Late last year, reports indicated that AMD's next-generation GPU had completed its design, was manufactured using TSMC's N3P process, and had already been taped out. Rumors originally suggested a debut at COMPUTEX 2027, but as the DRAM supply crisis intensified over the past year, news emerged that AMD might delay the launch of its next-generation RDNA 5 architecture products, potentially pushing it to late 2027 or even early 2028, aligning more closely with NVIDIA's next-generation GPU release window.
According to Wccftech, insiders recently revealed that next-generation graphics cards based on the RDNA 5 architecture GPU will feature very high core clocks, with AMD targeting between 3.1GHz and 3.4GHz. Since the RDNA 5 architecture covers a broad range, including discrete GPUs, integrated graphics, and custom SoCs, it remains unclear whether this applies to specific products or all designs.
Among the RDNA 4 architecture products currently on the market, the Radeon RX 9070 XT boasts the highest core clock, with a boost frequency of 2.97GHz. For comparison, if an RDNA 5 architecture discrete graphics card reaches 3.1GHz, that would be a 4.4% increase, while 3.4GHz would represent a 14.5% jump, which is quite significant.
Of course, higher core clocks should not be viewed as an equivalent indicator of performance improvement, as GPU performance is determined by multiple factors, including compute unit design, overall architectural improvements, memory bandwidth and capacity, and power limits, with core clock being just one influencing factor.