Advanced Micro Devices is preparing to release its second-quarter financial results, with the outcome highlighting questions about AI spending. Since late June, the industry has collectively lost over $1 trillion in market capitalization. This has raised concerns that major cloud service providers might be struggling to maintain their aggressive spending on artificial intelligence infrastructure. The uncertainty has created a ripple effect throughout the sector. Even so, chipmakers have still managed to deliver strong overall performance over the past year.
Earnings and revenue expectations
AMD’s stock closed at $476.15 last Friday, down from over $550 just two weeks prior. This reflects investor caution. Analysts are anticipating earnings per share of $1.62 on revenue of $11.3 billion for the quarter. This would be a substantial increase compared to $0.48 and $7.6 billion in the same period last year. The company itself has guided to approximately $11.2 billion in revenue. A non-GAAP gross margin is expected to remain near 56%.
Sector performance and competitive pressures
The Philadelphia Semiconductor index has declined roughly 20% since hitting its June peak. This signals investor uncertainty. One factor contributing to the downturn is the rising cost of memory chips. This is negatively affecting the personal computer market. It also presents challenges for both AMD and Intel in the quarters ahead. Gaming revenue is also under pressure. It is projected to drop 30% to $781 million due to industry-specific headwinds.
The Helios rack-scale system stands out as one of AMD’s most significant product developments. This system integrates 72 of the company’s MI455X accelerators with Venice server processors and Pensando networking, and it is positioned as a direct competitor to Nvidia’s offerings. Priced at around $5 million to $5.5 million per unit, the system is designed to deliver enhanced performance and efficiency compared to rival platforms.
AMD has already secured commitments for the Helios system from OpenAI for up to six gigawatts of capacity. It also has commitments from Anthropic for up to two gigawatts. Additional deployments are planned at Microsoft. The system is expected to see first customer shipments later in the third quarter. However, meaningful revenue is likely to emerge only in the fourth quarter. The availability of HBM4 memory, a critical and scarce component in AI systems, remains a potential bottleneck for AMD’s production efforts.