AMD today announced that its next-generation AMD EPYC processor, codenamed “Venice,” has entered production ramp on TSMC’s advanced 2nm process technology in Taiwan. The company also confirmed plans to expand future production at TSMC’s fabrication facility in Arizona, further strengthening its global manufacturing footprint for high-performance computing and AI infrastructure.
The milestone marks a major advancement in AMD’s data center CPU roadmap and positions “Venice” as the industry’s first high-performance computing product to move into production using TSMC’s 2nm process technology. The processor has been developed to address the increasing performance, efficiency, and scalability requirements of cloud computing, enterprise workloads, and next-generation artificial intelligence applications.
Dr. Lisa Su, Chair and CEO of AMD, said the launch represents an important step in accelerating AI infrastructure innovation as organisations increasingly deploy larger and more complex AI models. “As AI workloads continue to scale rapidly, customers require platforms capable of delivering high performance, efficiency, and faster deployment cycles. Our partnership with TSMC enables AMD to bring advanced compute technologies to market with the scale and speed needed to support this transformation,” she said.
AMD highlighted the growing importance of CPUs within AI infrastructure as enterprises expand beyond traditional AI training and inference toward more sophisticated agentic AI workloads. In modern data centers, CPUs play a critical role in managing data movement, networking, storage, security, and system orchestration across large-scale computing environments.
The company stated that the “Venice” production ramp reflects increasing market demand for AMD EPYC processors across cloud, enterprise, high-performance computing, and AI deployments. The initiative also aligns with AMD’s broader strategy to diversify advanced semiconductor manufacturing across multiple geographies.
AMD also announced plans to expand its use of TSMC’s 2nm technology across future EPYC processors, including the upcoming sixth-generation “Verano” platform, which is expected to deliver improvements in performance-per-dollar-per-watt efficiency. The roadmap includes support for advanced memory technologies such as LPDDR to address increasingly power-constrained workloads.
The AMD and TSMC partnership also extends to advanced packaging technologies including SoIC-X and CoWoS-L, which are being used across AMD’s broader AI and data center portfolio to deliver highly integrated compute platforms designed for next-generation infrastructure demands. AMD develops high-performance and adaptive computing technologies powering cloud computing, AI infrastructure, gaming, embedded systems, and enterprise applications worldwide.
For over five decades, AMD has been at the forefront of innovation in high-performance computing, graphics, and visualization. Its products are relied upon by billions of people worldwide, from leading global enterprises to groundbreaking scientific institutions, shaping the way we live, work, and innovate. AMD’s team remains committed to developing leadership products that push the limits of possibility.
