$AMD Q1 2024 AI-Generated Earnings Call Transcript Summary
The operator welcomes participants to the AMD First Quarter 2024 Conference Call and introduces the host, Mitch Haws. Non-GAAP financial measures will be primarily referenced and can be found on the Investor Relations page of amd.com. Participants on the call include Dr. Lisa Su, Chair and CEO, and Jean Hu, CFO and Treasurer. The call will be replayed via webcast and upcoming events where AMD executives will attend are mentioned. The discussion contains forward-looking statements and involves risks and uncertainties.
In the first quarter, AMD's revenue increased to $5.5 billion, with a 2% sequential increase in gross margin. The data center and client segments saw significant growth, with data center revenue reaching a record high of $2.3 billion. This growth was driven by strong sales of AMD's Instinct MI300X GPU and server CPUs, leading to a gain in server CPU revenue share. Hyperscalers continued to adopt AMD's fourth-gen EPYC processors, with nearly 900 AMD-powered public instances available globally. In the enterprise market, there is increasing demand for general purpose and AI compute capacity, which aligns with the value proposition of AMD's EPYC processors.
AMD's EPYC CPUs offer high core count and energy efficiency, resulting in a 45% reduction in servers needed and significant cost savings for enterprises. The adoption of EPYC CPUs is increasing, with major companies such as American Airlines and Oracle using them. The next generation of EPYC processors, Turin, is expected to launch later this year and offers even greater performance and efficiency. In the data center GPU market, AMD's MI300 has become the fastest ramping product in the company's history, with over $1 billion in sales in less than two quarters. MI300x is being deployed by major cloud providers and is also entering volume production for enterprise use. Over 100 enterprise and AI customers are currently using or developing MI300x.
In the AI software front, AMD has made significant progress by adding support for AMD hardware in the OpenAI Triton compiler and releasing a major update to their ROCm software stack. This has resulted in strong performance and significant TCO advantages for partners using MI300x GPUs. AMD is also working with partners to enable the broad ecosystem needed for the next generation of AI systems. Demand for MI300 continues to strengthen, and AMD expects data center GPU revenue to exceed $4 billion in 2024. In the client segment, revenue increased by 85% year-over-year due to strong demand for the latest generation Ryzen processors.
AMD saw strong growth in sales of their Ryzen desktop and mobile CPUs, with a double-digit increase in desktop sales and nearly doubling in mobile sales. The company also launched their Ryzen Pro 8000 series processors, which offer industry-leading performance and battery life for commercial notebooks. AMD sees opportunities for further growth in the commercial PC market, driven by the start of an enterprise refresh cycle and the adoption of AI PCs. They are working closely with Microsoft and other partners to enable AI experiences powered by Ryzen processors. Later this year, they will launch their next-generation Ryzen mobile processors, which are expected to have significant performance and energy efficiency improvements. AMD expects to grow their revenue share this year due to the strength of their Ryzen CPU portfolio.
In the Gaming segment, revenue declined due to the console cycle and the introduction of new technology. In the Embedded segment, revenue also decreased as customers focused on inventory levels, but there is potential for growth in the AI market. AMD announced their second generation of Versal adaptive SoCs, which have improved AI and compute capabilities and have already been adopted by Subaru for their iSIGHT ADAS system.
AMD had a successful first quarter with strong execution and momentum in their Embedded Design wins. They are expecting strong annual revenue growth and expanded gross margin due to the growing adoption of their instinct, EPYC, and Ryzen products. Their priorities for 2024 include accelerating data center growth, launching next-generation processors, and expanding their adaptive computing portfolio. They see AI as a significant growth opportunity and have increased investments in this area. Jean then reviewed the financial results and provided an outlook for the second quarter, highlighting strong year-over-year revenue growth and gross margin expansion.
In the first quarter of 2024, AMD's revenue was $5.5 billion, up 2% from the previous year due to growth in the Data Center and Client segments. However, revenue declined 11% sequentially due to lower gaming and Embedded segment revenues. Gross margin increased by 230 basis points, but operating expenses also increased by 10% due to investments in R&D and marketing for AI growth opportunities. The Data Center segment saw record quarterly revenue of $2.3 billion, driven by the ramp of AMD Instinct GPUs and server process revenue growth. The Client segment also saw a significant increase in revenue, up 85% year-over-year, driven by Ryzen 8000 series processors.
In the second quarter, the company's client revenue decreased by 6% on a sequential basis, while their gaming segment revenue decreased by 48% year-over-year and 33% sequentially. The embedded segment also saw a decline in revenue of 46% year-over-year and 20% sequentially. However, the company's operating income for the client segment improved due to higher revenue, while the gaming segment and embedded segment saw a decrease in operating income. The company generated $521 million in cash from operations and has $6 billion in cash and investments. For the second quarter of 2024, the company expects data center and client segment revenue to increase, while embedded and gaming segment revenue is expected to decline significantly. Overall, the company expects a significant increase in data center and client segment revenue year-over-year, but a significant decline in embedded and gaming segment revenue.
The company expects second quarter non-GAAP gross margin to be 53%, operating expenses to be $1.8 billion, effective tax rate to be 13%, and diluted share count to be 1.64 billion shares. They have made progress on their strategic priorities and believe their investments will position them well for AI opportunities. The audience can ask questions in the Q&A session. The full year outlook for MI300 has been increased from $3.5 billion to $4 billion due to strong ramp and qualification in production data centers. The company is working closely with customers and things are going well. There have been concerns about supply constraints, but the company is handling it.
The company is confident in their MI300 product and has over 100 customers engaged in development and deployment. The supply chain for the product has ramped up quickly and they expect to continue ramping up supply every quarter this year. They are currently tight on supply due to high demand. The company is pleased with the ramp-up of both demand and supply. When asked about their Data Center GPU roadmap, the company's CEO could not provide too much detail but expressed confidence in their ability to compete in the market.
Lisa Su discusses the company's multiyear, multigenerational road map for their AI products, with follow-ons to MI300 and future generations in development. They are receiving feedback from top AI customers and are confident in their ability to remain competitive. They also have a strong road map for the non-AI side of their data center business, with good traction in the enterprise market. The EPYC business has performed well, but the market is mixed.
In the first quarter, the enterprise space showed signs of improvement with large customers starting refresh programs. The company expects strong double-digit growth in the data center and server business in the second quarter, with even more improvement in the second half of the year due to market conditions and the Turin launch. On the client side, the company expects sequential growth in the second quarter and an increase in units and ASPs in the second half of the year, driven by the AI PC products which are well suited for the premium segments of the market.
Lisa Su, CEO of AMD, addresses the company's strong position in the AI compute market. She mentions that the company's AI PC content will be strongest in the beginning and will expand across the portfolio by 2025. In response to a question about competition from internal ASIC programs at primary customers, Su highlights the growing TAM for AI compute and the company's strong relationships with top AI companies. She emphasizes the importance of innovation and collaboration with customers.
Lisa Su, CEO of AMD, discusses the use of large language models and the company's preferred architecture for training and inferencing. She also mentions their partnerships and commitment to innovation. In terms of short-term demand, Su states that it remains strong despite recent noise in the market.
Customers are moving from initial POCs to full-scale production and deployment of multiple workloads. The company is seeing strong demand and is building backlog and engagements with new customers. Supply momentum is also strong and the company is pleased with the speed of ramp. They have secured enough capacity to support significant upside to the ramp of the MI300 and have supply visibility beyond their current guidance of $4 billion. There is still potential for supply capacity upside. There is also strong demand for traditional servers.
Lisa Su discusses the growth potential for traditional server CPUs in the coming years, noting a need for refresh cycles and an increase in AI head nodes. AMD's strength lies in high-performance, energy-efficient processors, and they are seeing an expansion into cloud third-party workloads. Su expects continued growth for their server business through 2024 and beyond. When asked about supply constraints, Su clarifies that the $4 billion outlook for the year is not due to supply limitations but rather a combination of both supply and demand factors. She also mentions that they have better visibility for the back half of the year and expects strong demand for their GPUs.
Lisa Su clarifies that their $4 billion number for MI is not supply capped and they have the capability to exceed that amount. However, demand has exceeded expectations in the industry and they are working on ramping up capacity. They have good visibility into customer progression and continue to bring on new customers and expand workloads. In terms of the embedded business, they are seeing a measured rebound in the back half, similar to what other auto and industrial peers are experiencing. The inventory clearing cycle is ongoing and the impact on gross margin expansion is uncertain.
The speaker, Jean Hu, responds to a question about the company's gross margin expansion and the performance of their embedded business. They attribute the decline in the embedded business to weaker demand in certain markets, but expect a gradual recovery in the second half. The main driver of gross margin expansion in the second half is expected to be the strong performance of the data center sector, with potential additional tailwinds if the embedded business improves. The next question is about the company's data center GPU roadmap, with the customer demand shifting towards optimizing performance per square foot in rack scale systems. The speaker, Lisa Su, confirms that their customers are engaged in multigenerational conversations and they have all the necessary pieces for the market shift towards rack scale systems.
AMD is focusing on system integration and working closely with customers and partners to ensure scale-out capability. They have the necessary pieces for this, including their Infinity Fabric and Pensando acquisition. They are also collaborating with networking partners such as Broadcom, Cisco, and Arista. The server CPU business is expected to be back-end loaded this year, with a strong ramp in the data center segment and growth in server CPUs and data center GPUs.
During the earnings call, Lisa Su, CEO of AMD, was asked about the profitability of their MI300 product and the transition to their new Turin server platform. Su stated that the MI300 ramp has been successful but there is still room for improvement in terms of yield and process improvement. She also mentioned that the gross margin for MI300 should eventually be above corporate average. On the topic of the Turin server platform, Su believes that the transition will be smoother and faster compared to previous transitions, as it is the same platform. She also expects a broader range of workloads and an expansion of their serviceable addressable market with the introduction of Turin.
Stacy Rasgon asks about the MI300 ramp in Q2 and how it compares to the overall corporate ramp in guidance. Jean Hu explains that the data center GPU ramp will be more than the overall company's ramp and that gaming will be down significantly. She also provides insight into the weak demand and inventory levels in the gaming business. Stacy then asks about the near-term data center profitability, noting a 19% decrease in operating profit despite a 2% increase in revenue.
Harlan Sur asks about the faster time to production shipments for AMD's data center GPU segment and whether it is due to faster bring up of customers' frameworks driven by AMD's latest software platform or a more aggressive build-out plan by customers. Lisa Su explains that it is a combination of both, with the success of the Rockham stack and close collaboration with customers to optimize their key models leading to increased confidence in broadening the ramp.
The company has passed important milestones in terms of verification and validation, and there is a high demand for AI compute. The ramp has been aggressive and complex, but has gone well. The server CPU business was down at the lower end of the seasonal range, but still performed better than the competition due to share gains and strength in enterprise. The company is confident in its momentum and expects continued growth in 2024.
During a Q&A session, Lisa Su, CEO of AMD, answers a question about the competitive environment in the CPU market and potential risks for pricing. She states that their server CPU ASPs are stable and they are indexed towards higher core counts. She also mentions the importance of the open ecosystem and the Ultra Ethernet consortium, but does not see it as a limitation to their ability to build large-scale systems.
The speaker discusses the potential for Ethernet to become the long-term solution for networking in systems, and mentions that they are working internally and with partners to enable this. They also touch on the topic of large data centers and whether there will be room for multiple competitors to play a role in these environments. The speaker believes that there will be opportunity for multiple players and views this as an attractive area for AMD.
AMD is well positioned to compete in large-scale systems and is expected to see growth in both the MI300 and EPYC segments in the June quarter. However, the company is not providing specific revenue guidance for these segments. The data center segment is expected to grow double-digit sequentially. There will not be any further questions on the call.
This summary was generated with AI and may contain some inaccuracies.