Advanced Micro Devices (AMD.O) reported weaker-than-expected data center revenue in its second quarter, disappointing investors betting on the company’s AI chip growth potential. Shares of the Santa Clara-based chipmaker fell roughly 4% in extended trading.
While AMD’s stock has climbed over 40% this year—outperforming the chip index’s 12% gain—its data center segment growth lagged behind rival Nvidia (NVDA.O), the dominant player in AI chips. Nvidia’s data center revenue surged 73% to $39.11 billion in its fiscal first quarter, driven by demand for its Blackwell GPUs and networking hardware.
AMD’s second-quarter data center revenue grew 14% to $3.2 billion, close to analysts’ estimate of $3.22 billion. This segment includes both server CPUs and Instinct AI chips. Portfolio manager Dan Morgan from Synovus Trust noted the “lackluster” data center results were concerning given AMD’s reliance on this segment.
CEO Lisa Su said the decline in AI chip revenue year-over-year was due to U.S. export restrictions on shipments to China and the transition to next-gen MI350 AI chips. Production of the MI350 series began ahead of schedule in June, with a planned steep production ramp in the second half of the year.
AMD also revealed that shipments of its MI308 AI chips to China remain on hold pending U.S. government export license approvals, impacting revenue. The company expects to resume shipments once licenses are granted. These export curbs are estimated to reduce AMD’s 2025 revenue by about $1.5 billion, mainly affecting Q2 and Q3.
For Q3, AMD forecast revenue of approximately $8.7 billion (±$300 million), above analyst expectations of $8.3 billion. The company projected adjusted gross margins around 54%, in line with estimates.
Adjusted earnings per share for Q2 were 48 cents on revenue of $7.69 billion, excluding stock-based compensation and other items.