Nova Lake to use TSMC N2P for all but Entry Configuration According to Moore's Law is Dead

[benb]

14:50 in the video

"There's been a lot of rumors out there that 18A isn't going well. And then I've seen the diehard Intel defenders say it's going great. Look, I've heard mixed things from people at Intel. Full disclosure. Some of them say it's going well. Some of them say they have no uh reliability that 18A is going to be good. But I'm going to use Occam's Razor here and say if Intel is only using 18A for the like non ultra like non-branded mega budget entry stuff, I don't think that tells me that 18A is going well at all. I mean, I would have thought they'd at least use it for the i3 budget configuration, but doesn't seem like it everybody."

 

[benb]

My take: If Intel Products won't back 18A in any meaningful way, it's another nail in the coffin of advanced node development at Intel in the USA. Sadly. But I understand why it's happening: Intel can transition to fab lite, (which is where they are today), eventually they will stop the huge expense of advanced node development, shed the fabs, and survive, while Samsung and TSMC can't do that. Intel has a relatively graceful survival option. The only flaw is it runs counter to National Security, which it absolutely does. But National Security is a separate topic from survival.

 

[Daniel Nenni]

My take: If Intel Products won't back 18A in any meaningful way, it's another nail in the coffin of advanced node development at Intel in the USA. Sadly. But I understand why it's happening: Intel can transition to fab lite, (which is where they are today), eventually they will stop the huge expense of advanced node development, shed the fabs, and survive, while Samsung and TSMC can't do that. Intel has a relatively graceful survival option. The only flaw is it runs counter to National Security, which it absolutely does. But National Security is a separate topic from survival.

Do you really think Intel has a chance against Nvidia and AMD as a design only house? Not to mention the dozens of start-ups gunning after Nvidia or the cloud companies making their own chips?

 

[blueone]

Do you really think Intel has a chance against Nvidia and AMD as a design only house? Not to mention the dozens of start-ups gunning after Nvidia or the cloud companies making their own chips?

I don't think so. I'd rather have two competitors than 200, and have the whole country cheering you on. I also can't think of one Intel chip design that I'd put in a leadership position.

IMO, Intel needs to figure out how to be a profitable competitive foundry. I don't think there's any other worthwhile future.

 

[benb]

Do you really think Intel has a chance against Nvidia and AMD as a design only house? Not to mention the dozens of start-ups gunning after Nvidia or the cloud companies making their own chips?

We're finding out that no, they can't. The Intel 7 stuff outsells the tile stuff which is all TSMC. People want the monolithic Intel products at the lower price points. Nevertheless, the Nova Lake strategy doubles down on TSMC. Time will tell I guess.

 

[Brady]

MLID in his typical fashion looks for the juicy take of this being more of “18A is bad” news. But what if we look at the whole game in context:

You have Intel still going full in on flagship mobile products in Panther Lake for end of 2025 early 2026, and AFAIK there is no news that Intel is outsourcing any datacenter CPUs. Is this just that 18A has a plan for fab capacity and ramping and the big client desktop chips weren’t a part of it? If 18A was in real trouble you would have already been long hearing that Diamond Rapids among others were taping out on TSMC as well.

Yeesh, it must be a pressure cooker in Intel right now. LBT needs to somehow convince whatever employees are left to put their heads down and focus and tune out the noise. Not sure that works when your coworkers are being let go left and right and the stock price is in free fall. At least shares have big upside now

 

[benb]

When will risk production end for Intel 18A Panther Lake chips?

Intel's risk production for the 18A process node, which will be used for the Panther Lake CPUs, is ongoing. While Intel has achieved the risk production milestone, volume production (HVM) is planned for 2026. The current focus is on improving yields of the 18A process to make production economically viable.

Here's a more detailed breakdown:

• Risk Production Started:
  Intel has announced that the 18A process node has entered risk production, which involves early, low-volume test manufacturing runs.
  
  
• Yield Challenges:
  Initial reports indicate that a small percentage of Panther Lake chips produced using the 18A process are meeting the required quality standards, highlighting yield issues.
  
  
• Economic Viability:
  For Intel to ramp up to high-volume production (HVM) and make the Panther Lake CPUs profitable, they need to significantly improve the yield of the 18A process.
  
  
• Volume Production (HVM) Planned for 2026:
  While risk production is underway, high-volume manufacturing is slated for 2026, according to the Futurum Group.
  
  
• Panther Lake Launch:
  Intel is targeting a launch of Panther Lake CPUs in late 2025, says TweakTown.
  
  
• Production Problems:
  Intel is facing challenges with the 18A process, which is impacting the production of Panther Lake CPUs, reports Reuters.

 

[siliconbruh999]

The Hub Die is 18A in the picture btw

 

[osnium]

MLID in his typical fashion looks for the juicy take of this being more of “18A is bad” news. But what if we look at the whole game in context:

You have Intel still going full in on flagship mobile products in Panther Lake for end of 2025 early 2026, and AFAIK there is no news that Intel is outsourcing any datacenter CPUs. Is this just that 18A has a plan for fab capacity and ramping and the big client desktop chips weren’t a part of it? If 18A was in real trouble you would have already been long hearing that Diamond Rapids among others were taping out on TSMC as well.

Yeesh, it must be a pressure cooker in Intel right now. LBT needs to somehow convince whatever employees are left to put their heads down and focus and tune out the noise. Not sure that works when your coworkers are being let go left and right and the stock price is in free fall. At least shares have big upside now

The benefit of having process R&D in Oregon is that there are not too many competing employers. If TSMC had built their fab in Oregon instead of Arizona, it would be a different story all together. Intel products has a larger issue with talent retention though.

 

[hist78]

Do you really think Intel has a chance against Nvidia and AMD as a design only house? Not to mention the dozens of start-ups gunning after Nvidia or the cloud companies making their own chips?

But if they stay together, they will drag each other down into the hole and risk Intel being sold off in pieces. Intel should seize the opportunity to split while it still has some technological, financial, and product leverage. The era of the advanced logic IDM is over.

 

[tomatoma]

Certainly, the high-end model is TSMC N2, and the one for laptops and the low-end is Intel 18A.
As expected of Core As expected of Core Ultra, Intel It's not that Intel 18A is not used in Core Ultra as expected.

 

[tomatoma]

Also, MLID is a liar, so it's best not to believe any information from Intel.

 

[essiah]

I don't think so. I'd rather have two competitors than 200, and have the whole country cheering you on. I also can't think of one Intel chip design that I'd put in a leadership position.

IMO, Intel needs to figure out how to be a profitable competitive foundry. I don't think there's any other worthwhile future.

I don't think intel knows how to be a profitable foundry.

 

[tomatoma]

In the first place, I don't think all SKUs are made by TSMC.

 

[tomatoma]

I don't think intel knows how to be a profitable foundry.

They are idiots

 

[siliconbruh999]

But if they stay together, they will drag each other down into the hole and risk Intel being sold off in pieces. Intel should seize the opportunity to split while it still has some technological, financial, and product leverage. The era of the advanced logic IDM is over.

I don't think intel knows how to be a profitable foundry.

Both are innovators dilemma's Apple was not a mobile Phone company Nvidia was not an AI company things need to be made only time can tell

 

[tooLongInEDA]

Do you really think Intel has a chance against Nvidia and AMD as a design only house? Not to mention the dozens of start-ups gunning after Nvidia or the cloud companies making their own chips?

This comment seems to imply two things:

1) Intel is currently uncompetitive as a design house vs the fabless companies
2) The "solution" to that problem is to double down and continue as an IDM

How can that possibly make sense ? Intel must operate as a foundry now to survive as an IDM. And it cannot do so if it withholds technologies from its customers (who are also competitors) in order to compensate for its [presumed in the comment] design weaknesses.

Some of the arguments being made about Intel are becoming increasingly circular. One person says that Intel products are carrying the fabs ... now this account says that the fabs need to carry the products.

If Intel design and products are weak, the solution to that is better design, innovation, product marketing and management.

 

[Xebec]

MLID has had a few good leaks, but also quite a few misses. Take this with a really big grain of salt.

He was wildly overoptimistic on Meteor Lake, Arrow Lake, and Zen 5. He is currently suggesting Zen 6 is aiming for 7 GHz (though he said he expects at least 6.4 GHz), and "double digit IPC increases over Zen 5", so I guess we'll see next year.

Also, Nova Lake on N2P doesn't necessarily mean 18A isn't performing or yielding at all, it could just mean that N2 is more cost effective overall. (scale + yields). Intel may also need TSMC packaging for what they're planning.

(P.S. I am personally disappointed if this comes out as a mainly N2P product. I would have liked to have seen 18A do something really good for desktop..)

 

[siliconbruh999]

Also, Nova Lake on N2P doesn't necessarily mean 18A isn't performing or yielding at all, it could just mean that N2 is more cost effective overall. (scale + yields). Intel may also need TSMC packaging for what they're planning.

(P.S. I am personally disappointed if this comes out as a mainly N2P product. I would have liked to have seen 18A do something really good for desktop..)

NVL is packed entirely at IFS as for choosing node ask Intel Products

 

[blueone]

How can that possibly make sense ? Intel must operate as a foundry now to survive as an IDM. And it cannot do so if it withholds technologies from its customers (who are also competitors) in order to compensate for its [presumed in the comment] design weaknesses.

You're making an assumption here that no one else advocating for Intel to become a successful foundry is proposing. Restricting foundry customers to N-1 technology, for example, is a recipe for failure.

Intel hasn't shown industry-leading innovation in chip design since they made the Pentium superscalar. Itanium, as misguided as it was, was designed by HP. Most of Intel's so-called leadership designs were barely competitive, often trailing, but achieved leadership due to a one or two generation fab process advantage. I think it's time Intel let the chip design innovation proceed. It's been a long time since that's been the case.