AMD has officially announced the launch of its new workstation graphics card, the Radeon AI PRO R9700, targeting the growing needs of artificial intelligence professionals and developers. The new GPU enters the market at a price point of $1299, offering a powerful solution for local AI inference, development, and other memory-intensive tasks.
The Radeon AI PRO R9700 is built on the latest AMD RDNA™ 4 architecture and is equipped with 32 GB of GDDR6 video memory, designed to handle large datasets and complex AI models directly on the desktop.
Key specifications of the new card include 128 dedicated AI Accelerators, 64 Ray Accelerators, and 64 Compute Units. This hardware configuration delivers significant performance, with the card boasting up to 47.8 TFLOPs of peak single-precision (FP32) performance. For AI-specific workloads, it offers up to 766 TOPs for 8-bit precision (INT8) operations and 1531 TOPs for 4-bit precision (INT4) operations, both with structured sparsity.
“The Radeon AI PRO R9700 is engineered to be a powerhouse for AI professionals and creators,” said a spokesperson for AMD. “With 32 GB of memory and our new RDNA 4 architecture, it provides the great capabilities needed for demanding AI development and content creation workflows, right from the workstation.”
The card supports the latest DisplayPort™ 2.1a standard and has a Total Board Power (TBP) of 300 W, with AMD recommending a minimum power supply of 750 W for stable operation. The Radeon AI PRO R9700 is positioned to compete with other high-end professional GPUs, offering a compelling blend of memory capacity and compute performance for the rapidly expanding field of local AI development.
