Intel foundry customers have successfully powered on 18A node engineering samples

Daniel Sims

Posts: 2,441   +74
Staff
Why it matters: The success of Intel's upcoming 18A process node is critical to the company's future. After refuting reports of abysmal yield rates, Intel recently confirmed that 18A has reached a crucial milestone in its effort to regain competitiveness against semiconductor rivals Samsung and TSMC.

Interim Intel co-CEO Michelle Johnston Holthaus announced that the first engineering samples of hardware manufactured with the company's 18A semiconductor node have been delivered to customers. Her comments aim to reassure industry observers that Intel's foundry business remains on track to compete with TSMC's and Samsung's 3nm and 2nm nodes starting next year.

At the Barclays Annual Global Technology Conference, Holthaus and co-CEO David Zinsner discussed Intel's upcoming Panther Lake processors, which will debut the 18A process node upon their expected launch in the second half of 2025. Holthaus revealed that eight foundry customers have powered on ES0 (likely "Engineering Sample 0") chips built on the 18A node, signaling significant progress compared to six months ago.

Intel released version 1.0 of the 18A process design kit in July, enabling customers to begin developing chips based on the node. In August, the company confirmed that internal samples of Panther Lake and Clearwater Forest processors, built on the 18A node, successfully powered on and booted Windows with satisfactory performance. The statements made at the Barclays event mark the first confirmation of 18A usage outside of Intel.

Also see: What's next for Intel: Split, sell, or shut down the fabs?

Semiconductors produced with the 18A node will integrate Intel's new RibbonFET gate-all-around technology and PowerVia power delivery architecture, both designed to improve performance-per-watt and power efficiency. Intel hopes these innovations will help it reenter the race for leadership against the most advanced nodes from TSMC and Samsung.

The successful launch of Panther Lake and the 18A node could provide a desperately needed win for Intel. The company has faced significant challenges recently, including the resignation of former CEO Pat Gelsinger following several disappointing quarters. Skepticism about Intel's foundry business persists, with even industry veterans like TSMC founder Morris Chang expressing doubts about its prospects.

Intel has also had to address concerns regarding 18A's yield rates, following reports suggesting that nine out of ten chips manufactured on the node were defective. However, when accounting for factors like die size, defects per square centimeter, and the types of products involved, the outlook appears more optimistic.

Panther Lake CPUs, targeted at laptops, are expected to feature up to six Cougar Cove performance cores and up to eight Skymount efficiency cores. These processors will also include integrated GPUs based on Intel's third-generation Xe graphics architecture, codenamed Celestial.

Permalink to story:

 
I really hope they recover, I bought the dip. Pat was a fool and bet too much on GAA and they had to adjust their nodes to FinFIT because GAA wasn't ready yet. This is where alot of the issues with the 13 and 14 series came from. Those nodes were meant for GAA. Mix that with just ramming as much power into the chips as possible and that lead to issues. Intel should have taken the performance hit and mixed it with a price drop.
 
Praying for them. Even though I never buy anything intel
I've had 2 Intel systems in my life. One was a 3770k and I recently got a fantastic deal on a used 12700 non-k with mobo and ram. Intel going under hurts everyone. Intel is the reason we got ryzen, it's why we got amazing tech like 3D vcache from AMD. I'll likely always be an AMD guy, but I really hope Intel makes a comeback
 
I also hope they come back. Users need the competition between manufacturers.
Yes, Intel will come back, there's no doubt about that. But I hope it takes them a long time to do so, just so that AMD has a chance to get a stronger foothold in the market. I would prefer a 50/50 or 60/40 market share split between the two companies. What Intel wants is a 90-95% share of the market like they had in the 90's. They were happy to push AMD to the edge of collapse/bankruptcy and given the chance they'd do the same thing again. Intel isn't looking for a healthy market where they have 55-65% share, they want it all.
 
" and booted Windows with satisfactory performance"

That's a high recommendation right there /s

Could it run anything , er no, but if managed to boot and didn't give a BSOD for at least 10 seconds
You know when corporates use weaselly words, it's not there yet
 
" and booted Windows with satisfactory performance"

That's a high recommendation right there /s

Could it run anything , er no, but if managed to boot and didn't give a BSOD for at least 10 seconds
You know when corporates use weaselly words, it's not there yet
It's better than it not working. That means they're getting SOMEWHERE at least with the node
 
The Western world needs Intel, and silicon manufacturing, outside of Taiwan so I really hope this works for them.
 
R&D in smaller nodes is terrible expensive. Billions of USD your talking about. AMD was clever - simply gave up and outsourced all that to TSMC. Only downside is how much TSMC can deliver for AMD exclusively.

They are getting their butts whipped by AMD on both enterprise and consumer level.
 
Praying for them. Even though I never buy anything intel

You don't need to pray, they have plenty of money + chip act money. Intel is too big to fail and won't go anywhere.

We don't need another company relying 100% on TSMC, so I hope Intel will get 18A working soon.
 
Back