AMD expects its data-center business revenue to grow by more than 100% as the company moves toward 2027 [5].
This growth signals a shift in the artificial intelligence landscape, suggesting that corporate AI adoption is moving from a testing phase into large-scale operational deployment.
Speaking during an interview on CNBC’s “Squawk on the Street,” CEO Lisa Su said the company sees tremendous momentum in its data-center business. Su said the AI market has reached an "inflection" in demand as enterprises move beyond experimentation and begin deploying AI in production [1].
Recent financial data highlights this trajectory. In the second quarter of 2026, AMD reported data-center revenue of $6.7 billion [3], representing a 107% increase year-over-year [1], [2]. This follows a period of steady climb, including Q3 2025 when the segment recorded $4.3 billion in revenue [4].
To meet this demand, AMD is expanding its hardware offerings. A company spokesperson said Helios is AMD's first rack-scale system for AI [7]. This infrastructure is designed to provide the massive compute capacity that Su said customers now require.
While AMD targets a doubling of data-center revenue by 2027 [5], other analysts provide different projections. A Morningstar forecast suggests the server-CPU business specifically could grow more than 70% by 2027 [6]. The discrepancy highlights the rapid volatility and scaling potential of the AI infrastructure market.
“The AI market has reached an "inflection" in demand”
The transition from AI experimentation to production deployment indicates that enterprises are finding tangible value in generative AI, moving beyond prototypes to integrate these tools into their core business processes. For AMD, this represents a critical window to capture market share from competitors by offering integrated rack-scale solutions like Helios, rather than just individual chips.
