Intel Claims 10nm is “Making Good Progress”
Once upon a time, Intel was the undisputed leader in the silicon foundry business. With the tick-tock strategy, a new process node rolled out every 2 years. However, since the start of 14nm, things have slowed down dramatically. 10nm has been even worse, with delays stretching nearly half a decade at this point. This week, more rumours started circulating about a potential cancellation of the 10nm node at Intel. In a rare move, the company is speaking out directly, denying that there are any plans to cancel 10nm.
The report alleged that Intel was in the process of cancelling the 10nm process node. The idea was that instead of sticking to the failing 10nm node, the company would move onto a new 7nm process instead. This strategy is not new as GlobalFoundries attempted a similar move, trying to go from 14nm direct to 7nm. The problem is why would 7nm be any easier than 10nm as 7nm would be expected to be tougher. The only exception is where the problems are specific to the implementation of 10nm and that is the hangup. Some reports indicate the problem is that Intel decided not to go with EUV for 10nm, instead opting for quad patterning. Unfortunately, that gamble has not paid off.
10nm Might Not Stick Around for Very Long
In a surprising move though, Intel has directly responded to the rumour. According to the company, things are on track for 10nm. At the very least, they are on track based on the last earnings call. This means that we will finally get 10nm sometime in 2019, likely well into the year. While Twitter is an uncommon mouthpiece, it is the official company account. We will likely hear a similar message for the upcoming earnings call as well. One problem is that Intel optimism in 10nm has not paid off.
Given the amount of work that has gone into 10nm, a cancellation at this point would be surprising. Even if yields are still terrible, by this point, enough working product would have built up for some kind of launch next year. Short of an unmitigated disaster, we should see some 10nm product in 2019. What is a likely possibility is that 10nm won’t be around for long, with perhaps a quick transition to 7nm to catch up to AMD’s TSMC 7nm CPUs.
Media reports published today that Intel is ending work on the 10nm process are untrue. We are making good progress on 10nm. Yields are improving consistent with the timeline we shared during our last earnings report.
— Intel News (@intelnews) October 22, 2018
