However, the company’s stock dropped nearly 9% in after-hours trading. This suggests that investors were expecting even stronger results following the stock’s sharp rise this year, driven by enthusiasm for artificial intelligence technologies.
On a post-earnings call, CEO Lisa Su highlighted the company’s ambitions for the future. She noted that data-center revenue is expected to more than double by 2027. She said AMD plans to exceed its previously announced revenue growth target. The target is over 35% per year.
Investor expectations outpace guidance
Jacob Bourne, an analyst at Emarketer, pointed out the challenge AMD now faces. He remarked, “AMD is now in a similar position to Nvidia and the hyperscalers. Where investors are looking for evidence that AI infrastructure investments will continue translating into accelerating returns.” His comments reflect the high expectations for consistent growth. This is in this competitive sector.
Even though AMD exceeded second-quarter expectations, investors were likely banking on stronger results. The data-center segment contributed $6.72 billion in revenue, beating forecasts of $6.48 billion. The company also reported $11.54 billion in total revenue for the quarter, which was ahead of the $11.28 billion anticipated.
Integrated AI systems push market reach
AMD is broadening its approach beyond individual chips by introducing full AI systems that integrate processors, networking components, and other hardware. These systems are engineered to rival Nvidia’s rack-scale solutions, which include the next-generation “Vera” CPU and “Rubin” GPUs. This shift aims to offer customers a more seamless and competitive option.
A significant development for AMD came in July. The company inked a deal to supply Anthropic with AI servers. The servers use up to 2 gigawatts of MI450 chips from early 2027. The agreement also includes AMD’s potential $5 billion investment in the startup. The investment is contingent on meeting deployment targets.
Additionally, AMD reached an agreement with Core Scientific last week for up to 2.5 gigawatts of data-center capacity. The deal includes stock warrants for AMD, offering further strategic value and influence.
At its AI-focused event in July, Su confirmed that AMD’s second-generation Helios AI servers are now in full production. The servers feature the MI455X AI accelerator and TSMC’s “Venice” processor. They are set to be delivered soon.
CPU demand aids in gaining market share
The rising use of central processing units (CPUs) in servers, especially for inference tasks like chatbot queries, has helped AMD expand its market share at the expense of Intel. As a response, Intel has revised its spending plans for the next two years to stay competitive amid the growing demand for such technologies.
AMD’s gross margin for the second quarter came in at 56%, matching expectations. Data-center sales rose 16.3% from the first quarter, and the company reported an adjusted earnings per share of $1.66, which beat the $1.62 forecast.
Looking ahead, AMD is preparing a new rack-scale AI platform for 2027. This platform will integrate MI500 GPUs, Verano CPUs, and Pensando networking. The system is built to support both copper and optical connectivity, offering a competitive alternative to rivals like Nvidia.
The company continues to build momentum in the AI sector, with its second-quarter data-center revenue more than doubling to $6.72 billion. This outperformed forecasts of $6.48 billion and showed a significant 16.3% sequential increase from the first quarter’s $5.78 billion.