AMD Server CPU Revenue Hits Record for 5th Straight Quarter: "Growth Proven, Moat Still Needed" [Click e-Stock]

Eugene Investment & Securities: "AMD Must Go Beyond Mere Top-Line Growth"

U.S. semiconductor giant AMD reported earnings on August 4 (local time) that exceeded market expectations, buoyed by strong data center demand. However, Eugene Investment & Securities commented on August 6 that AMD's profitability is yet to catch up with its rapid growth and did not provide an investment rating or target price for AMD. The closing price on the 4th was USD 518.6.

For the second quarter of fiscal year 2026, AMD posted revenue of USD 11.54 billion, a 50.1% increase year-on-year. Of this, data center revenue soared by 107% to USD 6.72 billion. This surge was driven by robust sales of both server CPUs for data centers and Instinct GPUs designed for artificial intelligence (AI) acceleration. The company’s revenue guidance for the third quarter also beat consensus (the average forecast by securities firms), coming in at USD 13 billion.

In particular, server CPU revenue hit a new all-time high for the fifth consecutive quarter. This trend is being fueled by continued strong server demand, driven by the spread of ‘agent AI’—AI that can make autonomous decisions and carry out tasks independently.

AMD Chief Executive Officer (CEO) Lisa Su is giving a speech at the Computex Forum held in Taipei, Taiwan, in March 2024. Photo by Reuters

Park Jaehwan, a researcher at Eugene Investment & Securities, explained that not only Turin, the latest fifth-generation EPYC product, but also the previous generation Genoa remain strong sellers. "Product lineup improvements through generational upgrades are driving up the average selling price (ASP), while stable supply chain management is boosting shipment volumes as well," he said. The company projected that server CPU revenue will grow more than 80% year-on-year in the second half of this year and over 70% in 2027.

The key to next year's growth lies in the rack-level product ‘Helios’. Production is already underway, with initial shipments scheduled to begin at the end of the third quarter and volumes expected to ramp up significantly from the fourth quarter. AMD has already secured OpenAI, Meta, Anthropic, and Microsoft as early customers. Thanks to the Helios effect, total data center segment revenues are projected to more than double by 2027. AMD reaffirmed its ambitious growth targets, anticipating that the high-performance computing (HPC) market will reach USD 2 trillion, the AI accelerator market USD 1.4 trillion, and the server CPU market USD 220 billion by 2030, outpacing overall market growth rates.

Gross profit margin (GPM) for both the second and third quarters stood at 56%, which fell slightly short of market expectations. The upcoming MI455X GPU, set for release in the second half of the year, will be equipped with 432 gigabytes (GB) of high-bandwidth memory (HBM4), which is significantly higher than Nvidia's Rubin (288GB) and consequently results in higher production costs.

Researcher Park pointed out, “AMD’s GPM, stuck in the mid-to-high 50% range, continues to be compared with Nvidia, which maintains around 75%.” While server CPUs contribute positively to margins, it is difficult to further increase CPU shipments because the new GPUs and CPUs must share the same leading-edge manufacturing and packaging facilities at Taiwan’s TSMC.

The competitive landscape ahead also presents variables. News that SpaceX will use only Nvidia products for its AI infrastructure confirms Nvidia‘s dominance, while the rise of the application-specific integrated circuit (ASIC) camp may leave AMD in a more precarious position. In addition, AMD is currently trading at a higher premium, with a forward price-to-earnings ratio (PER) of 48 times based on earnings forecasts for the next 12 months—higher than Nvidia.

While Park voiced optimism about AMD’s growth trajectory, given its competitive advantage in server CPUs and momentum from Helios shipments, he noted that for the stock to outperform the market, “AMD must go beyond mere top-line growth and establish its own competitive moat within the AI infrastructure value chain.”

This content was produced with the assistance of AI translation services.