New RDNA™ 4 architecture GPU brings scalable, cost-effective AI acceleration to desktops worldwide
AMD has announced that its new Radeon™ AI PRO R9700 workstation GPU will be available globally beginning October 27 at an MSRP of USD 1,299. Designed to meet the growing computational demands of developers, researchers, and content creators, the R9700 delivers scalable AI performance and data-center-class efficiency directly to the desktop.
Powered by the advanced AMD RDNA™ 4 architecture, the Radeon AI PRO R9700 enables medium to large AI and machine learning models to run locally on workstations. The GPU is optimized for both Linux and Windows environments and integrates with AMD ROCm™, the company’s open software ecosystem for high-performance computing. This allows users to accelerate large language models (LLMs), diffusion models, and other compute-intensive AI workloads with higher efficiency and flexibility.
“With the Radeon AI PRO R9700, AMD brings data-center-class AI performance to the desktop.”
— David McAfee, Corporate VP & GM, Radeon Graphics, AMD
Key features include:
- 32GB of GDDR6 memory, enabling seamless handling of large AI projects
- Multi-GPU scalability through ROCm for parallel training and inference
- PCIe Gen 5 support for faster data throughput and reduced latency
“With the Radeon AI PRO R9700, AMD brings data-center-class AI performance to the desktop, helping professionals train, tune, and deploy large models faster,” said David McAfee, Corporate Vice President and General Manager, Radeon Graphics at AMD. “We’re enabling researchers, engineers, and creators to innovate without the infrastructure limitations of traditional setups.”
The launch underscores AMD’s commitment to democratizing access to advanced AI and GPU computing. By merging workstation affordability with enterprise-grade power, the Radeon AI PRO R9700 bridges the gap between local compute flexibility and cloud-scale AI acceleration, setting a new standard for professional AI development.
