AMD Hits Record $13.4B Revenue as Data Center Sales Nearly Double

Santa Clara, California — U.S. chipmaker AMD posted record third-quarter revenue of $13.4 billion, up 45% from a year earlier, as sales of its Instinct graphics processors and EPYC server processors almost doubled the size of its Data Center business, the company said on November 3.

AMD's third-quarter results show GAAP net income of $2.4 billion for the period, equal to $1.46 per diluted share, with operating income of $2.8 billion and a gross margin of 54%. On the non-GAAP basis the company also publishes, which excludes certain items, gross margin reached 56%, operating income $3.9 billion, net income $3.5 billion and diluted earnings per share $2.09.

The top line compares with $9.2 billion in the third quarter of 2025 and $11.5 billion in the second quarter of 2026, a sequential gain of 16%. GAAP operating income jumped 123% from $1.27 billion a year earlier, while net income climbed 95% from $1.24 billion. Revenue also cleared the outlook AMD gave in August, when it guided to approximately $13 billion, plus or minus $300 million.

Dr. Lisa Su, AMD chair and CEO, said: “We delivered a record quarter with revenue exceeding the high end of our guidance as Data Center segment sales nearly doubled year-over-year,” adding: “Demand for our Instinct and EPYC products continued to strengthen, and we expect growth to accelerate further in the fourth quarter as we scale Helios and ‘Venice’ production. With our leadership product portfolio and deep customer and supply chain relationships, we are confident in our ability to sustain very strong revenue growth and expanding profitability over the coming years.”

Looking to the last three months of the year, AMD expects revenue of approximately $17 billion, plus or minus 3%. At the midpoint, that would mean growth of about 66% from a year earlier and about 27% from the third quarter, with non-GAAP gross margin forecast at approximately 55%. The company stressed that the outlook reflects current expectations and that actual results could differ, and executives discussed the results on a conference call at 2:00 p.m. PT (5:00 p.m. ET) on the day of the release.

The Data Center segment brought in a record $8.5 billion, nearly two-thirds of the quarter's sales and 97% more than a year earlier. AMD put the increase down to strong demand for EPYC processors and the launch of its Instinct MI450 Series GPUs. Three months earlier, the same segment had reported $6.7 billion in revenue after growing 107%.

AMD's other two segments moved in different directions. Client and Gaming revenue slipped 8% to $3.7 billion: the client business grew 7% to $2.9 billion on sales of Ryzen mobile processors, but gaming revenue fell 40% to $773 million as semi-custom and graphics sales declined. Embedded revenue rose 29% to $1.1 billion, with aerospace and defense, automotive, communications and industrial customers leading the growth.

Jean Hu, AMD executive vice president, chief financial officer and treasurer, said: “Third quarter revenue grew 45% year-over-year to a record $13.4 billion, while non-GAAP earnings per share grew 74%, demonstrating the significant operating leverage in our business model as our Data Center business drives scale,” adding: “Our fourth-quarter outlook reflects accelerating revenue and earnings growth, driven by strong demand across our Data Center portfolio.”

On the customer side, AMD said Anthropic, Meta, Microsoft, OpenAI and Oracle Cloud Infrastructure are already running models and workloads on its first Helios rackscale systems, which pair Instinct MI450 Series GPUs with 6th Gen EPYC CPUs. HPE has said cloud provider Vultr will be the first to deploy Helios systems built by HPE. Under a partnership AMD disclosed in August, Anthropic plans to use up to 2 gigawatts of MI450 Series GPUs housed in Helios racks.

Oracle Cloud Infrastructure also said its OCI Compute E7 instances will run on 6th Gen EPYC CPUs, handling enterprise applications and the CPU services that sit around AI models. Google Cloud and Microsoft each made storage-optimized cloud instances based on 5th Gen EPYC chips generally available, and the EuroHPC Joint Undertaking chose Instinct MI430X GPUs and 6th Gen EPYC processors to power its LUMI-AI supercomputer.

AMD also introduced Helion, a reference architecture it describes as an open blueprint for building rack-scale agentic systems, in which customers decide the balance of CPUs and GPUs according to their workloads. On the software side, it released ROCm 10, which brings AI-native developer tools to AMD platforms through ROCm.AI, together with faster performance and updated libraries, compilers, frameworks, tools and model support.

The company announced agreements to acquire two companies during the period. World Labs, AMD said, brings model research expertise, while Taalas adds specialized inference silicon that AMD expects to set its system-level products apart on inference performance and efficiency. The World Labs deal is an all-stock transaction valued at approximately $8.2 billion, expected to close by the end of 2026, after which co-founder Fei-Fei Li is set to become AMD's executive vice president and chief scientist.

In embedded and physical AI, AMD started sampling the Versal AI Core XQRVC1902, a space-grade adaptive system-on-chip built to support missions of up to 15 years. Kodiak AI said its next-generation autonomous truck platform runs on EPYC processors, and AMD launched Ross, an agentic design tool intended to help customers bring embedded systems to market faster and get more performance from AMD embedded products.

In PCs, the Ryzen AI Halo platform, based on Ryzen AI Max PRO 400 Series processors, is now available to developers and creators who want to build and run models on their own machines. HP, Lenovo, Framework and GMKtec widened the range of systems using those chips, and AMD unveiled the Threadripper Halo Station, which combines a Threadripper 9955WX CPU with an Instinct MI350p GPU and, according to the company, can run trillion-parameter models locally.

The fourth-quarter ramp of Helios and “Venice” that Su described could also matter for buyers in the Gulf. Saudi AI company HUMAIN agreed with AMD in 2025 on a $10 billion collaboration to deploy 500 MW of AI compute capacity over five years, and AMD named HUMAIN among the AI labs and cloud providers deploying Helios when it reported second-quarter results in August.