news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

After the U.S. stock market closed on Tuesday, July 30, AMD, a semiconductor giant that is striving to catch up with Nvidia in the data center GPU field, released its second-quarter financial report for fiscal year 2024. Investors will pay close attention to the full-year sales guidance for its MI300 AI accelerator chip.

As AMD's second-quarter revenue and profits both slightly exceeded expectations, its data center AI revenue doubled year-on-year and set a new record for two consecutive quarters. The midpoint of its third-quarter revenue guidance range was $6.7 billion, which was also higher than the market expectation of $6.62 billion, causing its stock price to rise 5% after the market, driving its competitor Nvidia's after-hours stock price from a 3% drop to an increase of about 3%.

On the earnings call at 5 p.m.AMD management raised its sales forecast for data center GPUs in 2024 from $4 billion to $4.5 billion, and said that Microsoft's use of MI300 chips has increased, intending to refute market rumors that Microsoft has cut orders for MI300. AMD's after-hours gains quickly expanded to 9%, while Nvidia maintained an after-hours gain of 3%.

1) Main financial figuresaccording to

Quarterly revenue:US$5.835 billion, a year-on-year increase of 9% and a month-on-month increase of 7%. The market expected US$5.73 billion.
Gross profit margin:Under GAAP, it was 49%, compared with 46% in the same period last year; under non-GAAP, it was 53%, compared with 50% in the same period last year.
operating profit:GAAP items were US$269 million, compared with a loss of US$20 million in the same period last year; non-GAAP items were US$1.264 billion, up 18% year-on-year and 12% month-on-month, and the market expected US$1.25 billion.
Net Profit:GAAP items were US$265 million, up 881% year-on-year; non-GAAP items were US$1.126 billion, up 19% year-on-year and 11% month-on-month.
Diluted EPSEarnings per share:GAAP items were $0.16, up 700% year-on-year; non-GAAP items were $0.69, up 19% year-on-year and 11% month-on-month, and the market expected $0.68.

2) Prospectssee

Third quarter revenue:It is expected to be between $6.4 billion and $7 billion, with the midpoint of the range being $6.7 billion, which would represent year-over-year growth of approximately 16% and month-over-month growth of approximately 15%. Analysts expected $6.62 billion.
Third Quarter Non-GAAPGross profit margin:The forecast is 53.5%, while analysts expect 53.8%.
MI300 accelerator sales:Data center GPU sales are expected to be $4.5 billion in 2024. Microsoft's use of the MI300 chip has increased.
Other highlights of the call:In the second quarter, MI300 chip revenue exceeded US$1 billion, and it is committed to launching new AI processors/chips every year.

3) Segmented business data

Data Center Division:Revenue hit a record high of $2.8 billion, up 115% year-over-year and 21% quarter-over-quarter.
Client Division:Revenue was $1.5 billion, up 49% year-over-year and 9% quarter-over-quarter.
Game Division:Revenue was $648 million, down 59% year-over-year and 30% quarter-over-quarter.
Embedded Business Unit:Revenue was $861 million, down 41% year-over-year and up 2% quarter-over-quarter.

AMDSecond quarter highlights: Data center AIRevenue doubled year-on-year, with a growth rate exceeding that of the previous quarter, and will also make outstanding contributions to the future

AMD's financial report also stated that the adjusted operating profit margin in the second quarter was 22%, basically in line with analysts' expectations of 21.8%. Capital expenditures in the second quarter were $154 million, higher than analysts' expectations of $127.1 million. R&D expenditures in the second quarter were $1.58 billion, in line with market expectations.

In terms of quarterly total revenue, the year-on-year growth of 9% was significantly narrowed from the growth of 18.2% in the second quarter of last year, while the non-GAAP gross profit margin continued to expand from 52% in the previous quarter, and non-GAAP adjusted earnings per share continued to be higher than US$0.62 in the previous quarter.

Most importantly, data center revenue hit a record high for two consecutive quarters, doubling the growth rate over the same period last year.This is stronger than the 80% year-on-year surge in the first quarter and significantly faster than the 2% month-on-month growth in the previous quarter.

AMDManagement emphasized AI in its earnings statementOutstanding contribution to current and future growth.In the earnings statement, the company's chairman and CEO Dr. Lisa Su said that the second quarter achieved strong revenue and profit growth, driven by record data center revenue:

“Our AI business continues to accelerate, and we are well positioned to deliver strong revenue growth in the second half of the year, led by demand for our Instinct, EPYC and Ryzen processors. The rapid advancement of generative AI is driving the need for more compute in every market, creating tremendous growth opportunities as we deliver leading AI solutions across our business.”

CFO Jean Hu also said that second-quarter revenue was higher than the midpoint of the company's guidance range, thanks to strong growth in the data center and client divisions. While steady profit growth was occurring, gross margins continued to expand, "increasing strategic AI investments to lay the foundation for future growth."

AMD said it is committed to launching new AI processors/chips every year, in an effort to catch up with Nvidia, which has previously moved up its new product release schedule from once every two years to once a year.

Financial report features: Data center AIRevenue growth accelerated, PCChip revenue grew steadily, but gaming and embedded systems continued to perform poorly

AMD said that its most important data center AI revenue doubled year-on-year in the second quarter, thanks to a significant increase in Instinct GPU shipments and strong growth in fourth-generation EPYC CPU sales. The month-on-month increase in revenue was due to strong growth in Instinct GPU shipments.

Meanwhile, the revenue growth of the client division, which includes PC chip sales, was mainly due to the sales of Ryzen processors. The revenue of the gaming division was cut in half year-on-year due to the decline in semi-custom revenue, and the revenue of the embedded division declined because customers continued to destock.

Before the financial report, the market expected the second quarter's gaming revenue to drop by more than 58% year-on-year to US$655 million, and the embedded business unit's revenue was expected to drop by nearly 42% year-on-year to US$850 million. It can be seen that AMD's gaming revenue in the second quarter was lower than expected, while the embedded revenue was higher than expected.

In the first quarter, its gaming revenue dropped 48% year-on-year and 33% month-on-month to $922 million, mainly due to a decrease in semi-custom business revenue and a decline in Radeon GPU sales. The embedded division's revenue dropped 46% year-on-year and 20% month-on-month to $846 million, mainly due to customers continuing to adjust their inventory levels.

That is to say,In the second quarter, AMD's chip sales in data centers and PC clients did continue to grow significantly year-on-year, while gaming and embedded revenues continued to decline significantly year-on-year. However, in the first quarter, client revenue surged 85% year-on-year to $1.4 billion, higher than the year-on-year growth in the second quarter, but the revenue in this area fell 6% month-on-month at that time, which was lower than the 9% month-on-month growth in the second quarter.

This year AMDThe stock price has fallen and seriously underperformed the market and chip index, but some analysts are optimistic about the potential of each business line

Before the earnings report was released, AMD fell as much as 4% on Tuesday and then closed down 0.9%, approaching the lowest level in nearly seven months since January 4, highlighting the cautious sentiment of the market. The stock has fallen nearly 7% so far this year, significantly underperforming the S&P 500 and Nasdaq, which have risen by about 14%, and the Philadelphia Semiconductor Index, which has risen by 18%.

However, Intel, a competitor in the CPU field that has been severely hit by AMD, has fallen nearly 40% this year, which is even worse than AMD. Nvidia, which has become the "darling of AI" with its GPU hardware and related software services, has doubled its stock price this year to a cumulative increase of 114%.

Since July 10, AMD has fallen 25% from its four-month high, becoming one of the worst performing stocks in the S&P 500. However, Wall Street's mainstream rating is still "buy", with an average target price of more than $190, representing about 40% room for growth.

Srini Pajjuri, an analyst at brokerage firm Raymond James, rated AMD as "outperform/overweight", believing that AMD's MI300 AI accelerator product revenue in the second quarter may reach US$900 million, and the full-year sales guidance in this area will also be "increased by several hundred million dollars." AMD currently expects MI 300 to bring incremental revenue of more than US$4 billion in 2024.

The analyst said the PC market looks relatively stable, the server (CPU) market shows signs of recovery, and the gaming market should rebound from depressed levels in the second half of the year, but suppliers of automotive/industrial integrated circuits hinted that Xilinx's embedded market recovery may be slower.

According to IDCstatistics,Global PC shipments grew 3% year-on-year in the second quarter, marking the second quarter of growth after eight consecutive quarters of year-on-year declines. As Nintendo prepares to launch its next-generation game console and video game company Take-Two is set to launch the highly anticipated Grand Theft Auto VI later next year, many organizations expect the gaming industry to rebound in late 2024 and 2025.

Why it matters and what to watch: AICan the annual outlook for revenue be revised upward?

AMDThe first of the three major chip companies to announce its second quarter financial report, Intel and Nvidia will report earnings on August 1 and August 28, respectively. Investors will closely monitor AMD's data center unit sales performance to speculate whether corporate AI infrastructure spending continues to rise, and any guidance and management comments on GPU sales will also be closely watched.

CJ Muse, an analyst at Cantor Fitzgerald, a US financial services company, pointed out that AMD has fallen more than 20% in the past two weeks, mainly due to rumors that Microsoft may cut orders for its MI300 AI accelerator. If AMD fails to raise its full-year sales guidance for this key chip in this financial report, its stock price will face the risk of further decline. If the guidance is raised to more than US$5 billion, it will be very beneficial.

AMD's current top GPU chip is the Instinct MI300X accelerator, which the company revealed in June has been adopted by partners and customers including Microsoft, Meta, Dell, Hewlett Packard Enterprise HPE and Lenovo. Its next-generation MI325X will be launched in the fourth quarter of this year, and the updated MI350X will be available in 2025, followed by the MI400 in 2026.

Currently, mainstream Wall Street analysts have different opinions on how much AMD's AI revenue is "good enough."

Wedbush, a brokerage firm, believes that the highest expected level for MI300 shipments this year is $6 billion. Barclays believes that even when there is a lot of negative news about Microsoft in the supply chain, annual sales of $5 billion can be achieved. Susquehanna is worried that after this year's Nvidia GTC AI Developer Conference, AMD's accelerator will be cut, which has caused a certain degree of concern.

Wall Street is most concerned about AMDCan it effectively compete with Nvidia?

This also shows that AMDGPUs from NVIDIAThe battle will be the focus of the market。

Rosenblatt Securities believes that the pressure on AMD's stock price this year reflects investors' concerns about whether its products can compete with Nvidia, which has an absolute lead in AI accelerators and data center GPUs with its H and more powerful B series chips:

"But contrary to popular belief, AMD's MI300 R&D roadmap is becoming more competitive due to software optimizations and higher configurations of high-bandwidth memory HBM3e, and AMD's open approach will also be rewarded compared to Nvidia's more closed technology stance."

AMD mentioned in its second quarter report that it and industry leaders announced the establishment of the Ultra Accelerator Link promoter group, which will use AMD Infinity Fabric technology to advance AI network infrastructure systems based on open standards.

Stacy Rasgon, an analyst at Bernstein, said that recent news involving possible problems with high-bandwidth memory and the planned retirement of AMD President Victor Peng may constitute two "less favorable" factors. However, AMD's valuation multiples have eased in recent weeks of selling, with a price-to-earnings ratio of about 30 times, which is closer to the agency's target level.

Another analyst wrote on SeekingAlpha that it is difficult to value AMD at present because there is "considerable uncertainty" as to whether AMD can successfully compete with Nvidia in the data center GPU market. The three data points that should be paid most attention to in the second quarter financial report are: the backlog of MI300 orders, the production schedule of the next-generation chip MI325, and customer engagement with data center GPUs.

The article stated that AMD's MI300 order backlog only increased by about $500 million in the first quarter, a slow and unsatisfactory growth rate, causing the stock price to fall nearly 9% the day after the earnings report, because "only $500 million in backlog orders in a quarter seems too small for AMD to maintain data center GPU revenue growth. The current surge in demand for data center GPUs should make it easier for AMD to win customers."

Backlogs are not projected sales, but "firm commitments made by customers to AMD regarding their needs." However, the slow growth of backlogs may not be related to the difficulty of competing with Nvidia, but rather to macro factors such as customers' investment considerations in AI infrastructure. If backlog growth picks up again, it will send a positive signal that AMD will be able to maintain and increase data center GPU sales next year.

At the same time, if AMD's data center GPU customer participation increases and is more easily converted into actual orders, it also means that AMD will be able to compete with Nvidia in the coming quarters. In the first quarter financial report, its management revealed that "more than 100 enterprises and AI customers are actively developing or deploying MI300x, up from dozens in the fourth quarter of last year."

The MI325, which is scheduled to be launched in the fourth quarter of this year, will upgrade the GPU memory from 192GB HBM3 to 288GB HBM3e. The speed of its production increase will also determine AMD's competitive position with Nvidia's current generation Hopper GPU and the upcoming Blackwell GPU. Some analysts believe that since there are no major changes in the architecture of MI325, it may be easier to increase production than MI300.