It looks like foundry issues are back to slow down the launch of NVIDIA's next generation high-end GPU, codenamed Kepler. The delay may push Kepler's launch to 2012, it was expected to launch by late 2011. The 28 nanometer silicon fabrication process at TSMC, a principal foundry partner of NVIDIA, is producing unsatisfactory yields. Add to that, Kepler's performance is lower than expected.

TSMC's 28 nm process at Fab 15 facility has already seen delays, which have even shaped AMD's designs in a big way. AMD had originally planned to design high-end VLIW4 chips on the 32 nm process at TSMC, but later decided to wait for the 28 nm process, leading to plans of 32 nm GPUs being scrapped by both GPU designers. TSMC was supposed to be in a position take orders of high-end 28 nm GPUs by Q4 2011, and was set to start pilot production for its 20nm process technology in the third quarter of 2012.

TSMC's major client apart from GPU vendors, that relies on new bulk manufacturing process, Qualcomm, hasn't yet adjusted its 28nm process schedule and is set to launch three new products, 8960. 8270 and 8260A using dual-core Krait architecture in the fourth quarter of 2011.

AMD's first 28 nm GPUs, under the "Southern Islands" family are on course for 1H 2012. Around the same time, AMD will also release APUs based on the 28 nm process, codenamed "Krishna" and "Wichita", that replace Ontraio and Zacate.