Intel talks about 14, 10, 7 and 5nm production. They will have 10nm production in 2015. They talk about their new WiFi chip and the potential of the Internet of Things. They talk about their building momentum in smartphones, tablets and all of mobility including an LTE roll-out in the US in 2013.
And ARM talks about, um, what? How the foundries can't even make enough 28nm product to satisfy demand. And what about 20nm? We hardly hear anything about that these days. We do hear about Apple's "new" 32nm processor. Yeah, 32nm production. Now that's exciting...
Intel researchers think they have cracked the 10nm manufacturing problem, paving the way for advanced chips that consume very little power.
[Meanwhile, ARM still hasn't cracked the problems that will allow them to make 20nm chips...]
The chip company's Ivy Bridge and Haswell chips are due to be built on its 22nm tri-gate process. After that, the company will move to 14nm and is expecting to start making chips with that process in late 2013 or early 2014. On Wednesday at the Intel Developer Forum, the company revealed that it thinks it knows how to build to 10nm as well.
Intel is exploring a spread of technologies for manufacturing new chips. Image: Jack Clark
The process a chip is made to dictates the density with which transistors can be packed together and the efficiency with which they use power, so new manufacturing methods at finer detail levels allow for better chips.
"The 14nm technology is in full development mode now and on track for full production readiness at the end of next year," Mark Bohr, director of process architecture and integration for Intel's technology manufacturing group, said. "Right now I'm spending my time personally on 10nm pathfinding and it looks like we have a solution there."
The 10nm solution may rely on a number of experimental technologies (pictured) potentially based around photonics, graphene, materials synthesis, dense memory, nanowires, extreme ultra violet lithography (EUV) and updated tri-gate transistors, Bohr said.
When it comes to how the 10nm chips will be manufactured, Intel has an immersion lithography method that works, though it would prefer to use EUV.
"I'd like to have EUV for 10, but I can't bet that it would be ready in time," Bohr said, hinting at the difficulties in using this method. EUV has much higher costs than immersion lithography.
Intel's research group are also exploring technologies for 7nm and 5nm solutions, though these are a very long way off as 10nm is not expected to go into production qualification until 2015.
Bohr capped off his talk by noting that Intel, unlike chip rivals AMD and ARM, owns and operates its manufacturing facilities.
"Yes, [process development] requires huge investments, but it also provides huge economic advantages," he said.
ARM licenses its chips and they are predominantly made in foundries operated by Samsung, the Taiwan Semiconductor Manufacturing Company, and GlobalFoundries. AMD uses GlobalFoundries. Intel has a three-year lead over these companies in terms of process, and with the news on Wednesday it looks like it could maintain that far into the future.