AMD 預測 2027 年收入增長逾 80%,伺服器 CPU 業務收入料倍增
AMD Forecasts Over 80% Revenue Growth in 2027, Server CPU Sales to Double
AMD 預計 2027 年公司收入較今年增長逾 80%,伺服器 CPU 業務收入約倍增,而 AI GPU 收入增長更快。

AMD CEO Lisa Su. (Photo: THE ELEC)
AMD expects annual revenue to grow more than 80% in 2027 from this year, with server central processing unit (CPU) revenue doubling and artificial intelligence (AI) graphics processing unit (GPU) revenue growing even faster.
"Strong demand continues across data center CPUs and AI, with demand currently exceeding supply," AMD CEO Lisa Su said during the company's third-quarter earnings conference call on Nov. 3 local time. "We expect company-wide revenue to grow more than 80% next year, while the server CPU business will roughly double."
AMD's data center segment comprises its server CPU business and data center AI GPU business. Most of the growth in server CPU revenue is expected to come from higher shipment volumes. Average selling prices (ASPs) will also rise somewhat as the sixth-generation Venice server CPU increases the number of cores per chip.
Growth in data center AI GPUs is expected to be even stronger. "We expect data center revenue to grow 50% sequentially in the fourth quarter, when revenue from the data center AI GPU business will surpass that of the server CPU business," said AMD Chief Financial Officer Jean Hu. Su added that data center AI revenue growth would significantly outpace server CPU growth next year.
AMD said major AI model developers and hyperscale cloud service providers, including OpenAI, Meta, Microsoft, Oracle and Anthropic, have begun initial operations of its Helios data center server racks. The company plans to significantly increase shipments starting in the fourth quarter.
AMD also said Helion, its CPU-focused rack design for AI data centers, has received positive feedback from server manufacturers and customers. According to AMD, the hyperscaler rack uses a CPU-only configuration and can accommodate up to 144 Venice processors per rack.
AMD cited total cost of ownership (TCO), cost efficiency per token and the flexibility of general-purpose GPUs as key reasons major customers choose its products.
The company is co-designing its next-generation Instinct MI500 and MI600 GPUs with major AI model developers, optimizing performance for deployment in gigawatt-scale data centers.
AMD plans to keep growth in selling, general and administrative expenses and research and development (R&D) spending below half its revenue growth rate next year to improve operating margins. The company also said it is working with key supply chain partners to secure supplies of wafers, substrates, packaging and memory.
AMD reported record third-quarter revenue of $13.377 billion and operating profit of $2.832 billion under generally accepted accounting principles (GAAP), up 45% and 123%, respectively, from a year earlier. Data center revenue nearly doubled to $8.546 billion, accounting for 64% of total revenue. The company projected fourth-quarter revenue of $17 billion, plus or minus 3%, representing a 66% year-on-year increase.
來源:The Elec · thelec.net