“I bet the whole company on 18A.” This is what Pat Gelsinger, CEO of Intel, says in the most deadpan and matter-of-fact way imaginable. He is not frivolous. He's not kidding. According to Gelsinger, everything runs on Intel's 18A process.
Intel's CEO made the comments in a recent interview with TechTechPotato. And we can't help but notice that Gelsinger was even less equivocal this time than when he addressed the exact same topic late last year.
As we noted at the time, he was about to say he was putting the company on 18A. “Bet the whole company? I don’t know if I would do it that way,” he said in November. “But this is the biggest bet we’ve ever made as a company because it also puts an incredible strain on the company’s finances.”
Whether Gelsinger has decided to turn Intel's biggest bet yet into an entirely existential one out of conviction that the plan is making progress or out of desperation at impending oblivion, we'll all have to wait and see.
To recap the technical details, 18A is the production process considered number five in Intel's accelerated roadmap back to technology leadership. We already have Intel 7 in Alder Lake and Raptor Lake CPUs, and Intel 4 just launched late last year with the Meteor Lake chips. 20A is reportedly coming later this year with the Arrow Lake CPU family and then it's the turn of the 18A node, which will be introduced in 2025. Phew.
But why is 18A and not the other landmarks along the way so important? The answer lies, in part, in the advanced technology it promises. 18A provides backside power, or what Intel calls PowerVia. In very simple terms, this means that the transistors of a chip are supplied with power from below and not from above.
If you're wondering why this matters, Gelsinger explained it last year and it all boils down to the problems caused by passing power through the many layers of wiring and connections that sit on top of the transistors in a modern chip.
“If you look at a stack of metal and a modern process, the cutting edge technology might have fifteen to twenty layers of metal. Metal one, metal two…and.” [then] The transistors down here. So it's just an incredible skyscraper design. Well, the top level of metals is used almost exclusively for energy transfer, so now you need to take signals and weave them through this grid. And then you want big, fat metals and why do you want them fat? So you get good RC characteristics, no inductance, but low IR drop across these large chips.”
The bottom line is that the power lines going into the chip cause interference. The solution is backside power and for Gelsinger it is a “Hallelujah” moment for the chip industry.
“Now you don't have to deal with the overall topology and chip planning considerations, and you'll get better metal characterization because now I can make them really fat, really big, and exactly where I want them on the transistor. So that’s really quite powerful,” he said.
The other reason 18A is so important to Intel is that it is the main target node for its ambitious plans to become the world's second-largest customer manufacturer by 2030. TSMC is of course the world's largest manufacturer of chips for customers. Samsung is currently a distant second.
Gelsinger's plan calls for Intel Foundry Services (IFS) to take second place away from Samsung, and the perceived attractiveness of its 18A node is one reason Intel believes it can poach customers like TSMC and Samsung.
At the moment, it's difficult to assess how well Intel is progressing with its plans to return to form. Microsoft has emerged as a customer for its 18A node, along with smaller companies Ericsson, Siemens and a handful of others. But the clock is ticking on whether the company will be able to consistently serve these new customers and then win even more.
Within two years everything will be revealed and the future of not only Intel, but arguably the entire chip industry, depends on this revelation.
