ASRock Releases Creator Radeon AI Pro R9700 GPU
The new 32GB workstation card leverages AMD Navi 48 architecture to deliver dedicated hardware acceleration for AI tasks.
Updated on Oct. 26, 2026 in Semiconductors
Live Poll
Is now a good time to invest in high-end workstation hardware for AI inference?
ASRock has launched the Creator Radeon AI Pro R9700, a new workstation-grade video card powered by AMD's Navi 48 GPU architecture. This card is officially released and available for professional computing applications requiring local AI acceleration.
Why it matters
The card offers a dedicated workstation solution for high-throughput matrix operations, integrating specific hardware support for AI acceleration in professional environments.
The card features 32GB of GDDR6 memory providing 640GB per second of bandwidth. It utilizes 64 compute units paired with 128 dedicated AI accelerator cores to hit a peak clock speed of 2.92GHz within a 300 Watt power envelope.
The players
ASRock
A hardware manufacturer specializing in motherboard and graphics card production for both consumer and enterprise markets.
AMD
A semiconductor designer known for its x86 CPUs and Navi-based GPU architectures used in workstation and gaming hardware.
The details
The R9700 uses Wave Matrix Multiply-Accumulate, a technique that allows the GPU to perform dense matrix operations directly within the compute units for faster neural network processing. To manage the thermal load of the 300-watt design, the card employs a vapor chamber cooling system and Honeywell PTM7950, a phase-change thermal interface material that transitions from solid to liquid to fill microscopic gaps between the chip and heatsink. For data integrity, the card supports soft ECC (Error Correction Code) memory error correction specifically when operating under Linux environments.
Timeline
The card was released on October 26, 2026.
The architecture was highlighted during the 2025 Advancing AI event.
The Tech Race
The R9700 serves as a new implementation of the AMD Navi 48 GPU architecture, expanding the professional workstation lineup. This release positions the card within the broader market of specialized AI accelerators that prioritize matrix-math throughput over general graphics rendering.
This card is designed for professional workstations requiring high-bandwidth memory for local AI workflows, utilizing a standard dual-slot blower configuration. Users requiring Linux-based ECC support or DisplayPort 2.1a multi-monitor output can integrate the card into existing chassis designs.
The takeaway
The R9700 brings specialized matrix acceleration to the workstation market, reflecting a broader trend of tailoring GPU silicon for specific AI math tasks. Watch for official pricing and benchmark comparisons against competitor workstation cards to gauge its real-world performance value.
Further reading
For broader trends in enterprise hardware, explore the latest developments in Semiconductors.
Source note: This article includes information reported by ServeTheHome.
Live Poll
Is now a good time to invest in high-end workstation hardware for AI inference?
