AMD Prepares "Gorgon Point" APU Refresh for 2026
AMD is set to advance its APU lineup with the introduction of the "Gorgon Point" generation, a refresh of the current "Strix Point" series, expected to launch in 2026. Recent benchmark leaks have provided an early look at the capabilities of this upcoming processor, specifically highlighting the AMD Ryzen AI 9 HX 470 model.
Key Specifications and Architecture
The "Gorgon Point" APU continues to utilize up to 12 cores and 24 threads, leveraging either the "Zen 5" or "Zen 5c" architecture. The integrated graphics are powered by the RDNA 3.5 architecture, while the neural processing unit (NPU) is based on XDNA 2 technology. The cache configuration includes 12×1 MB of L2 cache and 3×16 MB of L3 cache, supporting efficient data handling and improved multitasking performance.
Performance Enhancements
While the core architecture remains consistent with its predecessor, the "Gorgon Point" APU introduces several performance improvements. Notably, the CPU, integrated GPU, and NPU all receive targeted enhancements. According to a recent SiSoftware benchmark submission, the Ryzen AI 9 HX 470 achieves a boost clock of 5.25 GHz, a slight but meaningful increase over the previous generation's 5.1 GHz. This performance gain is achieved while maintaining a 28 W default TDP, ensuring energy efficiency alongside higher speeds.
AI and NPU Upgrades
One of the standout features of the "Gorgon Point" APU is its upgraded NPU, which now delivers over 55 AI TOPS (Tera Operations Per Second). AMD has enabled full NPU performance across all tiers of Ryzen AI SKUs built on this chip, providing enhanced AI capabilities for a wide range of applications. These improvements are designed to meet the growing demand for AI-driven workloads in modern computing environments.
Leaked Benchmarks and Industry Insights
Leaked partner slides and confidential presentations have offered additional insights into the "Gorgon Point" APU's performance. AMD's internal benchmarks indicate incremental yet impactful gains in both single-threaded and AI processing tasks compared to the existing "Strix Point" chips. These enhancements position the "Gorgon Point" generation as a compelling choice for users seeking advanced performance in both traditional and AI-accelerated workloads.
