AMD now expects its revenue to climb "substantially above" the company's earlier forecast of more than 35 percent growth, CEO Lisa Su announced Tuesday during the chipmaker's second-quarter earnings call. The upgraded outlook comes as demand for AMD's AI accelerators and server processors continues to surge, while the company successfully navigates rising component costs and supply constraints. Su told investors that AMD anticipates the data center AI accelerator market will expand by more than 55 percent each year to reach approximately $1.4 trillion by 2030, with the server CPU market growing more than 50 percent annually to about $220 billion by decade's end.
During the three months ended June 27, AMD brought in $11.5 billion in revenue, climbing 50 percent from the same period last year and 13 percent from the previous quarter. The company's data center segment delivered a record $6.7 billion, more than doubling year over year and jumping 16 percent quarter on quarter. That segment now accounts for 58 percent of AMD's total revenue, up from 42 percent a year earlier. Server CPU revenue hit a fifth consecutive quarterly record, with both cloud and enterprise sales surging more than 70 percent year over year. Sales of Instinct data center GPU accelerators more than doubled compared to the prior year. Gross profit doubled to $6.2 billion, while gross margin reached 54 percent, up 14 percentage points year over year. The client business generated $3.1 billion in revenue, growing 23 percent annually, with Ryzen Pro business-class processor sales climbing more than 50 percent. AMD's embedded business posted $977 million in revenue, marking 19 percent annual growth and the segment's strongest expansion in over three years. Net income reached $2.3 billion, more than doubling from the same quarter last year and jumping 66 percent from the prior quarter.
"We are still in the early innings of a multiyear AI adoption cycle, and the opportunity ahead is enormous," Su said on the call. The CEO stated that AMD is "exceptionally well positioned to capitalize on this opportunity and deliver significant growth in the coming years." The company disclosed that over 3 million AI models now operate directly on its product lineup. AMD now projects its data center segment revenue will more than double year over year in 2027, according to the earnings announcement. Su described the overall market for high-performance AI computing as growing roughly 40 percent annually over coming years, approaching $2 trillion by 2030, adding that "we expect to grow well above the market."
AMD credits several factors for its ability to raise guidance despite industry-wide supply pressures. Su told analysts the company worked with memory suppliers years in advance to secure High Bandwidth Memory allocation needed to meet 2027 delivery commitments. The chipmaker's use of chiplet technology allows it to ramp production with fewer wafers on new process nodes compared to competitors, providing flexibility during the current supply crunch. AMD adjusted quickly as server demand hit an inflection point over recent quarters, ensuring adequate supplies of wafers, back-end capacity, substrate, and other components, according to Su. The company expects improved supply availability going into 2027 to support projected data center growth exceeding 70 percent year over year. AMD forecast third-quarter revenue of about $13 billion, which at the midpoint would represent 41 percent annual growth and 13 percent sequential growth, with CFO Jean Hu projecting "very strong" double-digit expansion in both data center and embedded segments.
Su identified agentic AI as the fastest-growing segment of the server market and expects it to become the largest over the next three to five years, though it remains the smallest today. The company's sixth-generation Venice Epyc server processor family is now in production, with every major OEM on track to launch platforms and leading cloud providers planning deployments beginning later this year. AMD's Helios integrated, rack-scale AI infrastructure is ramping at the end of the third quarter, with more substantial growth anticipated in the fourth quarter and beyond, running ahead of the company's 2027 demand forecasts. While Su acknowledged a softer PC market in the second half due to higher memory and component costs, she maintained optimism that local AI will grow in share over time and that AMD's client business will outperform the broader market. The inference market should expand "very substantially" from this year into next, fueling AMD's recent partnership with Cerebras to offer fast inference capabilities in Cerebras' cloud starting in the fourth quarter. Partners will determine whether the chipmaker can sustain growth rates that outpace an AI market already expanding at historic speed, particularly as memory constraints and component costs test the economics of deploying ever-larger models at scale.