Advanced Micro Devices reports news on high-performance and AI computing products used across cloud and AI infrastructure, embedded systems, AI PCs and gaming. Recurring updates cover AMD Instinct GPUs, AMD EPYC processors, Ryzen AI processors, networking and software, along with customer and ecosystem collaborations tied to AI infrastructure and full-stack computing platforms.
Company news also includes quarterly financial results, earnings-call materials, annual stockholder meeting notices, product portfolio launches and strategic technology collaborations involving memory, compute infrastructure and AI platforms. These developments reflect AMD’s role as a semiconductor operating company serving data center, PC, gaming, embedded, industrial and automotive markets.
AMD (NASDAQ: AMD) and Eviden will build Alice Recoque, France’s first and Europe’s second exascale supercomputer, funded at 544 million euros by EuroHPC JU and partners on Nov 18, 2025.
The system uses next-gen AMD EPYC "Venice" CPUs, Instinct MI430X GPUs (432 GB HBM4, 19.6 TB/s per GPU, FP4/FP8 support), AMD FPGAs, Eviden BXI networking, BullSequana XH3500 platform and DDN storage. Alice Recoque spans 94 racks and targets >1 exaflop HPL with claims of 25% fewer racks and up to 50% better energy efficiency per GPU.
AMD (NASDAQ: AMD) presented its long-term strategy at Financial Analyst Day on Nov 11, 2025, targeting accelerated growth driven by data center and AI products.
Key numerical targets: >35% revenue CAGR company-wide, non-GAAP operating margin >35%, and non-GAAP EPS >$20 over the next 3–5 years. AMD expects >60% revenue CAGR in data center and >80% CAGR in data center AI, a plan for >50% server CPU revenue share, and product launches including Helios (MI450) in Q3 2026 and MI500 in 2027.
AMD (NASDAQ:AMD) reported record Q3 2025 revenue of $9.246 billion, up 36% year‑over‑year, with GAAP gross margin of 52% and non‑GAAP gross margin of 54%. GAAP operating income was $1.27 billion and GAAP diluted EPS was $0.75; non‑GAAP operating income was $2.238 billion and non‑GAAP diluted EPS was $1.20. Data Center revenue was $4.3 billion (+22% YoY) and Client & Gaming revenue was $4.0 billion (+73% YoY). Embedded revenue declined to $857 million (-8% YoY). AMD expects Q4 2025 revenue of ~$9.6 billion ± $300 million and non‑GAAP gross margin of ~54.5%. The results exclude any revenue from AMD Instinct MI308 shipments to China.
AMD (NASDAQ: AMD) and the U.S. Department of Energy will deliver two next‑generation supercomputers at Oak Ridge National Laboratory: the Lux AI supercomputer and the Discovery supercomputer. Together these systems represent a combined $1 billion of public and private investment to build a secure, federated U.S. AI and science infrastructure.
Lux, powered by AMD Instinct MI355X GPUs and EPYC CPUs, will deploy in early 2026 as the nation’s first dedicated AI Factory for science. Discovery, using EPYC "Venice" CPUs and AMD Instinct MI430X GPUs, is expected for delivery in 2028 with user operations in 2029.
AMD (NASDAQ: AMD) announced completion of the divestiture of the ZT Systems U.S.-headquartered data center infrastructure manufacturing business to Sanmina on October 27, 2025. AMD retains ZT Systems’ rack-scale AI solutions design and customer enablement teams to accelerate quality and time-to-deployment for cloud AI customers.
Sanmina becomes a preferred new product introduction (NPI) manufacturing partner for AMD cloud rack and cluster-scale AI solutions, reinforcing AMD’s ecosystem of ODM and OEM partners and aiming to speed manufacturing and deployment for cloud customers.
Sanmina (NASDAQ: SANM) completed the acquisition of the ZT Systems data center infrastructure manufacturing business from AMD (NASDAQ: AMD) on Oct 27, 2025. The deal adds large, state-of-the-art manufacturing facilities in New Jersey, Texas and an operation in the Netherlands, including high-power and advanced liquid-cooling capabilities.
Sanmina and AMD also agreed a strategic partnership naming Sanmina a U.S.-based preferred new product introduction (NPI) manufacturing partner to accelerate AMD cloud rack and cluster-scale AI deployments. A conference call to discuss the announcement and Sanmina fiscal Q4/2025 results is scheduled for Nov 3, 2025 at 5:00 p.m. ET.
AMD (NASDAQ: AMD) announced that senior executives will present at two investor conferences in December 2025 and that presentations will be available via live webcast and replay on the AMD Investor Relations website.
Key details: Dr. Lisa Su, chair and chief executive officer, will present at UBS 2025 Global Technology and AI Conference on Wednesday, Dec. 3, 2025. Jean Hu, executive vice president and chief financial officer, will present at Barclay’s 2025 Global Technology Conference on Wednesday, Dec. 10, 2025. Webcasts and replays are accessible at ir.amd.com.
AMD (NASDAQ: AMD) unveiled its Helios rack-scale reference platform publicly on Oct 14, 2025 at the OCP Global Summit. Built to the new Open Rack Wide (ORW) specification introduced by Meta, Helios combines AMD Instinct GPUs, EPYC CPUs, and AMD Pensando networking to target next-generation AI workloads.
Helios is a double-wide, ORW-based rack designed for gigawatt-scale data centers with quick-disconnect liquid cooling, improved serviceability, and standards-based Ethernet for multi-path resiliency. As an open reference design, it aims to help OEMs, ODMs, and hyperscalers adopt and customize scalable AI and HPC infrastructure more quickly.
Oracle (NYSE:ORCL) and AMD announced an expanded collaboration to deliver publicly available AI superclusters on Oracle Cloud Infrastructure starting in calendar Q3 2026.
The initial deployment will use an AMD "Helios" rack design with an initial 50,000 AMD Instinct MI450 Series GPUs, Helios 72-GPU liquid-cooled racks, next-gen EPYC "Venice" CPUs, and AMD Pensando "Vulcano" DPUs. Each MI450 GPU offers up to 432 GB HBM4 and 20 TB/s memory bandwidth. The zettascale OCI Supercluster can scale to 131,072 GPUs. Features include UALink/UALoE fabric, ROCm software stack, fine-grained partitioning, and up to three 800 Gbps AI-NICs per GPU.
AMD (NASDAQ: AMD) and OpenAI announced a multi‑year, multi‑generation agreement for OpenAI to deploy 6 gigawatts of AMD GPUs. The initial 1 gigawatt deployment of AMD Instinct MI450 series GPUs is slated to begin in 2H 2026. AMD issued OpenAI a warrant for up to 160 million shares that vests by deployment and performance milestones, including share‑price targets. AMD management characterized the deal as driving “tens of billions of dollars” in revenue and being accretive to non‑GAAP EPS. AMD will host a teleconference today and will report fiscal Q3 2025 results on Nov 4, 2025.