So, as we look at how to deploy
our free cash flow, investing in our core business opportunities, obviously,
remains our primary focus. We see compelling investment opportunities
ahead, including both with AI and Reality Labs.
And with that said, we generally believe that our strong financial position is
going to enable us to support both of those initiatives while continuing to
return capital to shareholders. Within the capital return program, obviously,

-- 8 of 13 --

9
we mostly rely on share repurchases. We initiated a modest dividend last
quarter.
One dynamic that is worth being aware of is, in Q1, that is when we issue our
annual equity refresher grants to employees. So we have typically bought
back more shares in Q1 in that period basically to neutralize the dilution that
comes from stock based compensation.
Operator: Your next question comes from the line of Dan Salmon from New Street
Research. Please go ahead.
Dan Salmon: Great. Thanks. Good afternoon everyone. Thanks, Ken, first for the color on
paid messaging earlier. I’ll maybe follow up at the end with more questions
on that but my question was for Susan and the color that you provided on
the public call around Lattice, which was very interesting.
In the past you’ve called out new surfaces, ad products, new ad formats that
have contributed $10 billion in revenue like Reels, the Advantage tools, click
to message ads in the past. Do you see Lattice as something that you think
can have that type of impact on your ad business or is there any color you
can add to how it’s impacted ad revenue growth? Thank you.
Susan Li: Thanks, Dan. So what I would say is in general we haven’t really quantified, I
think, the gains that come from the ranking architecture, whether it’s on the
ad side or on the engagement side since both of them -- I mentioned we
have sort of a similar ranking model unification project going on, underway in
the organic side as well as Meta Lattice on the ad side.
But in general these represent the next generation of continued refinement
on our models. And for as long as I’ve been here, which is quite a long time,
we’ve made the ad models more sophisticated over time, leveraging both
more capacity as we’ve built that out over time, as well as new and more
refined machine learning techniques and then more creative ways that we
apply those techniques against the different surfaces that we have.
So I think the thing that I would say about these types of architecture
unifications is they’re very scalable because they make both the cost

-- 9 of 13 --

10
structure more efficient and they enable us to sort of accrue gains across the
whole platform instead of sort of something that is specific to Reels or
specific to Stories or specific to Feed.
So I would say -- again, we’re not quantifying this, but Lattice as well as the
organic analog are good examples of where we believe that our core AI work
is generating very positive returns for the business.
Operator: Your next question comes from James Lee from Mizuho. Please go ahead.
James Lee: Great. Thanks for taking my questions. One more question on CapEx here.
Should we think about maybe CapEx increase for computing specifically is
most related to model training or maybe model inferencing?
And also related to that, how should we think about your expectation for
custom silicon? Is that included in the ’24 CapEx guidance or is it more
beyond that? Thank you.
Susan Li: Yes. So in the near-term, the acceleration of infrastructure capacity that
we’re building is oriented more toward being able to train bigger and more
complex next generation foundation models.
But having said that, again, the capacity is fungible. So you can use it both
for training, and then you can make adaptations to use it for inference and
other use cases. But I expect that the first use case will likely be for training
future generations of foundation models.
On the second part of your question, which is about custom silicon, so MTIA
is factored in our 2024 guidance. Our in-house MTIA program is initially
focused on supporting our core ranking and recommendation inference
workloads.
We’re starting to support ads and organic content recommendations with
MTIA this year, and we are expecting to increase MTIA’s share of
recommendation workloads over the course of this year and into 2025.

-- 10 of 13 --

11
And I hope in the future that we will also be able to expand and to support
training workloads and, over time, some targeted GenAI use cases as well.
Operator: Your next question comes from the line of Brent Thill from Jefferies. Please
go ahead.
Brent Thill: Thanks. I was curious if you could just speak to China and what you're seeing
in these multinational advertisers. Anything change? I think last year you
mentioned it was roughly 10% of revenue. Have you seen any pickup or
degradation in terms of activity there?
Susan Li: We didn’t quantify this this quarter, and this won't be a quarterly
quantification for us. But what I would say is growth and spend from China
advertisers remains strong. Online commerce and gaming are still strong
verticals there, and you can see that our Asia-Pacific advertiser segment
continues to be the fastest growing region at 41% year-over-year in Q1.
Well, what I would also say is, again, this is just a tougher period of comps on
the China-based advertiser growth because of the demand reacceleration in
2023 as they came out of their pandemic-driven headwinds. But over the
long-term, this has been a long run trend of growth for us with this pandemic
blip.
And so, again, I think we feel good about the long run opportunity here but
recognize that it is a more volatile -- it has been a more volatile revenue
stream for us, and it could be subject to more volatility going forward.
Operator: Your next question comes from the line of Rob Sanderson from Loop Capital
Markets. Please go ahead.
Rob Sanderson: Yes. Good afternoon. Thank you for taking the question. It’s been quite
awhile now since the IDFA was deprecated, and obviously there are still
other layers of cookies and other identifiers that may become more limited
over time.
So the big focus, obviously, for Meta over the past several years to become
less dependent on third parties for measurement. Susan, how would you

-- 11 of 13 --

12
characterize how well insulated the business is these days after all this work
from dependence on third parties or signal loss or any other sort of platform
changes that might be outside the company’s control?
Susan Li: Yes. I mean, I think in the period of time since Apple implemented the iOS
14.5 changes, we have really rebuilt our systems to perform well even when
we have less access to granular data. And we feel pretty confident that both
our rebuilt systems and the advancements that we've made in AI and lot of
the model and ranking advancements we’ve made will be more resilient to
future changes.
Having said that, mobile advertising obviously remains an evolving landscape.
We continue to stay closely on top of this. We continue to invest heavily in
the tools such as Conversions API that we have built, and our focus on those
tools and continuing to onboard advertisers to those tools has not lessened.
Kenneth Dorell: Krista, we have time for one last question.
Operator: And that question comes from the line of Jason Helfstein from Oppenheimer.
Please go ahead.
Chad Larkin: Hey, thanks and this is Chad Larkin for Jason. So, Mark talked a lot about
positive datapoints around AI glasses. Just anything you share there that is
kind of helping you lean in. Thanks.
Susan Li: Hi, thanks, Chad. Yes, so we’ve been quite happy with sales and just
generally high demand for the Ray-Ban Meta smart glasses. We’ve seen
many styles and colors sell out periodically and we’ve been working with our
retail and channel partners to keep those in stock.
We’ve also been seeing strong engagement and retention with the smart
glasses. Upgrades that we made to core features from the Ray-Ban Stories
have really resonated with users. Things like open ear audio, hands-free
photo/video capture, livestreaming to Facebook and Instagram, et cetera.
So, the positive reception we think is also a testament to our decision to
embed Meta AI into our smart glasses being a good one.

-- 12 of 13 --

13
We’re excited to have rolled out multimodal AI to Ray-Ban Meta Glasses in
the U.S. and Canada. That’s going to allow folks to ask Meta AI questions
about content that’s in their field of view. So, can it translate a sign, or what
can I make with this ingredient in front of me? And we believe that glasses
will be a really good form factor for carrying around an AI assistant, because,
obviously, it can see what you see and hear what you hear.
And I think we’re both going to learn a lot from kind of the initial feedback
loop we have on usage, in terms of making it better, and then, we hope to
really leverage this platform to build really useful and delightful experiences
for users who are wearing them.
Kenneth Dorell: Great. Thank you all for joining us today. We appreciate your time and we
look forward to speaking with everyone again soon.
Operator: This concludes today’s conference call. Thank you for joining us. You may
now disconnect your lines.

-- 13 of 13 --