Q4 2024 Amazon.com Inc Earnings Call
Thank you for standing by.
Good day, everyone and welcome to the amazon.comfourth quarter 2024 financial results teleconference.
Speaker Change: At this time all participants are in a listen only mode. After the presentation. We will conduct a question and answer session. Today's call is being recorded and for opening remarks I'll be turning the call over to the Vice President of Investor Relations. Dave files. Thank you Sir Please go ahead.
Speaker Change: Hello, and welcome to our Q4 2024 financial results conference call joining us today to answer your questions as Andy Jaffe, our CEO and Brian <unk>, Our CFO, how did you listen to today's conference call. We encourage you to have our press release in front of you, which includes our financial results as well as metrics and commentary on the quarter. Please note unless otherwise stated.
Speaker Change: All comparisons in this call will be against our results for the comparable period of 2023, our comments and responses to your questions reflect management's views as of today February six 2025 only and.
Speaker Change: We will include forward looking statements actual results may differ materially.
Speaker Change: Additional information about factors that could potentially impact our financial results is included in today's press release, and our filings with the SEC, including our most recent annual report on Form 10-K, and subsequent filings. During this call. We may discuss certain non-GAAP financial measures in our press release slides accompanying this webcast and our filings with the SEC each of which is posted.
Speaker Change: On our IR website, you will find additional disclosures regarding these non-GAAP measures, including reconciliations of these measures with comparable GAAP measures our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our results are inherently unpredictable and maybe materially affected by many factors, including fluctuations in foreign exchange.
Speaker Change: Rates changes in global economic and geopolitical conditions and customer demand and spending including the impact of recessionary fears inflation interest rates regional labor market constraints world events the rate of growth of the Internet online Commerce cloud services, and new and emerging technologies and the various factors detailed in our filings with the SEC.
Speaker Change: C.
Speaker Change: Our guidance assumes among other things that we don't conclude any additional business acquisitions restructurings or legal settlements, it's not possible to accurately predict demand for our goods and services and therefore, our actual results could differ materially from our guidance.
Andy Jaffe: Now I'll turn the call over to Andy.
Andy Jaffe: Thanks, Dave.
Andy Jaffe: We're reporting a 187 $8 billion in revenue up 10% year over year, given the way the dollar strengthened throughout the quarter, we saw a $700 million more a foreign exchange headwind than we anticipated a guidance without that headwind revenue would have been 11% year over year and exceeded the top end of our guidance.
Andy Jaffe: Operating income was $21 $2 billion up 61% year over year, and trailing 12 month free cash flow adjusted for equipment Finance leases was $36 2 billion up $700 million year over year.
Andy Jaffe: We're pleased with the invention customer experience improvements and results delivered in 2024 and have a lot more planned in 2025.
Andy Jaffe: I'll start by talking about our stores business.
Andy Jaffe: We saw a 10% year over year revenue growth in our North America segment, and 9% year over year in our international segment, excluding the impact from foreign exchange rates. Our continued focus on expanding selection lowering prices and improving convenience drove strong unit growth that even outpaced our revenue growth.
Andy Jaffe: We continue to add to our broad range of selection, giving customers choice across a variety of price points.
Andy Jaffe: We welcome notable brands to our store throughout 2024, including Clinique, Este, Lauder or rings and Armani beauty.
Andy Jaffe: We continue to add to the hundreds of millions of products offered from our selling partners, who made up 61% of items that we sold in 2024, our highest annual mix of third party seller units ever.
Andy Jaffe: We also launched Amazon a hall for U S customers in Q4, which offers customers and engaging shopping experience they bring ultra low price products into one convenient destination.
Andy Jaffe: It's off to a very strong start.
Andy Jaffe: Customers continue to want Amazon to be the place they rely on for sharp pricing in the fourth quarter consumers save more than $15 billion with Argo everyday prices and record setting events during the prime Big deal days in October and Black Friday, and cyber Monday around Thanksgiving. Additionally.
Andy Jaffe: Additionally, profiteroles annual pricing study found that entering the holiday season, Amazon had the lowest online prices for the eighth year in a row, averaging 14% lower prices on average than other leading retailers in the U S.
Andy Jaffe: Our speed of delivery continues to accelerate in 2024. It was another record setting year for prime members we.
Andy Jaffe: We expanded the number of same day delivery sites by more than 60% in 2024, which now serve more than 140 metro areas and overall, we delivered over 9 billion units the same or next day around the world.
Andy Jaffe: Our relentless pursuit of better selection price and delivery speed is driving accelerated growth in prime membership.
Andy Jaffe: $14 99, a month prime members get unlimited free shipping on 300 million items, often the same day or one day delivery exclusive shopping events like Prime day access to a vast collection of premium programming and live sports in Prime video AD.
Andy Jaffe: AD free listening of 100 million songs in podcast with Amazon music.
Andy Jaffe: Access to unlimited generic prescriptions for only $5 a month unlimited grocery delivery on orders over $35 from whole foods, Mark and Amazon fresh for $9 99, a month.
Andy Jaffe: Our free Grubhub, plus membership with free unlimited delivery and our latest benefit of a 10 cents per gallon fuel discount of BP Amoco and A&P am stations.
Andy Jaffe: When you think about this as a whole and also compared to many other membership services that are comparably a more expansively priced offer just one benefit like video Prime is a screaming deal and we have more coming for our prime members in 2025.
Andy Jaffe: We also remain squarely focused on cost to serve in our fulfillment network, which has been a meaningful driver of our increased operating income we've talked about the regionalization of our U S network. We've also recently rolled out a redesign U S inbound network.
Andy Jaffe: While still in its early stages, our inbound efforts have improved our placement of inventory so that even more items are closer to end customers.
Ahead of Black Friday in November we had improved the percentage of ordered units available in the ideal building by over 40% year over year.
Andy Jaffe: We've also spent considerable time optimizing the number of items, we think customers in the same package, which reduces packaging is more convenient for customers and less expensive for us to fulfill.
Andy Jaffe: And our per unit transportation cost continue to decline as we build out and optimize our last mile network.
Andy Jaffe: Overall, we've reduced our global cost to serve on a per unit basis for the second year in a row, while at the same time, increasing speed improving safety and adding selection.
Andy Jaffe: As we look to 2025 and beyond we see opportunity to reduce costs again as we further refine inventory placement grow our same day delivery network accelerate robotics and automation throughout the network.
Andy Jaffe: In advertising, we remain pleased with the strong growth on a very large base generating $17 $3 billion of revenue in the quarter growing 18% year over year.
Andy Jaffe: That's a $69 billion annual revenue run rate more than double what it was just four years ago at 29 billion.
Andy Jaffe: Sponsored products the largest portion of that revenue are doing well and we see runway for even more growth.
Andy Jaffe: We also have a number of newer streaming offerings that are starting to become significant new revenue sources and the streaming video side, we've wrapped up our first year of Prime video ads and we're quite pleased with the early progress and head into this year with momentum.
Andy Jaffe: We've made it easier to do full funnel advertising with us.
Andy Jaffe: The funnel is from the top of the funnel with broad reach advertising that drives brand awareness to mid funnel.
Andy Jaffe: Sponsored brands the company specify certain keywords and audiences to attract people to their detailed pages or brand store and Amazon to bottom of the funnel, where sponsored products help advertisers surface relevant product ads to customers at the point of purchase.
Andy Jaffe: We make this easy for brands to sign up for and to play across our growing advertising.
Andy Jaffe: We also have differentiated audience features as average billions of customer signals across our stores and media destinations from Amazon marketing cloud secure clean rooms, providing advertisers the ability to analyze data produced core marketing metrics and understand how their marketing performs across various channels to our new multi touch attribution.
Andy Jaffe: Houston model.
Andy Jaffe: Advertisers understand how their marketing is working.
Andy Jaffe: If an advertiser uses streaming TV display sponsored products and other AD types in their campaign multi touch attribution will show the relative contribution of each to their sales.
Speaker Change: Moving on to AWS.
Speaker Change: In Q4, AWS grew 19% year over year, and now has a $115 billion annualized revenue run rate.
Speaker Change: AWS is a reasonably large business by most folks standards and though we expect growth will be lumpy over the next few years as enterprise adoption cycles capacity considerations and technology advancements impact timing, it's hard to overstate how optimistic we are about what lies ahead for AWS customers and business.
Speaker Change: I spent a fair bit of time thinking several years out.
Speaker Change: Well, maybe hard for some to fathom a world where virtually every app has generative AI infused in it with inference being a core building block just like compute storage and database and most companies having their own agents that accomplish various tasks and interact with one. Another this is the world. We're thinking about all the time and we continue to believe that this world with <unk>.
Speaker Change: Mostly be built on top of the cloud with the largest portion of it on AWS.
Speaker Change: To best help customers realize this future need powerful capabilities, all three layers of the stack.
Speaker Change: The bottom there for those building models unique compelling chips chips to the key ingredient in the compute that drives training and inference.
Speaker Change: AI compute has been driven by Nvidia chips, and we obviously have a deep partnership with Nvidia and will for as long as we can see into the future. However, there arent that many generative AI applications of large scale yet and.
When you get there as we have with apps like <unk> and Rufus cost can get steep quickly customers want better price performance and it's why we built our own customer AI silicon training them to just squash at our AWS re invent conference in December and easy to instances with these chips are typically 30% to 40% more price performance.
Speaker Change: Other current GPU powered instances available that's very compelling and scale several technically capable companies like Adobe data bricks poolside and Qualcomm have seen impressive results in early testing of training them too.
Speaker Change: It's also why Youre seeing anthropic build their future frontier models and training them to.
Speaker Change: We're collaborating with anthropic to build project Rainier.
Speaker Change: Cluster of training them to ultra servers containing hundreds of thousands of training in two chips. This cluster is going to be five times the number of extra thoughts as the cluster. The anthropic used to train their current leading set of cloud models.
Speaker Change: We're already hard at work on training <unk>, III, which we expect to preview late in 'twenty, five and defining training them for thereafter.
Speaker Change: <unk> outstanding performance chips that deliver leading price performance has become a core strength of AWS is starting with our nitro and graviton chips in our core business and now extending to training them and AI is something unique to AWS relative to other competing cloud providers.
Speaker Change: The other key component for model builders is services and make it easier to construct their models I won't spend a lot of time in these comments on Amazon's H maker, AI, which has become the go to service for AI model builders to manage their AI data build models experiment and deploy these models except to say this HMA.
Speaker Change: Occurs hyper pod capability, which automatically splits training workloads across many AI accelerators prevents interruptions by periodically saving checkpoints and automatically repairing faulty instances from their last saved checkpoint and saving training time by up to 40%. It continues to be a differentiator received several new.
Speaker Change: <unk> capability that reinvent, including the ability to manage costs, a cluster level and prioritize which workloads should receive capacity when budgets are reach and is increasingly being adopted by model builders.
Speaker Change: At the middle there for those wanting to leverage frontier models to build jet AI apps Amazon bedrock, because our fully managed service that offers the broadest choice of high performing foundation models with the most compelling set of features that make it easy to build a high quality generative AI application.
Speaker Change: We continue to iterate quickly on bedrock announcing lunar AI pool side and over 100 other popular emerging models to bedrock and reinvent it.
Speaker Change: In short order. We also just added deep seeks our one models to bedrock and Sage maker and Additionally, we delivered several compelling new bedrock features of reinvent, including prompt caching intelligent prompt routing and modeled installation all of which help customers achieve lower cost and latency and their influence.
Speaker Change: Like Sage maker AI bedrock is growing quickly and resonating strongly with customers.
Speaker Change: Related we also just launched Amazon's own family of frontier models in bedrock called Nova.
Speaker Change: These models compare favorably in intelligence against the leading miles the world put offer lower latency lower price about 75% lower than other models in bedrock and our integrated with key bedrock features like fine tuning.
Speaker Change: Modeled installation knowledge bases, Iraq, Hey, Jed take capabilities.
Speaker Change: <unk> of AWS customers are already taking advantage of the capabilities and price performance of Amazon, Nova models, including Palin tier two really.
Speaker Change: Dentsu Fortinet.
Speaker Change: Alex and Robinhood and we've just gotten started.
Speaker Change: At the top layer of the stack.
Speaker Change: Amazon Q is the most capable generative AI powered assistant for software development and to leverage their own data you.
Speaker Change: You may remember that on the last call I shared the very practical use case, where Q transform helps save Amazon's teams $260 million and 4500 developer years in migrating over 30000 applications to new versions of the Java J D K.
Speaker Change: This is real value in companies asked or more which we obliged with a recent deliveries of Q transformations that enable moves from windows net applications to Linux Vmware, it's easy to accelerates mainframe migrations early customer testing indicates the queue can turn was going to be a multiyear effort to do a mainframe.
Speaker Change: <unk> into a multi quarter effort cutting by more than 50% time to migrate mainframes.
Speaker Change: Big deal in these transformations are good examples of practical AI well.
Speaker Change: AI continues to be a compelling new driver in the business. We haven't lost our focus on core modernization of companies technology infrastructure from on premises to cloud.
Speaker Change: We signed new AWS agreements with companies, including Intuit, Paypal Norwegian cruise line holdings.
Speaker Change: Northrop Grumman The Guardian life Insurance Company of America Red It Japan Airlines Baker Hughes, The Hertz Corporation Redfin chime financial.
Donna and many others.
Speaker Change: Consistent customer feedback from our recent AWS reinvent gathering was appreciation that we're still inventing rapidly and non AI infrastructure areas like storage.
Speaker Change: Database analytics are.
Speaker Change: Our functionality leadership continues to expand and there were several key launches customers were above about including.
Speaker Change: Amazon Aurora D sequel, our New server list distributors sequel database that enables applications with the highest availability.
Speaker Change: <unk> consistency post dress compatibility and four times faster reason rates compared to other popular distributed sequel databases.
Speaker Change: Amazon has three tables, which make us three the first cloud object store with fully managed support for Apache iceberg for faster analytics.
Speaker Change: Amazon is III metadata, which automatically generates query Abel metadata simplifying data discovery business analytics, and real time inference to help customers unlock the value of their data and has three.
Speaker Change: And the next generation of Amazon Sage maker, which brings together all the data analytics services and AI services into one interface to do analytics and AI more easily scale.
Speaker Change: As 2024 comes to an end I want to thank our teammates and partners for their meaningful impact throughout the year.
Speaker Change: It was a very successful year across almost any dimension you pick.
Speaker Change: We're far from done and look forward to delivering for customers in 2025.
Speaker Change: With that I'll turn it over to Brian for a financial update.
Brian: Thanks Sandy.
Speaker Change: Starting with our top line financial results Worldview.
Speaker Change: Worldwide revenue was $187 8 billion, an 11% increase year over year, excluding the impact of foreign exchange.
Speaker Change: This equates to an approximate $900 million headwind from FX in the quarter, which is about $700 million higher than what we'd anticipated in our Q4 guidance range.
Speaker Change: Excluding that additional FX headwind, we would have exceeded the top end of our revenue guidance range.
Speaker Change: Worldwide operating income was $21 2 billion, our largest operating income quarter ever and was $1 2 billion above the high end of our guidance range.
Speaker Change: All segments, we continue to innovate for customers, while operating more efficiently at the same time.
Speaker Change: In the North America segment fourth quarter revenue was $115 6 billion, an increase of 10% year over year.
Speaker Change: International segment revenue was $43 4 billion.
Speaker Change: An increase of 9% year over year, excluding the impact of foreign exchange.
Speaker Change: Worldwide paid units grew 11% year over year.
Speaker Change: Our focus on low prices broad selection and fast shipping continues to resonate with customers.
Speaker Change: Shifting to profitability North America segment operating income was $9 3 billion.
Speaker Change: An increase of $2 $8 billion year over year.
Speaker Change: Operating margin was 8% up 190 basis points year over year.
Speaker Change: In the International segment operating income was $1 3 billion, an improvement of $1 $7 billion year over year.
Speaker Change: Operating margin was 3%.
Speaker Change: A 400 basis points year over year.
Speaker Change: Marks the eighth consecutive quarter, where we've seen year over year margin improvement in both North America and international segments.
Speaker Change: 2024 also marks the second year in a row.
Speaker Change: We've lowered our global cost to serve on a per unit basis.
Speaker Change: In the fourth quarter, we saw strong productivity in our transportation network from improved inventory placement higher units per package and reduced travel distances.
Speaker Change: We also saw improved productivity in our fulfillment centers.
Speaker Change: Overall, our teams executed extremely well throughout the quarter and particularly during our peak seasons.
Speaker Change: I want to thank them for all they do to deliver for our customers.
Speaker Change: Looking ahead, we have several opportunities to keep lowering our costs to even better inventory placement, which also allows us to deliver items to customers faster.
Speaker Change: In the U S for tuning, our inbound network and continuing to expand our same day delivery network.
Speaker Change: Globally, we're adding automation and robotics throughout our network.
Speaker Change: While these efforts will take time to implement and progress may not be linear we have a good plan to continue to drive improvements in our cost structure.
Speaker Change: Advertising remains an important contributor to profitability in the North America and international segments.
Speaker Change: This quarter, we saw strong advertising revenue growth on an increasingly large base.
Speaker Change: We will also continue to invest in experiences that have potential to be important to customers and Amazon long term.
Speaker Change: In areas like Alexa healthcare and grocery as.
Speaker Change: As well as kuiper, including the planned launches of our production satellites in the coming months.
Speaker Change: As a reminder, we currently expense the majority of the costs associated with the development of our satellite network.
Speaker Change: We will capitalize certain costs once the service chiefs commercial viability, including sales to customers.
Speaker Change: Moving next to our AWS segment revenue was $28 8 billion.
Speaker Change: An increase of 19% year over year.
Speaker Change: AWS now has an annualized revenue run rate of $115 billion.
Speaker Change: During the fourth quarter, we continued to see growth in both generative AI and non generative AI offerings as companies turn their attention to newer initiatives bring more workloads to the cloud.
Speaker Change: We start or accelerate existing migrations from on premise to the cloud and tap into the power of generative AI.
Speaker Change: Customers recognize to get the full benefit of generative AI they have to move to the cloud.
Speaker Change: AWS reported operating income of $10 6 billion, an increase of $3 $5 billion year over year.
Speaker Change: This is a result of strong growth.
Speaker Change: Innovation in our software and infrastructure to drive efficiencies.
Speaker Change: Continued focus on cost control across the business.
Speaker Change: As we've said in the past.
Speaker Change: Specced AWS operating margins to fluctuate over time, driven in part by the level of investments we're making.
Speaker Change: Additionally, we increased the estimated useful life of our servers, starting in 2024, which contributed approximately 200 basis points to the AWS margin increase year over year in Q4.
Speaker Change: Now turning to our capital investments.
Speaker Change: As a reminder, we define these as a combination of cash capex plus equipment finance leases cap.
Speaker Change: Capital investments were $26 3 billion in the fourth quarter.
Speaker Change: And we think that run rate will be reasonably representative of our 2025 capital investment rate.
Speaker Change: Similar to 2020 for the majority of the spend will be to support the growing need for technology infrastructure.
Speaker Change: This primarily relates to AWS, including to support demand for our AI services as well as tech infrastructure to support our North America and international segments.
Speaker Change: Additionally, we're continuing to invest in capacity for fulfillment and transportation network to support future growth.
Speaker Change: We're also investing in same day delivery facilities, and our inbound network as well as robotics and automation to improve delivery speeds and to lower our cost to serve.
Speaker Change: These capital investments will support growth for many years to come.
Speaker Change: Turning to our revenue guidance for Q1 net sales are expected to be between $151 billion and $155 5 billion.
Speaker Change: I'd like to highlight two items impacting our Q1 revenue guidance.
Speaker Change: First we estimate the year over year impact of changes in foreign exchange rates based on current rates, which we expect to be a headwind of approximately $2 $1 billion in Q1 year over year or 150 basis points.
Speaker Change: As a reminder, global currencies can fluctuate during the quarter and just as we saw in Q4 with the strengthening of the dollar versus most other currencies.
Speaker Change: Second a reminder, that we are comping the impact of last years leap year.
Speaker Change: The extra day contributed approximately $1 5 billion of additional net sales across our businesses in Q1, 2024 or about 120 basis points to the year over year growth rate, which impacted all segments.
Speaker Change: Q1 operating income is expected to be between $14 billion $18 billion.
Speaker Change: This guidance includes the estimated impact of certain updates to the useful life of our fixed assets.
Speaker Change: I'll provide a bit more detail in a moment, but on an aggregate basis. We estimate this will decrease full year 2025 operating income by approximately $400 million for the assets on our balance sheet as of December 31, 2024.
First in Q4, we completed a useful life study for our servers and networking equipment and observed an increased pace of technology development, particularly in the area of artificial intelligence and machine learning.
Speaker Change: As a result, we are decreasing the useful life for a subset of our servers and networking equipment from six years to five years beginning in January 2025.
Speaker Change: We anticipate this will decrease full year 2025 operating income by approximately $700 million.
Speaker Change: In addition, we also early retired a subset of our servers and network equipment.
Speaker Change: Q4, 2024 expense of approximately $920 million from accelerated depreciation and related charges.
Speaker Change: And expect this will also decrease full year 2025 operating income by approximately $600 million.
Speaker Change: Both of these server and network equipment useful life changes primarily impact our AWS segment.
Speaker Change: Lastly, we also completed the useful life study for certain types of heavy equipment used in our fulfillment centers and are increasingly useful life from 10 years to 13 years beginning in January 2025.
Speaker Change: We anticipate this will increase full year 2025 operating income by approximately $900 million.
Speaker Change: As we turn the page to 2025, we're energized by the great work our teams have delivered.
Speaker Change: We'll remain focused on driving even better customer experience and we believe putting customers first is the only reliable way to create lasting value for our shareholders.
Speaker Change: With that let's move on to your questions.
Speaker Change: At this time, we will now open the call up for questions. We ask each caller to please limit yourself to one question.
Speaker Change: I would like to ask a question. Please press Star then one on your keypad.
Speaker Change: When you pose your question you pick up your handsets to provide optimum sound quality.
Speaker Change: Once again to initiate a question. Please press Star then one on your Touchtone telephone at this time.
Speaker Change: Please hold while we pull for questions.
Speaker Change: Yeah.
Speaker Change: Thank you. Our first question comes from the line of Mark Mahaney with Evercore ISI. Please proceed with your question.
Speaker Change: Two quick questions. So Brian that's 100 billion Capex, we should think about <unk> 25, and then Andy was there any so would you describe the AWS growth is being currently moderated down by supply constraints do you see those across the industry or do you see those materially impacting AWS today. Thank you very much.
Speaker Change: So I'll take both those sandy.
Speaker Change: On the Capex side.
Brian: As Brian mentioned earlier.
We spent $26 3 billion in Capex in Q4, and I think that is reasonably representative of what you would expect an annualized capex rate in 2025.
Brian: The vast majority of that Capex spend is on AI for AWS.
Brian: It's.
Brian: The way the AWS.
Brian: <unk> works with cash cycle works is that the faster we grow the more capex. We ended up spending because we have to procure data center and hardware and chips and networking gear ahead of when we're able to monetize it we don't procure it unless we see significant signals of demand.
Brian: And so when AWS is expanding its capex, particularly in what we think is one of these once in a lifetime type of business opportunities like AI represents I think it's actually quite a good sign medium to long term for the AWS business.
Brian: And I actually think that spending this capital to pursue this opportunity which from our perspective, we think.
Brian: Every application that we know of today is going to be reinvented with AI inside of it and with influence being a core building block just like compute and storage and database.
Brian: If you believe that plus it altogether new experiences that we have only dreamed about are going to actually be available to us with AI.
Brian: AI represents for sure the biggest opportunity since cloud and probably the biggest technology shift in opportunity in business since the internet and so.
Brian: That both our business our customers and shareholders will be happy medium to long term that we're pursuing the capital opportunity and the business opportunity in AI. We also have capex that we're spending this year in our stores business really with an aim towards trying to continue to improve.
Brian: The delivery speed and our cost to serve and so youll see us expanding the number of same day facilities from where we are right now Youll also see us expand the number of delivery stations that we havent rural areas, we can get items to people who live in rural areas much more quickly and then a pretty significant investment as well on robotic.
Brian: And automation, so we can take our cost to serve down and continue to improve our productivity.
So thats the Capex piece I think the second question you asked Mark.
Brian: It's really around AWS growth and whether this is being moderated down at all by supply chain constraints.
Brian: Well, it's it is hard to complain when you have a.
Brian: Multibillion dollar annualized revenue run rate business and AI.
Brian: We do.
Brian: It is growing triple digit percentage year over year, it's hard to complain.
Brian: However, it is true that we could be growing faster if not for some of the constraints on capacity.
Brian: Come in the form.
Of.
Brian: I would say chips from our third party partners come in a little bit slower than before with a lot of midstream changes that take a little bit of time to get the.
Brian: The hardware actually yielding.
Brian: The percentage healthy and high quality.
Brian: Servers, we expect.
Brian: It comes with our own big.
Brian: New launch of our own hardware and our own ships and training them too.
We just went to general availability at reinvent, but the majority of the volume is coming in really over the next couple of quarters and next few months.
Brian: It comes in the form of.
Brian: Power constraints, where I think the world is still constrained on power from from where I think we all believe we could serve customers. If we were unconstrained theres some components in the supply chain like motherboards two that are a little bit.
Brian: Short in supply for various types of servers. So.
Brian: I think the team has done a really good job scrapping and providing capacity for our customers. So they can grow we're still growing at a pretty reasonable clip as I mentioned earlier, but I do think we could be growing faster if we were unconstrained.
Brian: Riddick those constraints really start to relax in the in the second half of 'twenty five.
Brian: And as I said, I think we could be growing faster, even though we're growing at pretty good clip today.
Speaker Change: And our next question is from Eric Sheridan with Goldman Sachs. Please proceed.
Eric Sheridan: Thanks, so much for taking the question I'll just ask one that's building on <unk> questions there.
Speaker Change: Andy when you think about the news that came out of China over the last couple of weeks and think longer term about bending the cost curve lower with AI I understood. The commentary around Capex for 2025, but when you look at where you sit in the industry. The move towards open source elements of custom Silicon how do you think about bending the cost curve in either.
Speaker Change: Speeding up our amplifying time deployment to market or possibly higher returns on capital.
Speaker Change: Thanks, so much.
Speaker Change: Yes.
Speaker Change: Well I'd say a few things because there are a few questions built into that first of all.
Speaker Change: I think like many others.
Speaker Change: We were impressed with what <unk> has done I think in part impressed with some of the training techniques primarily in slipping.
Speaker Change: The sequencing of.
Speaker Change: Reinforcement training and reinforcement learning being earlier without the human in the loop.
Speaker Change: We thought that was interesting ahead of the supervised fine tuning.
Speaker Change: We also thought some of the influence.
Speaker Change: Optimization. They did we're also quite interesting for those of US who are building frontier models, who are all working on the same types of things. We're all learning from one another I think you have seen and we will continue to see a lot of leapfrogging between US there is a lot of innovation to come and I.
Inc.
Speaker Change: If you run a business like AWS and you have a core belief like we do.
Speaker Change: That.
Speaker Change: Virtually all the big generative AI apps are going to use multiple model types.
Speaker Change: Different customers are going to use different models.
Speaker Change: For different types of workloads, you're going to provide as many.
Speaker Change: Leading frontier models as possible for customers to choose from and that's what we've done with services like Amazon bedrock and it's why we moved so quickly to make sure that deep sic was available both on bedrock and Sage maker.
Speaker Change: Fast and you saw from others and we already have customers starting to experiment with that I think one of the interesting things over the last.
Speaker Change: A couple of weeks is sometimes people.
Speaker Change: The assumptions there.
Speaker Change: That if you're able to.
Speaker Change: Decrease the cost of any type of technology component in this case, what we're really talking about in France.
Speaker Change: Somehow it is going to lead to less total spend in technology and we just we have never seen that to be the case.
Speaker Change: We did the same thing in the cloud, where we launched AWS in 2006, where we offered S. Three object storage for 15, gigabyte and compute for 10 cents narrow which of course is much lower now.
Speaker Change: Years later.
Speaker Change: People thought that people would span a lot less money in the infrastructure technology, what happens is key.
Speaker Change: Companies will span a lot last per unit of infrastructure and that is very very useful for their businesses, but then they get excited about what else. They could build that they always thought with cost prohibited before and they usually end up spending a lot more in total.
Speaker Change: On technology once you make the per unit cost less and I think that is very much what's going to happen here.
Speaker Change: Which is the cost in France will substantially come down.
Speaker Change: <unk> heard the last couple of weeks at <unk> as a piece of it but everybody is working on this I believe the cost of inference will meaningfully come down.
Speaker Change: It will make it much easier for companies to be able to fuse all of them.
Speaker Change: Their applications within France, and with generative AI and I think it's kind of if you run a business like we do where we want to make it as easy as possible for customers to be successful building customer experiences on top of our various infrastructure services.
Speaker Change: Cost of insurance coming down is going to be very positive for customers and for our business.
Speaker Change: And the next question is from Doug Anmuth with Jpmorgan. Please proceed.
Doug Anmuth: Thanks for taking the questions I'll stick with AWS to start just Brian maybe you can talk a little bit more about margins. There just given that they've kind of moved between the mid <unk> to high <unk> over the past two years, how should we think about that more normalized, especially as you're investing that much more in general with AI and then just on this.
Doug Anmuth: Store side can you talk about the impact of less volume going through.
Shipping partner EPS going forward and are you able to manage that incremental shipping that's required. Thanks.
Doug Anmuth: Yes sure Doug Thanks for your question.
Doug Anmuth: First on AWS, yes, we have seen a lot of.
Doug Anmuth: Fluctuation in operating margin in AWS, and we've said historically that they will be lumpy as I say over time.
Doug Anmuth: <unk>.
Doug Anmuth: The stage, we're in right now.
Doug Anmuth: AI is still early stage it does come.
Originally with lower margins in a heavy investment load is as we've talked about.
And then the short term over time that that should have a headwind on on margins, but over the long term, we feel that margins will be comparable to non AI business as well so.
Doug Anmuth: We're very pleased with the strong growth.
Doug Anmuth: Focus on driving efficiencies.
Doug Anmuth: In all of our data centers saving power, we're using power new generative AI applications, and just generally reducing costs. So.
Doug Anmuth: We're pleased with the performance of the AWS team and look forward to strong 2025.
Doug Anmuth: I'll take the U P S one which is really.
Doug Anmuth: UBS has been a partner of ours for many years and we expect that that will continue to be partners with UBS for many years.
Doug Anmuth: As you know increasingly over the last several years.
Doug Anmuth: Particularly accelerated by the pandemic.
Doug Anmuth: We have.
Doug Anmuth: Ship to much larger percentage of our shipments through logistics network, our own last mile Transportation network.
Doug Anmuth: Yes, I think thats in part.
Doug Anmuth: Because we needed to scale up so fast and the pandemic with with everything being shut down.
Doug Anmuth: And needing to serve more of the total.
Doug Anmuth: <unk> segment share of retail units during that time and needing to do it at a.
Doug Anmuth: Our low cost structure because of course, our customers expected oil prices and that's the nature of the business.
Doug Anmuth: UBS has decided that.
Oh, sorry.
Doug Anmuth: Serving Amazon is a lower margin for them and so I think they've walked away from some of the volume that they otherwise could have had in the partnership we're able to handle it with our own logistics capability and we'll see how it continues to evolve.
Speaker Change: And the next question comes from the line of Brian Nowak with Morgan Stanley. Please proceed.
Brian Nowak: Thanks for taking my questions, Andy maybe to drill a little bit into the robotics acceleration that you talked about any any new data points you can share on learnings from from Shreveport, and how do we think about sort of the scalability of savings of the timing that we could see a real impact on profitability from from Robotics, and then maybe just.
Brian Nowak: A bigger picture Gen AI GPU enabled changes if other examples of how you see the Amazon retail shopping experience changing throughout 2025 is your better using gen AI or GPU on a machine learning and things.
Brian Nowak: Yeah.
Brian Nowak: Okay, well on the robotics piece, what I would tell you is since we've been pretty substantially integrating robotics into our fulfillment network over the last many years, we have seen cost savings and we've seen productivity improvements and we've seen safe.
Brian Nowak: Safety improvements and so we have already gotten a significant amount of value out of our robotics.
Brian Nowak: Innovations, what we've seen recently and I think maybe part of what you're referencing in Shreveport is that the next.
Brian Nowak: Trosch of robotics.
Brian Nowak: Initiatives have started hitting production and we've put them all together for the first time as part of an experienced in our Shreveport facility.
Brian Nowak: We are very very encouraged by what we're seeing there by both the speed improvements that we're seeing the productivity improvements the cost to serve improvements.
Brian Nowak: It's still relatively early days in and these all being put together only in Shreveport at this point, but we have plans now to start to expand that and rolled it out to a number of other facilities in the network some of which will be our new facilities and others of which will we will retrofit existing facilities to be able to use there.
Brian Nowak: Same robotics innovations I'll also tell you that.
Brian Nowak: This group of call it a half dozen or so new initiatives is not close to the end of what we think is possible with respect to being able to use robotics to improve.
Brian Nowak: Productivity cost to serve its safety in our fulfillment network and we have kind of the next wave that we're starting to work on now, but I think this will be a many year effort as we continue to tune different parts of our fulfillment network, where we can use robotics and we we actually don't think there are that many things that we can't improve the experience.
Speaker Change: <unk> with robotics.
Speaker Change: On your other question, which is I think really about how we might use AI and other areas of the business and AWS, maybe more than I think you asked about our retail business.
Speaker Change: The way I would think about it is that theres kind of two macro buckets of how we see people both ourselves inside of Amazon as well as other companies using AWS, how we see them getting value head of AI today, the first macro bucket I would say is.
Speaker Change: Around productivity and cost savings.
Speaker Change: And in many ways. This is the lowest hanging fruit in AI.
Speaker Change: And you see that all over the place in our retail business for instance, if you look at customer service.
Speaker Change: And you look at the Chatbot that we've built we completely re architected it with generative AI.
Speaker Change: It's delivering it already had pretty high satisfaction is delivering 500 basis points better satisfaction from customers with the new generative AI.
Speaker Change: <unk> Chatbot, if you look at our.
Speaker Change: Third party selling partners.
Speaker Change: One of their biggest pain points is because we put a high premium on really organizing our marketplace. Today, it's easy to find things. There is a bunch of different fuels you have to fill out when you're creating a new product detail page, but we've built the generative AI application for them, where they can either fill in just a couple of lines.
Speaker Change: Taxed or take a picture of an image or point to where your al and the generative AI App will fill in most of the rest of the information they have to fill out which speeds are getting selection on the website and easier for sellers.
Speaker Change: If you look at.
Speaker Change: How we do inventory management and trying to understand what inventory, we need and what facility at what time.
Speaker Change: The tenure of AI applications, we've built there.
Speaker Change: 10% better forecasting on our part and 20% better regional predictions.
Speaker Change: <unk> robotics, we were just talking about the brains and a lot of those robotics, our generative AI infused that do things like tell the <unk>.
Speaker Change: The robotic claw whats in have been what it should pick up how it should move it where it should place it in the in the other Ben that it's feeling like so it's really in the brains of most of our robotics. So we have a number of very significant.
Speaker Change: Call it productivity and cost savings efforts in our retail business, they're using generative AI and again, it's just a fraction of what we have going.
Speaker Change: The other big macro bucket are really altogether, new experiences and again you see lots of those in our retail business ranging from Ruth This which is our AI.
Speaker Change: Infused shopping assistant which continues to grow very significantly.
Speaker Change: Two things like Amazon lens wear.
Speaker Change: You can take a picture of a product that's in front of me check it out in the App you can you can find it in the little box at the top you take a picture of an item in front of you and us.
Speaker Change: It uses computer vision agenda of AI to pull up the exact item.
Speaker Change: In a search result.
Speaker Change: Two things like sizing, where we basically have taken the catalogs of all of these different clothing manufacturers and then compare them against one another one another so we know which brands tend to run bigger small relative to each other so when you come to buy a pair of shoes for instance, it can recommend what sai.
As you need to even what we're doing in Thursday football, where we're using generative AI for really an inventive features like that.
Speaker Change: Defensive alerts, where we predict which players going to puts a quarterback or defensive vulnerabilities, where we were able to show viewers what area of the field is vulnerable. So we're using it really all over.
Speaker Change: Our retail business in all the businesses in which we're in we've got about 1000 different generative AI applications, we've either built or in the process of building right now.
Speaker Change: And the next question comes from the line of John Blackledge with TD Cowen. Please proceed.
John Blackledge: Great. Thanks.
Speaker Change: You talk about the current speed of delivery, maybe how much more room to go there and how is that driving the everyday essentials business and then somewhat relatedly. It any further color on inbound network efficiencies you would expect to see this year as you guys try to continue to lower the cost to Sir Thank you.
Speaker Change: Yes, I would say on speed of delivery that we measure this very carefully and we measure both.
Speaker Change: What the conversion rate is of somebody who views a product detail page with a faster delivery promise.
Speaker Change: Versus those that are slower as well as what we see downstream from customers once they've bought with a SaaS promise and what they end up buying throughout the year.
Speaker Change: And we have not yet seen.
Speaker Change: Diminishing returns and being able to continue to improve the speed of delivery.
Speaker Change: Doesn't mean that there won't be instances in which people are happy to take products later.
Have a program, where if people want what we want to pick a day during the week, where they want to combine a bunch of their shipments and have it delivered then to to.
More sustainable.
Speaker Change: More environmentally friendly they can and we have plenty of customers, who choose that we time in and time out.
Speaker Change: See the people choose to buy from us more frequently when we're able to deliver to their homes or wherever they are much more quickly and it leads to theyre actually using us for more of their everyday purchases.
When we can deliver more quickly I think that.
Speaker Change: If you look at what we're doing with Prime air.
Speaker Change: The promise there is for a number of items that we'll be able to deliver items to customers inside an hour.
Speaker Change: And I think when you're ordering everyday essentials, where you need something more quickly. It's a big deal and you see it had it had a big impact on our everyday essentials, it's had a big impact on our pharmacy business, where people are able to get items same day now in lots of cities throughout the U S and they're just using us much more.
Speaker Change: Frequently than than they had before.
Speaker Change: On the inbound network efficiencies.
Speaker Change: Well I would tell you is.
Speaker Change: We've made a pretty significant architectural change in our inbound network.
Speaker Change: We've been working on for the better part of the year that we rolled out just a few months ago.
Speaker Change: Again, it's what we find when we make big architectural changes like this is that.
Speaker Change: You tend to get some low hanging fruit efficiencies early but then there is all sorts of tuning and refinement you have to do once you actually see it working live across the really vast network.
Speaker Change: There are all sorts of ways here that where I think it's early and I think we're going to get.
Speaker Change: Additional efficiencies throughout the year, but I expect that we will have opportunities to keep taking our cost to serve down this year and that'll be a big part of it.
Speaker Change: And our final question comes from the line of Michael Morton with Moffett Nathan. Please proceed.
Michael Morton: Alright. Thank you so much for the question I wanted to follow up on Andy's remarks about how you are using AI within the Amazon E Commerce experience.
Michael Morton: I wanted to talk about the other side of the coin and Thats really the E Commerce discovery process that leads people to Amazon.
Michael Morton: A lot of companies rolling out agents and assistant and I would love to hear how Amazon is planning for potential disruption in the funnel and what the plans are you spoke about Rufus maybe make that more prominent but what do you think the changes coming to the e-commerce funnel over the next several years would be great.
Michael Morton: Yeah.
Michael Morton: Well I would say that.
Michael Morton: I think retailers ourselves and probably lots of other retailers are all going to have.
Michael Morton: Let's say on how they want to interact with agents I think that it's an emerging space and if you think about the investments that we've made.
Michael Morton: In our fulfillment network and our website.
All the selection that we've built how we organize it.
Michael Morton: Most retailers are going to have.
Michael Morton: Kind of terms in which theyre going to interact with agents and will be no different that way.
Michael Morton: I think that.
Michael Morton: I do think that Rufus.
Michael Morton: Look at what it how it.
Michael Morton: <unk> the customer experience and if you if you actually use it month to month continues to get better and better it's already if you're buying something and youre on our product detail page our product detail pages provides so much information that.
Michael Morton: Sometimes it's hard if you're trying to find something quickly to scroll through and figure and find that little piece of information. So we have so many customers now who just use room for us to help them find a quick fact about our product.
Michael Morton: They also use <unk>.
<unk> is to figure out how to.
Michael Morton: I had to summarize customer reviews, so they'd have to read 100 customer reviews to get a sense of what people think about that product.
Michael Morton: If you look at the personalization really most prominently today your ability to go into Rufus and and.
Michael Morton: And ask what's happened to an order or what did I just order or can you can you pull up for me. This item they were two months ago the personalization.
Michael Morton: <unk> keeps getting much better and so we expect throughout 2025 that the number of occasions, where youre not sure. What you want to buy and you want help from real FIS are going to continue to increase and be more and more helpful to customers.
Michael Morton: Thank you for joining us on the call today and for your questions. A replay will be available on the Investor Relations website for at least three months. We appreciate your interest in Amazon and look forward to talking with you again next quarter.
Speaker Change: And ladies and gentlemen, this does conclude today's teleconference. You may disconnect. Your lines at this time. Thank you for your participation.
Speaker Change:
Speaker Change: [music].
Speaker Change: Okay.
Speaker Change: [music].
Speaker Change: Hmm.
Speaker Change: [music].