Intel Stuck With Ageing 14nm For Another Year
Samuel Wan / 2 years ago
Intel 10nm Launch Slips to Holiday 2019
Once the dominant semiconductor foundry, Intel has now fallen on tougher times. Intel first debuted their current mainstream 14nm process back in 2014. The plan was that the company would transition to 10nm sometime in 2016 or 2017. Unfortunately, those plans did not play out as planned for Chipzilla. According to the latest investor earnings calls, 10nm may be a lot further away than expected.
According to the Q2 2018 investor earnings call, 10nm products won’t really arrive till Holiday 2019. This means about another year before 14nm gets retired. This ageing process will have to hold the line as the mainstream process until likely October 2019. This is more than a year from now. This is a massive slippage from the original 2015 plans, delayed to 2016, and now to 2019. This may make 14nm the longest-leading process at Intel, a total of 5 years on the frontline.
Intel Delays Allows AMD to Catch Up Fast
To make sure there is something new for 2018, Intel is reportedly pushing out Whiskey Lake. This will slot in as an updated architecture for the 14nm process. To keep up the competition against a resurgent AMD, the company has been pumping up the core counts. For now, the company still holds a slight lead as their 14nm process is at least on par with the 12nm process AMD is using. Intel has also tweaked their 14nm process and improved from the original Broadwell launch in 2014.
Back in 2015 when the delays first started, the plans were for a late 2016 or 2017 launch. After starting risk production in 2016, the delays just got worse. Even with additional investment, things have not worked out well. Despite having a Cannon Lake mobile Core i3 10nm chip, it looks like yields are just so bad, a full launch is impossible this year. The only ray of sunshine is the possibility of earlier limited launches for enterprise and enthusiast markets.
However, these delays do point to major problems at Intel. Without their process node advantage, Chipzilla will find it harder to compete. In fact, AMD may even be ahead early in 2019. AMD is planning to launch 7nm chips with Zen 2 in 1H 2018. This should provide a process advantage against Chipzilla. It will be interesting to see if this will even the playing field and allow AMD to surge past Intel at last. Either way, it looks like Intel has squandered their massive process lead they’ve built up over the years.