Today: Dec 26, 2024

Complex Micro Units (AMD) Q1 2024 Income Name Transcript | The Motley Idiot

Complex Micro Units (AMD) Q1 2024 Income Name Transcript | The Motley Idiot
May 1, 2024



AMD income name for the duration finishing March 31, 2024.
Complex Micro Units (AMD) Q1 2024 Income Name Transcript | The Motley Idiot
Symbol supply: The Motley Idiot.

Complex Micro Units (AMD -1.14%)
Q1 2024 Income Name
Apr 30, 2024, 5:00 p.m. ETContents: Ready Remarks Questions and Solutions Name Members Ready Remarks: OperatorOver greetings and welcome to the AMD first-quarter 2024 convention name. [Operator instructions] As a reminder, this convention is being recorded. It’s now my excitement to introduce your host, Mitch Haws, Vice President, Investor Family members. Thanks, Mitch. You could start.Mitch Haws — Head of Investor Family members Thanks, and welcome to AMD’s first-quarter 2024 monetary effects convention name. By means of now, you will have had the chance to study a duplicate of our income press liberate and the accompanying slides. You probably have now not had the danger to study those fabrics, they are able to be discovered at the Investor Family members web page of AMD.com. We can refer basically to non-GAAP monetary measures throughout these days’s name, and the total non-GAAP to GAAP reconciliations are to be had in these days’s press liberate and the slides posted on our web page. Members on these days’s name are Physician Lisa Su, our chair and leader govt officer; and Jean Hu, our govt vp, leader monetary officer, and treasurer. It is a reside name and can be replayed by way of webcast on our web page. Earlier than we commence, I want to be aware that Mark Papermaster, govt vp and leader generation officer, will attend the TD Cowen Generation Media and Telecom Convention on Might twenty ninth. And Jean Hu, govt vp, leader monetary officer, and treasurer, will attend the J.P. Morgan International Media and Communications Convention on Tuesday, Might twenty first; the Financial institution of The us International Generation Convention on Wednesday, June fifth; and the Jefferies Nasdaq Investor Convention on Tuesday, June eleventh. Nowadays’s dialogue accommodates forward-looking statements in line with present ideals, assumptions, and expectancies, discuss handiest as of these days, and as such, contain dangers and uncertainties that might purpose exact effects to fluctuate materially from our present expectancies. Please consult with the cautionary observation in our press liberate for more info at the elements that might purpose exact effects to fluctuate materially. With that, I can hand the decision over to Lisa. Lisa Su — President and Leader Govt Officer Thanks, Mitch, and excellent afternoon to all the ones listening these days. That is a shockingly thrilling time for the {industry} because the well-liked deployment of AI is riding call for for considerably extra compute throughout a extensive vary of markets. Underneath this backdrop, we’re executing really well as we ramp our information heart industry and permit AI functions throughout our product portfolio. Taking a look on the first quarter, income larger to $5.5 billion. We expanded gross margin through greater than two proportion issues and larger profitability as information heart and shopper section gross sales each and every grew through greater than 80% yr over yr. Knowledge heart section income grew 80% yr over yr and a couple of% sequentially to a file $2.3 billion. The considerable year-over-year expansion used to be pushed through the robust ramp of AMD Intuition MI300X GPU shipments and a double-digit proportion building up in server CPU gross sales. We imagine we received server CPU income percentage within the seasonally down first quarter led through expansion in undertaking adoption and expanded cloud deployments. In cloud, whilst the total call for atmosphere remained combined, hyperscalers persisted adopting fourth-gen EPYC processors to energy extra in their inner workloads and public cases. There are actually just about 900 AMD-powered public cases to be had globally as Amazon, Microsoft and Google all larger their fourth-gen EPYC processor choices with new cases and regional deployments. Within the undertaking, now we have noticed indicators of making improvements to call for as CIOs want to upload extra normal objective and AI compute capability whilst keeping up the bodily footprint and tool wishes in their present infrastructure. This situation aligns completely with the price proposition of our EPYC processors. Given our excessive core depend and effort potency, we will ship an identical quantity of compute with 45% fewer servers in comparison to the contest, chopping preliminary capex through as much as 1/2 and decreasing annual opex through greater than 40%. In consequence, undertaking adoption of EPYC CPUs is accelerating, highlighted through deployments with extensive enterprises, together with American Airways, DBS, Emirates Financial institution, Shell, and SD Micro. We are additionally development momentum with AMD-powered answers, powering the preferred ERP and database packages. As one instance, the most recent technology of Oracle Exadata, the main database answer utilized by 76 of the Fortune 100, is now powered completely through fourth-gen EPYC processors. Taking a look forward, we are very occupied with our next-gen Turin circle of relatives of EPYC processors that includes our Zen 5 core. We are broadly sampling Turin, and the silicon is searching nice. Within the cloud, the numerous functionality and potency will increase of Turin place us neatly to seize a fair higher percentage of each first- and third-party workloads. As well as, there are 30% extra Turin platforms in building from our server companions in comparison to fourth-gen EPYC platforms, expanding our undertaking SAM with new answers optimized for added workloads. Turin stays not off course to release later this yr. Turning to our broader information heart portfolio, we delivered our moment instantly quarter of file information heart GPU income as MI300 become the fastest-ramping product in AMD historical past, passing $1 billion in general gross sales in not up to two quarters. In cloud, MI300X manufacturing deployments expanded at Microsoft, Meta, and Oracle to energy generative AI coaching and inferencing for each inner workloads and a extensive set of public choices. For the undertaking, we are operating very intently with Dell, HP, Lenovo, Tremendous Micro, and others as more than one MI300 platforms input quantity manufacturing this quarter. As well as, now we have greater than 100 undertaking and AI consumers actively creating or deploying MI300X. At the AI instrument entrance, we made superb development including upstream strengthen for AMD {hardware} within the OpenAI Triton compiler, making it even more uncomplicated to expand extremely performant AI instrument for AMD platforms. We additionally launched a significant replace to our ROCm instrument stack that expands strengthen for open-source libraries together with VLLM and frameworks together with JAX, provides new options like video decode and considerably will increase generative AI functionality through integrating complicated consideration algorithms strengthen for sparsity and FP8. Our companions are seeing very robust functionality of their AI workloads. As we collectively optimize for his or her fashions, MI300X GPUs are handing over management, inferencing functionality and considerable TCO benefits in comparison to H100. For example, a number of of our companions are seeing important will increase in tokens in keeping with moment when working their flagship LLMs on MI300X in comparison to H100. We are additionally proceeding to permit the extensive ecosystem required to energy the following technology of AI programs, together with as a founding member of the Extremely Ethernet Consortium operating to optimize the generally followed Ethernet protocol to run AI workloads at information heart scale. MI300 call for continues to fortify. And in line with our increasing buyer engagements, we now be expecting information heart GPU income to exceed $4 billion in 2024, up from the $3.5 billion we guided in January. Long run, we’re an increasing number of operating nearer with our cloud and undertaking consumers as we amplify and boost up our AI {hardware} and instrument roadmaps and develop our information heart GPU footprint. Turning to our shopper section, income used to be $1.4 billion, an building up of 85% yr over yr, pushed through robust call for for our newest technology Ryzen cell and desktop processors with OEMs and within the channel. Consumer section income declined 6% sequentially. We noticed robust call for for our newest technology Ryzen processors within the first quarter. Ryzen desktop CPU gross sales grew through a robust double-digit proportion yr over yr, and Ryzen cell CPU gross sales just about doubled yr over yr as new Ryzen 8040 pocket book designs from Acer, Asus, HP, Lenovo, and others ramped. We expanded our portfolio of management undertaking PC choices with the release of our Ryzen Professional 8000 processors previous this month. Ryzen Professional 8040 cell CPUs ship industry-leading functionality and battery existence for business notebooks, and our Ryzen Professional 8000 sequence desktop CPUs are the primary processor to provide devoted on-chip AI accelerators in business desktop PCs. We see transparent alternatives to achieve further business PC percentage in line with the functionality and potency benefits of our Ryzen Professional portfolio and an expanded set of AMD-powered business PCs from our OEM companions. Taking a look ahead, we imagine the marketplace is not off course to go back to annual expansion in 2024, pushed through the beginning of an undertaking refresh cycle in AI PC adoption. We see AI as the largest inflection level in PCs because the web being able to ship unparalleled productiveness and value good points. We are operating very intently with Microsoft and a extensive ecosystem of companions to permit the following technology of AI stories powered through Ryzen processors with greater than 150 ISVs not off course to be creating for AMD AI PCs through the top of the yr. We can additionally take the following primary step in our AI PC roadmap later this yr with the release of our next-generation Ryzen cell processors codenamed Strix. Buyer passion in Strix may be very excessive in line with the numerous functionality and effort potency uplifts we’re handing over. Design win momentum for top rate notebooks is outpacing prior generations as Strix allows next-generation AI stories in laptops which can be thinner, lighter, and quicker than ever sooner than. We are occupied with the expansion alternatives for the PC marketplace. And in line with the power of our Ryzen CPU portfolio, we think to develop income percentage this yr. Now turning to our gaming section. Earnings declined 48% yr over yr and 33% sequentially to $922 million. First-quarter semi-custom SoC gross sales declined consistent with our projections as we are actually within the 5th yr of the console cycle. In gaming graphics, income declined yr over yr and sequentially. We expanded our Radeon 7000 sequence circle of relatives with the worldwide release of our Radeon RX 7900 GRE and in addition offered our motive force primarily based AMD Fluid Movement Frames generation that may give extensive functionality will increase in hundreds of video games. Turning to our embedded section. Earnings reduced 46% yr over yr and 20% sequentially to $846 million as consumers stay taken with normalizing their stock ranges. We introduced our Spartan UltraScale+ FPGA circle of relatives with excessive IO counts, energy potency, and state of the art security measures, and we are seeing a robust pipeline of expansion for our value optimized embedded portfolio throughout more than one markets. Given the present embedded marketplace stipulations, we are now anticipating second-quarter embedded section income to be flat sequentially with a gentle restoration in the second one 1/2 of the yr. Long run, we see AI on the edge as a big expansion alternative that can pressure larger call for for compute throughout a variety of units. To deal with this call for, we introduced our moment technology of versal adaptive SoCs that ship a thrice building up in AI tops in keeping with watt and a ten occasions better scalar compute functionality in comparison to our prior technology of {industry} main adaptive SoCs. Versal gen two adaptive SoCs are the one answer that mixed more than one compute engines to maintain AI, pre-processing, inferencing, and post-processing on a unmarried chip, enabling consumers to abruptly upload extremely performant and environment friendly AI functions to a extensive vary of goods. We have been happy to be joined at our release through Subaru, who introduced they followed versal AI Edge Collection Gen two units, to energy the following technology in their EyeSight ADAS machine. Embedded design win momentum stays very robust as consumers undertake our complete portfolio of FPGAs, CPUs, GPUs, and adaptive SoCs to deal with a bigger portion in their compute wishes. In abstract, we completed neatly within the first quarter, surroundings us as much as ship robust annual income expansion and expanded gross margin pushed through rising adoption of our Intuition, EPYC and Ryzen product portfolios. Our priorities for 2024 are very transparent: boost up our information heart expansion through ramping Intuition GPU manufacturing and gaining percentage with our EPYC processors, release our next-generation Zen 5 PC and server processors that reach our management functionality, and amplify our adaptive computing portfolio with differentiated answers. Taking a look additional forward, AI represents an unparalleled alternative for AMD. Whilst there was important expansion in AI infrastructure build-outs, we’re nonetheless within the very early levels of what we imagine goes to be a duration of sustained expansion pushed through an insatiable call for for each high-performance AI and normal objective compute. We’ve got expanded our investments around the corporate to seize this massive expansion alternative, from abruptly increasing our AI instrument stack to accelerating our AI {hardware} roadmaps, expanding our go-to-market actions, and partnering intently with the most important AI corporations to co-optimize answers for his or her maximum necessary workloads. We’re very occupied with the trajectory of the industry and the numerous expansion alternatives forward. Now I would like to show the decision over to Jean to supply some further colour on our first-quarter effects. Jean?Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Thanks, Lisa. And excellent afternoon, everybody. I will get started with a assessment of our monetary effects after which supply our present outlook for the second one quarter of fiscal 2024. We delivered a robust year-over-year income expansion in our information heart and shopper segments within the first quarter and drew 230 foundation issues of gross margin enlargement. For the primary quarter of 2024, income used to be $5.5 billion, up 2% yr over yr, as income expansion within the information heart and shopper section used to be in part offset through decrease income in our gaming and embedded segments. Earnings declined 11% sequentially as upper information heart income as a consequence of the ramp of our AMD Intuition GPUs used to be offset through decrease gaming and embedded section revenues. Gross margin used to be 52%, up 230 foundation issues yr over yr, pushed through upper income contribution from the information heart and the customer section, in part offset through decrease embedded and gaming section income contribution. Working bills have been $1.7 billion, an building up of 10% yr over yr as we proceed to take a position aggressively in R&D and advertising actions to deal with the numerous AI expansion alternatives forward folks. Working source of revenue used to be $1.1 billion, representing a 21% running margin. Taxes, passion expense, and different used to be $120 million. For the fourth quarter of 2024, diluted income in keeping with percentage used to be $0.62, an building up of three% yr over yr. Now turning to our reportable segments, beginning with the information heart. Knowledge heart delivered file quarterly section income of $2.3 billion, up 80%, a $1 billion building up yr over yr. Knowledge heart accounted for greater than 40% of general income, basically led through the ramp of AMD Intuition GPUs from each cloud and undertaking consumers and a powerful double-digit proportion expansion in our server processing income on account of expansion throughout our Zen 4 merchandise. Our sequential foundation income larger 2% pushed through the ramp-up of our AMD Intuition GPUs, in part offset through seasonal decline in server CPU gross sales. Knowledge heart section running source of revenue used to be $541 million or 23% of income, in comparison to $148 million or 11% a yr in the past. Working source of revenue used to be up 266% yr over yr because of running leverage whilst we considerably larger our funding in R&D. Consumer section income used to be $1.4 billion, up 85% yr over yr, pushed basically through Ryzen 8000 sequence processors. On a sequential foundation, shopper income declined 6%. Consumer section running source of revenue used to be $86 million or 6% of income in comparison to an running lack of $172 million a yr in the past, pushed through upper income. Gaming section income used to be $922 million, down 48% yr over yr and down 33% sequentially because of a lower in [semi] buyer and Radeon GPU gross sales. Gaming section running source of revenue used to be $151 million or 16% of income in comparison to $314 million or 18% a yr in the past. Embedded section income used to be $846 million, down 46% yr over yr and 20% sequentially as consumers proceed to control their stock ranges. Embedded section running source of revenue used to be $342 million or 41% of income in comparison to $798 million or 51% a yr in the past. Turning to the stability sheet and the money float. All over the quarter, we generated $521 million in money from operations, and our loose money float used to be $379 million. Stock will increase sequentially through $301 million to $4.7 billion, basically to strengthen the ongoing ramp of knowledge heart and shopper merchandise in complicated procedure nodes. On the finish of the quarter, money, money equivalents, and non permanent funding used to be $6 billion. As a reminder, now we have $750 million of debt maturing this June. Given our plentiful liquidity, we plan to retire the debt using present money. Now turning to our moment quarter 2024 outlook. We predict income to be roughly $5.7 billion plus or minus $300 million. Sequentially, we think our information heart section income to extend through double-digit proportion, basically pushed through the information heart GPU ramp; shopper section income to extend; embedded section income to be flat; and within the gaming section in line with present call for indicators, income to say no through important double-digit proportion. 12 months over yr, we think our information heart and shopper section income to be up considerably pushed through the power of our product portfolio, the embedded and gaming section income to say no through a vital double-digit proportion. As well as, we think second-quarter non-GAAP gross margin to be roughly 53%, non-GAAP running bills to be roughly $1.8 billion, non-GAAP efficient tax fee to be 13%, and the diluted percentage depend is anticipated to be roughly 1.64 billion stocks. In last, we began the yr robust. We made important development on our strategic priorities, delivered year-over-year income expansion in our information heart and shopper section, and expanded the gross margin. Taking a look forward, we imagine the funding we’re making will place us really well to deal with the massive AI alternatives forward. With that, I will flip it again to Mitch for the Q&A consultation.Mitch Haws — Head of Investor Family members Thanks, Jean. Paul, we are satisfied to ballot the target market for questions. Questions & Solutions:Operator[Operator instructions] Our first query is from Toshiya Hari with Goldman Sachs. Please continue together with your query.Toshiya Hari — Goldman Sachs — Analyst Hello, Thanks such a lot for taking the query. Lisa, my first query is at the MI300. You take up the full-year outlook from $3.5 billion to $4 billion. I am curious, what is riding that incremental $500 million in income? Is it new consumers? Is it further bookings from present consumers? Is it extra cloud? Is it extra undertaking? If you’ll be able to form of supply colour there that will be useful. After which at the provide facet, there may be been headlines or chatter that CoWos underneath HBM is usually a beautiful critical constraining issue for you guys. If you’ll be able to discuss to how you might be dealing with the provision facet of the equation, that will be useful, too. After which I’ve a snappy follow-up.Lisa Su — President and Leader Govt Officer Nice. Thanks, Toshiya for the query. Glance, the MI300 ramp goes in point of fact neatly. If we take a look at simply what is came about during the last 90 days, we have been operating very intently with our consumers to qualify MI300 of their manufacturing information facilities, each from a {hardware} viewpoint, instrument viewpoint. Thus far, issues are going relatively neatly, and what we see now’s simply better visibility to each present consumers in addition to new consumers committing to MI300. In order that provides us the arrogance to move from $3.5 billion to $4 billion. And I view this as you already know very a lot — as an excessively dynamic marketplace, and there are many consumers. We mentioned at the — within the ready remarks that we have got over 100 consumers that have been engaged with in each building in addition to deployment. So general the ramp goes in point of fact neatly. Because it pertains to the provision chain, if truth be told I might say I am very happy with how provide has ramped. It is the absolute quickest product ramp that we have got carried out. It is a very complicated product. Chiplets, CoWos, 3-D integration, HBM. And up to now it is long gone extraordinarily neatly. We have now gotten nice strengthen from our companions. And I might say even within the quarter that we simply completed, we if truth be told did a little bit bit higher than anticipated once we first began the quarter. I believe Q2 can be every other important ramp, and we are going to ramp provide each quarter this yr. So I believe the provision chain goes neatly. We’re tight on provide. So there is no query within the close to time period that if we had extra provide now we have call for uh for that product. And we are going to proceed to paintings on the ones parts as we move throughout the yr. However I believe each at the call for facet and the provision facet, I am very happy with how the ramp goes.Toshiya Hari — Goldman Sachs — Analyst Thanks for all of the main points. After which as my follow-up, I used to be hoping you have to discuss in your information heart GPU roadmap past the MI300. The opposite worry that we listen is your nearest competitor has been beautiful clear with their roadmap, and that extends into ’25 and oftentimes ’26. So possibly this is not the precise venue so that you can give an excessive amount of. However past the MI300, how will have to we take into consideration your roadmap and your skill to compete in information heart? Thanks.Lisa Su — President and Leader Govt Officer Yeah, certain. So glance, Toshiya, once we get started with the roadmap, I imply we all the time take into consideration it as a multi-year, multi-generational roadmap. So now we have the follow-ons to MI300 in addition to the following subsequent generations neatly in building. I believe what is right is we are getting a lot nearer to our best AI consumers. They are if truth be told giving us important comments at the roadmap and what we want to meet their wishes. Our chiplet structure is if truth be told very versatile, and in order that permits us to if truth be told make adjustments to the roadmap as essential. So we are very assured in our skill to proceed to be very aggressive. Frankly, I believe we are going to get extra aggressive. At the moment, I believe MI300X is in a candy spot for inference, very, very robust inference functionality. I see as we carry in more merchandise later this yr into 2025, that can proceed to be a robust spot for us, after which we are additionally improving our coaching functionality and our instrument roadmap to move at the side of it. So extra main points to come back within the coming months. However now we have a robust roadmap that is going thru the following couple of years, and it’s knowledgeable through simply a large number of finding out in operating with our best consumers.Toshiya Hari — Goldman Sachs — Analyst Admire it. Thanks.Lisa Su — President and Leader Govt Officer Certain.OperatorOur subsequent query is from Ross Seymore with Deutsche Financial institution. Please continue together with your query.Ross Seymore — Deutsche Financial institution — Analyst Hello, Thank you for having me ask a query. The non AI facet of the information heart industry, it sounds just like the undertaking facet has some excellent traction although the sequential drop came about seasonally, Lisa. However I used to be simply questioning, what is implied for your second-quarter steering for the information heart CPU facet of items? And usually talking, how are you seeing that entire more or less GPU as opposed to CPU crowding out dynamic taking part in out for the remainder of 2024?Lisa Su — President and Leader Govt Officer Yeah, certain. Ross, thank you for the query. I believe the — our EPYC industry has if truth be told carried out beautiful neatly. The marketplace is a little combined. I believe one of the most cloud guys are nonetheless operating thru form of their optimizations. I believe it is other through buyer. We did see right here within the first quarter if truth be told some really nice early indicators within the undertaking area, form of extensive consumers beginning refresh systems. The price proposition of Genoa may be very, very robust, and we are seeing that pull thru around the undertaking. In the second one quarter, we think general information heart to be up robust double digits. After which inside of that, we think server to be up as neatly. And as we move into the second one 1/2 of the yr, I believe there are a few drivers for us. We do be expecting some development within the general marketplace stipulations for the server industry, however we even have our Turin release in the second one 1/2 of the yr. That may also, we imagine, lengthen our management place inside the server marketplace. So general, I believe the industry is appearing neatly, and we imagine that we are proceeding to be really well located to achieve percentage during the yr.Ross Seymore — Deutsche Financial institution — Analyst Thank you for that. And I suppose as my follow-up, simply switching over to the customer facet, I famous you guided it up sequentially. Any form of magnitudes round that for the second one quarter? And possibly extra importantly, while you discuss the entire AI PC facet of items, do you imagine that is extra of a unit motive force for you, an ASP motive force? Or will it’s each?Lisa Su — President and Leader Govt Officer Yeah, so I believe, once more I believe we are beautiful occupied with the AI PC, each alternative in, let’s name it, the close to time period and much more so within the medium time period. I believe the customer industry is appearing neatly, each at the channel and at the M&C facet. We predict shopper to be up sequentially in the second one quarter. And as we move into the second one 1/2 of the yr, in your query about gadgets as opposed to ASPs, I believe we think some building up in gadgets in addition to ASPs. The AI PC merchandise, once we take a look at the Strix merchandise, it is — they are in point of fact neatly suited to the top rate segments of the marketplace. And I believe that is the place you’ll see one of the most AI PC content material most powerful at first. After which as we move into 2025, you can see it extra throughout the remainder of the portfolio.Ross Seymore — Deutsche Financial institution — Analyst Thanks.Lisa Su — President and Leader Govt Officer Thanks. Thank you, Ross.OperatorThank you. Our subsequent query is from Matt Ramsay with TD Cowen. Please continue together with your query.Matt Ramsay — TD Cowen — Analyst Sure. Thanks very a lot. Excellent afternoon, everyone. Lisa, I’ve form of a longer-term query after which a shorter-term follow-up. One, I suppose is the — one of the vital questions that I have been getting from other people so much is, clearly your number one competitor has introduced, I suppose, a multiyear roadmap. And we proceed to listen to increasingly more from folks about inner ASIC systems at a few of your number one consumers, whether or not they be for inference, or coaching, or each. I suppose it could be in point of fact useful if you have to to speak to us about how your conversations move with the ones consumers, how dedicated they’re in your long-term roadmap, multi-generation, as you described it, how they juxtapose doing investments in their inner silicon as opposed to the usage of a service provider provider like yourselves and possibly what benefits the revel in throughout a big footprint of shoppers may give your corporate that the ones guys doing inner ASICs would possibly now not get. Thank you.Lisa Su — President and Leader Govt Officer Yeah, certain, Matt. Thank you for the query. So glance, I believe one of the vital issues that we see and we have mentioned is that the TAM for AI compute is rising extraordinarily briefly, and we see that proceeding to be the case in all conversations. We had highlighted a TAM of, let’s name it, $400 billion in 2027. I believe some other people concept that used to be competitive on the time. However the general AI compute wishes, as we communicate to consumers, may be very, very robust. And you’ve got noticed that during one of the most bulletins, even not too long ago with one of the most greatest cloud guys. From my view, there are a number of facets of it. To start with, now we have nice relationships with all of form of the highest AI corporations. And the speculation there’s we need to innovate in combination. Whilst you take a look at those extensive language fashions and the entirety that you want for coaching and inferencing there, even though there can be many answers, I don’t believe there may be only one answer that can are compatible all. The GPU continues to be the most well liked structure, particularly because the algorithms and the fashions are proceeding to adapt through the years, and that favors our structure and in addition our skill to in point of fact optimize CPU with GPU. So from my viewpoint, I believe we are more than happy with the partnerships that we have got. I believe it is a large alternative for all folks to in point of fact innovate in combination, and we see that there is a very robust dedication to operating in combination over more than one years going ahead. And that is the reason, I believe, a testomony to one of the most paintings that we have carried out prior to now, and that very a lot is what came about with the epic roadmap as neatly.Matt Ramsay — TD Cowen — Analyst Admire that, Lisa. As my follow-up, a little bit bit shorter time period. And I suppose having adopted the corporate tremendous intently for a very long time, I believe there may be been — there may be all the time been noise within the machine from whether or not the inventory worth is $2 a percentage or $200. There is been more or less all the time constant noise come what may. However the remaining 1.5 months has been excessive in that sense. And so I sought after to simply — I have were given random studies in my inbox about adjustments in call for from a few of your MI300 consumers or deliberate call for for eating your product. I believe you spoke back previous concerning the provide state of affairs and the way you might be operating together with your companions there. However has there been any exchange from the shoppers that you are in ramp with now or that you simply quickly can be of what their purpose is for call for? Or in truth has that possibly bolstered moderately than long gone down in contemporary classes? As a result of I stay getting questions on it. Thank you.Lisa Su — President and Leader Govt Officer Certain, Matt. Glance, I believe I would possibly have mentioned it previous, however possibly I will repeat it once more. I believe the call for facet is if truth be told in point of fact robust. And what we see with our consumers and what we’re monitoring very intently is consumers shifting from, let’s name, it preliminary POCs to pilots, to complete scale manufacturing, to deployment throughout more than one workloads. And we are shifting thru that series really well. I think excellent concerning the deployments and ramps that we have got ongoing presently, and I additionally really feel excellent about new consumers who’re form of previous on in that procedure. So from a requirement viewpoint, we proceed to construct backlog in addition to construct engagements going ahead. And in a similar way at the provide viewpoint, we are proceeding to construct provide momentum. However from a speed-of-ramp viewpoint, I am if truth be told in point of fact proud of the development.Matt Ramsay — TD Cowen — Analyst All proper. Thanks very a lot.OperatorThank you. Our subsequent query is from Aaron Rakers with Wells Fargo. Please continue together with your query.Aaron Rakers — Wells Fargo Securities — Analyst Yeah. Thank you for taking the query, and I express regret if I overlooked this previous. However I do know remaining quarter, you mentioned having a — securing sufficient capability to strengthen important upside to the ramp of the MI300. I do know that you simply upped your information now to $4 billion. I am curious to how you can represent the provision relative to that context presented remaining quarter as we take into consideration that new more or less goal of $4 billion. Would you represent it as nonetheless having provide capability upside possible? Thanks.Lisa Su — President and Leader Govt Officer Sure, Aaron. So we have mentioned sooner than that our objective is to be sure that now we have provide that exceeds the present steering, and that’s true. In order we have upped our steering from $3.5 billion to $4 billion, we nonetheless — now we have provide visibility considerably past that.Aaron Rakers — Wells Fargo Securities — Analyst Yeah. OK. Thanks. After which as a snappy follow-up, going again to the sooner query on server call for, extra conventional server. As you spot the ramp of possibly percentage alternatives in additional conventional undertaking, I am curious to how you can represent the expansion that you are expecting to look in additional conventional server CPU marketplace as we transfer thru ’24 and even long run, the way you represent that expansion development.Lisa Su — President and Leader Govt Officer Yeah, I believe, Aaron, what I might say is there are — the will for refresh of, let’s name it, older apparatus is without a doubt there, so we see a refresh cycle coming. We additionally see AI head nodes, as every other position the place we see expansion in, let’s name it, the extra conventional marketplace. Our candy spot is in point of fact within the absolute best functionality, form of excessive core depend, power potency area, and that’s taking part in out neatly. And we are additionally — we have historically been very robust in, let’s name it, cloud first-party workloads. And that’s now extending to cloud third-party workloads the place we see enterprises who’re, let’s name it, in additional of a hybrid atmosphere adopting AMD each within the cloud and on-prem. So I believe general, we see it as a persisted excellent development for us with the server industry going thru 2024 and past.Aaron Rakers — Wells Fargo Securities — Analyst Yeah. Thanks.Lisa Su — President and Leader Govt Officer Thank you.OperatorThank you. Our subsequent query is from Vivek Arya with Financial institution of The us Securities. Please continue together with your query.Vivek Arya — Financial institution of The us Merrill Lynch — Analyst Thank you for taking my query. Lisa, I simply sought after to return to the provision query and the $4 billion outlook for this yr. I believe someday, there used to be a proposal that the $4 billion quantity, proper, that there are nonetheless provide constraints. However I believe at a unique level, you mentioned that you’ve provide visibility considerably past that. For the reason that we’re virtually on the center of the yr, I might have concept that you’d have significantly better visibility concerning the again 1/2. So is the $4 billion quantity a provide constraint quantity? Or is it a requirement constraint quantity? Or however, if you have to give us some sense of what the go out fee of your GPU gross sales may well be. I believe at the remaining name, $1.5 billion used to be instructed. May it’s much more than that relating to your go out fee of MI for this yr?Lisa Su — President and Leader Govt Officer Yeah, Vivek. Let me attempt to ensure that we resolution this query obviously. From a full-year viewpoint, our $4 billion quantity isn’t provide capped. I am sorry, sure, it isn’t provide capped. It’s — we do have provide capacity above that. It’s extra again half-weighted. So if you are searching at form of the close to time period, the close to time period, I might say as an example in the second one quarter, we do have extra call for than now we have provide presently, and we are proceeding to paintings on pulling in a few of that offer. By means of the best way, I believe that is an general {industry} factor. This isn’t in any respect associated with AMD. I believe general, AI call for has exceeded somebody’s expectancies in 2024. So you’ve got heard it from the reminiscence guys. You will have heard it from the foundry guys. We are all ramping capability as we move throughout the yr. And because it pertains to visibility, we do have excellent visibility into what is taking place. As I mentioned, now we have nice buyer engagements which can be going ahead. My objective is to ensure that we cross all the milestones as we are ramping merchandise. And as we cross the ones milestones, we put that into the um, the total full-year steering for AI. However relating to how buyer development issues are going, they are if truth be told going relatively neatly. And we proceed to carry new consumers on, and we proceed to amplify workloads with our present consumers. And so confidently, that clarifies the query, Vivek.Vivek Arya — Financial institution of The us Merrill Lynch — Analyst OK. Thanks, Lisa. Possibly one nod on MI, however possibly at the embedded industry. I believe you sound a little extra measured about Q2 and the second one 1/2 rebound, which has similarities to what now we have heard from a large number of the car commercial friends. However the place are you within the stock clearing cycle? And if embedded has a relatively extra measured rebound within the again 1/2, what implication does that experience on gross margin enlargement? Are we able to proceed to be expecting, I have no idea, 100 foundation issues 1 / 4 relating to gross margin enlargement on account of the information heart combine? Or simply any places and takes of embedded, after which what it method for gross margins within the again 1/2. Thanks.Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Hello, Vivek. Thanks for the query. I believe the embedded industry declined a little bit bit greater than anticipated in point of fact because of the weaker call for in one of the most markets. Very in particular, conversation has been vulnerable. And a few wallet of business and car, as you discussed, it is if truth be told relatively in line with the friends. 2d 1/2, we do assume the primary 1/2 is the ground of embedded industry, and we’re going to begin to see sluggish restoration in the second one 1/2. And going again in your gross margin query is while you take a look at our gross margin enlargement in each Q1 and guided Q2, the main motive force is the robust functionality at the information heart facet. The knowledge heart will proceed to ramp in moment 1/2. I believe that can proceed to be the most important motive force of gross margin enlargement in moment 1/2. In fact, if embedded is doing higher, we’re going to have a extra tailwind in moment 1/2.Vivek Arya — Financial institution of The us Merrill Lynch — Analyst Thanks.OperatorThank you. Our subsequent query is from Timothy Arcuri with UBS. Please continue together with your query.Timothy Arcuri — UBS — Analyst Thank you very a lot. I additionally sought after to invite about your information heart GPU roadmap. The shoppers that we communicate to mention that they are engaged, now not simply on account of MI300 however in point of fact on account of what is coming, and it sort of feels like there is a giant call for shift to rack scale programs that attempt to optimize functionality in keeping with sq. foot given one of the most information heart and tool constraints. So are you able to simply discuss how necessary programs are going to be for your roadmap And do you’ve gotten all of the items you want because the marketplace shifts to rack scale programs?Lisa Su — President and Leader Govt Officer Yeah, certain. Timothy, thank you for the query. Needless to say, glance, our consumers are engaged within the multi-generational conversations. So we have — unquestionably going out over the following couple of years. And because it pertains to the total machine integration, it’s relatively necessary. It’s one thing that we are operating very intently with our consumers and companions on. That is a vital funding in networking, operating with quite a lot of networking companions as neatly to ensure that the scale-out capacity is there. And in your query of do now we have the items, we do completely have the items. I believe the paintings that we have all the time carried out with our Infinity Material in addition to with our Pensando acquisition, that is introduced in a large number of networking experience. After which we are operating around the networking ecosystem with key companions like Broadcom, and Cisco, and Arista who’re with us at our AI information heart tournament in December. So our paintings presently in long term generations isn’t just specifying a GPU. It’s specifying, let’s name it, complete machine reference designs, and that is the reason one thing that can be relatively necessary going ahead.Timothy Arcuri — UBS — Analyst Thank you so much. After which simply as a snappy stick to up, I do know you already know this yr seems like it’ll be beautiful again 1/2 loaded for your server CPU industry simply adore it used to be remaining yr. And I do know you more or less held our fingers at about this time remaining yr, form of on what the total yr may seem like and the way again finish loaded it may well be. So I more or less marvel may you give us some milestones relating to how a lot server CPU may develop this yr, how again finish loaded. It may well be is it’s like up 30% this yr on your server CPU industry yr over yr. Is {that a} cheap bogey? I simply wonder whether you’ll be able to more or less give us any steering on that piece of the industry. Thank you.Lisa Su — President and Leader Govt Officer Yeah, I imply, I believe Tim, I believe the easiest way to mention it’s our information heart section, is on an excessively, very robust ramp as we move throughout the again 1/2 of the yr. Server CPUs, without a doubt. Knowledge heart GPUs, needless to say. So I have no idea that we are going to get into specifics. However I may say, on the whole, you will have to be expecting general on the section degree to be very robust double digits.Timothy Arcuri — UBS — Analyst Thanks, Lisa.Lisa Su — President and Leader Govt Officer Thanks.OperatorThank you. Our subsequent query is from Joe Moore with Morgan Stanley. Please continue together with your query.Joe Moore — Morgan Stanley — Analyst Nice. Thanks. I wonder whether you have to cope with the profitability of MI300. I do know you mentioned a few quarters in the past that it could in the end be above company reasonable, however it could take you a couple of quarters to get there. Are you able to discuss the place you’re in that?Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Yeah. Thanks, Joe. Our crew has carried out a fantastic process to ramp MI300. As you most likely know, it is a very complicated product, and we’re nonetheless on the fourth yr of the ramp each from the checking out time and the method development. The ones issues are nonetheless ongoing. We do assume through the years, the gross margin will have to be accretive to company reasonable.Joe Moore — Morgan Stanley — Analyst Nice, thanks. After which a separate follow-up at the Turin transition on server. I do know while you had transitioned into Genoa, you mentioned it would take a short time that there have been important platform shifts and such things as that. Turin appears to be a lot more more or less ecosystem-compatible. How briefly do you assume you may see that product ramp inside of your server portfolio?Lisa Su — President and Leader Govt Officer Yeah, Joe, I believe from what we see, glance, I believe Turin is similar platform, in order that does make it an more uncomplicated ramp. I do assume that Genoa and Turin will coexist for some period of time as a result of consumers are deciding when they will carry out their new platforms. We predict Turin to provide us get admission to to a broader set of workloads. So our SAM if truth be told expands with Turin each in undertaking and cloud. And from our from our revel in, I believe you can see a quicker transition than, as an example, once we went from Milan to Genoa.Joe Moore — Morgan Stanley — Analyst Nice. Thanks.OperatorThank you. Our subsequent query is from Stacy Rasgon with Bernstein Analysis. Please continue together with your query.Stacy Rasgon — AllianceBernstein — Analyst Hello, guys. Thank you for taking my questions. For my first one, I sought after to deal with the MI300 ramp into Q2. So that you mentioned you’ve got carried out $1 billion give or soak up cumulative gross sales, which places it at possibly, I have no idea, possibly $600 million in Q1. You might be guiding general revenues up about $225 million into Q2. However you have got shopper up. You have got conventional information heart up. You have got embedded flat. Gaming goes to be down, however I would danger a bet that the customer and standard information heart like offset it, if now not extra. Does the MI300 ramp into Q2, is it roughly than the full company ramp that you have got constructed into steering presently? Is it that you are anticipating?Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Hello, Stacy. Thank you for the query. You all the time ask a math query. So I believe on the whole, it’s extra, the information heart GPU ramp. It’ll be greater than the total corporate’s $207 million ramp.Stacy Rasgon — AllianceBernstein — Analyst OK. In order that method gaming should be down like so much, proper, if shopper and standard information are all up. OK.Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Yeah, you might be proper. Gaming is down equivalent zip code like a Q1.Stacy Rasgon — AllianceBernstein — Analyst Were given it. Were given it. And that is the reason useful.Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer So possibly — yeah, possibly let me come up with some colour concerning the gaming industry, proper. It is — in the event you take a look at the gaming, the call for has been relatively vulnerable. That is relatively very widely recognized, and in addition their stock degree. So in line with the visibility now we have, the primary 1/2, each Q1 and Q2, we guided down sequentially greater than 30%. We if truth be told assume the second one 1/2 can be not up to first 1/2. That is principally how we are searching at this yr for the gaming industry. And on the identical time, gaming’s gross margin is not up to our corporate reasonable. So general, it is going to lend a hand the combo at the gross margin facet. That is just a few colour at the gaming facet. However you might be proper, Q2 gaming is down so much.Stacy Rasgon — AllianceBernstein — Analyst Were given it. That is useful. Thanks. For my moment query, I sought after to have a look at the near-term information heart profitability. So running benefit used to be down 19% sequentially on 2% income expansion. Is that simply the margins of the GPUs filtering in relative to the CPUs? And I do know you mentioned GPUs would in the end be above company reasonable? Are they under the CPU reasonable in — I imply they obviously are, I suppose, within the close to time period. However are they going to stick that means?Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Yeah. I believe you might be proper, is the GPU gross margin presently is under the information heart gross margin degree. I believe there are two causes. In reality, the most important reason why is that we if truth be told building up the funding relatively considerably to — as Lisa discussed, to amplify and accelerating our roadmap within the AI facet. That is one of the vital primary drivers of the running source of revenue coming down relatively. At the gross margin facet, going again in your query is, we mentioned prior to now and we proceed to imagine the case is information heart GPU gross margin through the years can be accretive to company reasonable, however it is going to take a little time to get to the silver degree of gross margin.Stacy Rasgon — AllianceBernstein — Analyst Were given it. That is useful. Thanks.OperatorThank you. Our subsequent query is from Harlan Sur with J.P. Morgan. Please continue together with your query.Harlan Sur — JPMorgan Chase and Corporate — Analyst Excellent, excellent afternoon. Thank you for taking my query. To your information heart GPU section and the quicker time to manufacturing shipments, given you simply upped your full-year GPU outlook, how a lot of it’s quicker carry up of your consumers frameworks pushed through your newest ROCm instrument platform and possibly more potent collaboration together with your consumers engineers simply to get them to name quicker? And what kind of of it is only a extra competitive build-out plan through consumers as opposed to their prior expectancies given what seems to be beautiful robust urgency for them to transport ahead with their necessary AI tasks?Lisa Su — President and Leader Govt Officer Yeah, Harlan, thanks for the query. What it in point of fact is is each us and our consumers feeling assured in broadening the ramp. As a result of in the event you take into consideration it, initially, the ROCm stack has carried out in point of fact neatly. And the paintings that we are doing is hand in hand with our consumers to optimize their key fashions. And it used to be necessary to get form of verification and validation that the entirety would run neatly, and we have now handed some necessary milestones in that space. After which I believe the opposite factor is, as you mentioned, there is a large call for for extra AI compute and so our skill to take part in that, and lend a hand consumers get that up and working is excellent. So I believe general as we take a look at it, this ramp has been very, very competitive in the event you take into consideration the place we have been only a quarter in the past. Every of those are beautiful complicated bring-ups, and I am more than happy with how they have long gone. And through the best way, we are handiest sitting right here in April, so there may be nonetheless a large number of 2024 to move, and there is a nice buyer momentum within the procedure.Harlan Sur — JPMorgan Chase and Corporate — Analyst Yeah, completely. Simply going again, simply more or less rewinding again to the March quarter. So very similar to the PC shopper industry, proper, which declined on the low finish of the seasonal vary, if I ensure assumptions round your information heart GPU industry, I be expecting that out of knowledge heart, it seems like your server CPU industry used to be additionally down on the decrease finish of the seasonal vary. By means of my math, it used to be down like 5%, 6% sequentially. Is that proper? And that is the reason not up to 1/2 the decline of your competitor. And if this is the case, like what drove the not up to seasonal declines? I suppose a few of it used to be percentage good points, feels like undertaking used to be additionally higher. Seems like you guys did pressure a little bit bit extra cloud example adoption. However the rest that drove the relatively higher seasonal trend in March for information heart server?Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Yeah, yeah. Harlan, that is Jean. I believe the server industry has been appearing in point of fact neatly yr over yr. It if truth be told larger very robust double digit. I believe sequentially, it’s extra seasonal, however we really feel beautiful excellent about proceed gaining percentage there.Harlan Sur — JPMorgan Chase and Corporate — Analyst Thanks.Lisa Su — President and Leader Govt Officer And if I simply upload, Harlan, in your query, we did see power in undertaking within the first quarter, and I believe that has offset possibly one of the most customary seasonality.Mitch Haws — Head of Investor Family members Yeah. Neatly, now we have time for 2 extra questions.OperatorThank you. Our subsequent query is from Tom O’Malley with Barclays. Please continue together with your query.Tom O’Malley — Barclays — Analyst Good day, thank you for taking my query. I simply sought after to invite at the aggressive atmosphere. Clearly at the CPU facet, you had a competitor discuss launching a excessive core depend product within the coming quarter, more or less ramping now and extra so into Q3. You will have noticed in point of fact excellent pricing tailwinds as a serve as of the upper core capital. Are you able to discuss what you might be seeing in that marketplace? Do you assume that there is any possibility for extra competitive pricing, which might affect your ASP ramp for the remainder of the yr?Lisa Su — President and Leader Govt Officer Yeah. Once we take a look at our server CPU, form of ASPs, they are if truth be told very strong. I believe we — once more, we have a tendency to be listed towards the upper core counts. Total, I might say the pricing atmosphere is strong. That is about form of TCO for form of the buyer environments. And form of our functionality and our functionality in keeping with watt, our management, and that generally interprets into TCO merit for our consumers.Tom O’Malley — Barclays — Analyst Useful. After which only a broader query to stick to up right here. So I believe it were given requested previous concerning the significance of programs. However for your finish, how necessary is the Open Ethernet consortium to you having the ability to transfer extra into programs? I do know that these days you clearly have some inner property after which you’ll be able to spouse with others. However is there some way that you’ll be able to be aggressive sooner than there’s an {industry} same old at the Ethernet facet? And are you able to discuss while you assume the timing of that more or less consortium involves marketplace and allows you to possibly boost up that roadmap? Thank you so much.Lisa Su — President and Leader Govt Officer Yeah, I believe it is crucial to mention we’re very supportive of the open ecosystem. We are very supportive of the extremely Ethernet consortium. However I do not imagine that that could be a limiter to our skill to construct large-scale programs. I believe Ethernet is one thing that many within the {industry} really feel would be the long-term resolution for networking in those programs. And now we have a large number of paintings that we are doing with internally in addition to with our consumers and companions to permit that.Mitch Haws — Head of Investor Family members Paul, we are in a position for our ultimate query.OperatorThank you. Our remaining query is from Harsh Kumar with Piper Sandler. Please continue together with your questions.Harsh Kumar — Piper Sandler — Analyst Good day, thanks for letting me ask a query. Lisa, I had two. One is for you and one possibly for Jean. So we not too long ago hosted an excessively extensive tradition GPU corporate for a decision, and so they mentioned more or less mega information facilities bobbing up within the on the subject of mid-term, speaking about nodes probably within the 100,000-plus vary and possibly as much as one million. In order we glance out at most of these information facilities, from an architectural viewpoint, is {that a} state of affairs the place winner takes all the place if any individual will get in, they more or less get all of the sockets? Or will they be strains the place your chip possibly or your board may also be positioned proper subsequent to any individual else’s board possibly on a separate line? Simply lend a hand us know how one thing like that will play out if there is a probability for a couple of competitor to play in this sort of extensive information heart.Lisa Su — President and Leader Govt Officer Yeah, so I will communicate possibly a little bit bit extra on the strategic degree. I believe as we take a look at form of how AI shapes up over the following couple of years, there are consumers who could be searching at very extensive coaching environments. And possibly, that is what you might be speaking about. I believe our view of this is primary, we view that as an excessively sexy space for AMD. It is a space the place, we imagine now we have the generation to be very aggressive there. And I believe the will could be to have optionality relating to the way you construct the ones out. So clearly so much has to occur between right here and there. However I believe your overarching query of is it winner takes all, I don’t believe. In order that being the case, we imagine that AMD may be very neatly located to play in the ones, let’s name it, very large-scale programs.Harsh Kumar — Piper Sandler — Analyst Glorious. Thanks. After which possibly a snappy one for Jean. So Jean, I put the entirety into the fashion that you simply mentioned for June. I am getting a few roughly a $400 million upward thrust within the June quarter over March. You discussed that each MI300 and EPYC will develop. Curious if you have to lend a hand us take into consideration the relative sizing of the ones two segments inside the quarter. I am getting — the purpose I am seeking to make is I am getting more or less a few $900 million quantity for MI300 for June. Is that — am I within the ballpark on my means off right here?Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Harsh, we aren’t going to steer a particular section under the section income. I believe an important factor is we did say information heart goes to develop double digit sequentially. I will go away it over there. We — subsegment, there are a large number of main points.Harsh Kumar — Piper Sandler — Analyst Honest sufficient, I had to check out. Thanks, guys. Thanks such a lot.Jean Hu — Govt Vice President, Leader Monetary Officer and Treasurer Yeah, yeah, yeah. Thanks.OperatorThank you. There are not any additional questions at the moment. I would like handy the ground again over to control for any last feedback.Mitch Haws — Head of Investor Family members Nice. That concludes these days’s name. Because of all of you for becoming a member of us these days.Operator[Operator signoff] Period: 0 minutesCall members:Mitch Haws — Head of Investor RelationsLisa Su — President and Leader Govt OfficerJean Hu — Govt Vice President, Leader Monetary Officer and TreasurerToshiya Hari — Goldman Sachs — AnalystRoss Seymore — Deutsche Financial institution — AnalystMatt Ramsay — TD Cowen — AnalystAaron Rakers — Wells Fargo Securities — AnalystVivek Arya — Financial institution of The us Merrill Lynch — AnalystTimothy Arcuri — UBS — AnalystJoe Moore — Morgan Stanley — AnalystStacy Rasgon — AllianceBernstein — AnalystHarlan Sur — JPMorgan Chase and Corporate — AnalystTom O’Malley — Barclays — AnalystHarsh Kumar — Piper Sandler — Analyst
Extra AMD research
All income name transcripts

OpenAI
Author: OpenAI

Don't Miss

This new AMD Ryzen CPU reputedly beats the Nvidia RTX 4060 and not using a graphics card

This new AMD Ryzen CPU reputedly beats the Nvidia RTX 4060 and not using a graphics card

The brand new AMD Ryzen CPU with built-in tough Radeon GPU turns
AMD’s subsequent GPU already has two giant issues

AMD’s subsequent GPU already has two giant issues

Desk of Contents Desk of Contents Symbol of what is to return