“We are shipping in low volume and yields are improving, though the rate of improvement is slower than we anticipated,” said Brian Krzanich, Intel’s chief executive, in a conference call with financial analysts. “We understand the yield issues and have defined improvements for them, but they will take time to implement and qualify.”
Intel became the world’s largest maker of computer processors by following the tenets of Moore’s Law, which says that the number of transistors that can be placed on a single silicon chip doubles roughly every two years. But the company has fallen behind the prescribed pace, with every new generation of chips separated by about three years.
The 10nm problems are the latest setback for the Santa Clara, California-based company, which has long used manufacturing prowess to stay ahead of rivals like Taiwan Semiconductor Manufacturing Corporation. Intel had said it would deliver the 10nm process in the first half of 2016, but it continues to extend the due date.
Under pressure from customers and investors, Intel has also been forced to update its tick-tock development model. For almost a decade, it has followed up every new process technology with a new processor architecture, which is reflected by an “+” in the node name. But in 2016, Intel added another step to the model, in which it optimizes the architecture.
Currently, Intel makes chips based on 14nm+ technology. But in the second half of the year, the company plans to release two new lines of chips likely based on 14nm++ technology. The product lines are Whiskey Lake for personal computers and Cascade Lake for data centers. That technology will provide 70 percent more performance than the original 14nm process.