Friday, March 7th 2025

Intel Panther Lake on Track for H2 2025 Launch, Company Exec Disregards Rumors of 18A Delays

Earlier in the week, online chatter pointed to a possible delay in the production of Panther Lake silicon. Well-known industry analyst—Ming-Chi Kuo—has kept tabs on the inner workings of several big semiconductor players. A previous insider tale revealed NVIDIA's allegedly revised "Blackwell" architecture roadmap. Kuo's latest insight focused on Intel and their 18A node process; rumored setbacks have reportedly pushed the launch of next-gen Panther Lake (PTL) mobile processors into 2026. Team Blue leadership has already reacted to the relatively fresh allegations—earlier in the week, John Pitzer sat down with Morgan Stanley Semiconductor Research's Joe Moore. During their conference fireside chat, Intel's Corporate Vice President of Investor Relations addressed recent internet whispers.

When asked about 18A being developed on schedule, Pitzer responded with: "yes, it is. I mean, I tend to wake up every morning trying to fish through rumors that are coming across on social media about Intel 18A. I want to be very clear. Panther Lake is on track to launch in the second half of this year. That launch date has not changed. We feel really good about the progress that we are making. In fact, if you look at where our yields are on Panther Lake today, they're actually slightly ahead at a similar point in time to Meteor Lake, if you look at the development process for Meteor Lake. I think a couple of weeks ago, there was a technical paper out that actually looked at our SRAM density on Intel 18A that compared well with TSMC's N2. Lots of different metrics you can compare technologies on. I think in general, we think about Intel 18A being an N3 type/N2 sort of comp with the external peers." Panther Lake is set to become the company's first product family that will utilize its own Foundry's 18A node process. Mid-way through February, we heard about the importance of PTL with Intel's portable gaming strategy.

Pitzer elaborated on the topic of production punctuality: "we're also on track, and we've announced this that we're going to tape out our first external design on Intel 18A in the first half of this year. So we feel good about the progress that we're making there. Now I'll remind you that we will launch Panther Lake in the second half of this year. In a similar vein, we launched Meteor Lake in the second half of 2023. It wasn't until 2024 that it became real volume. We launched Lunar Lake in the second half of last year, and this is the year where it becomes real volume. And as we've said on Panther Lake, we launched in the second half of this year, but it's really not until next year that we get to that volumes where it really starts to help improve the margin profile of the overall company...I think as we turn the page into next year, and we get Panther Lake out the second half of this year and then Nova Lake the next year that we're in a very strong position in that market." Later on in the talk, he reiterated that everything is going according to plan—one of Team Blue's key milestones is the: "launch of Panther Lake in the second half of the year, which I said we're absolutely on track to do."
Sources: Investing.com News, Wccftech, Seeking Alpha
Add your own comment

8 Comments on Intel Panther Lake on Track for H2 2025 Launch, Company Exec Disregards Rumors of 18A Delays

#1
Daven
Split off IFS
License x86 like ARM

Until then STFU Intel!
Posted on Reply
#2
dyonoctis
DavenSplit off IFS
License x86 like ARM

Until then STFU Intel!
Licensing x86 is useless, if AMD doesn't agree to license x64 as well. x86 won't run anything modern. x86 to talk about the modern CPUs is an abuse of language, those are x86-64 CPUs.
Posted on Reply
#3
john_
DavenSplit off IFS
License x86 like ARM

Until then STFU Intel!
Bad idea. I am with Pat in this one. Intel needs to keep it's fabs, especially if their 18A process works as expected, it could sustain their product lines for the next 5 years while also attracting customers.

An x86 in best case scenario will interest no one, in worst case scenario it could interest Nvidia and that could be bad news for Intel CPUs down the line. In any case AMD also has a saying about x86 today, so licensing the x86 architecture could be way more complicated. AMD would probably say no in licensing x86-64 to anybody who could become a competitor. Also with China turning all it's focus on RISC-V, i would expect the popularity of this architecture to expand greatly in the next 10 years.
Posted on Reply
#4
Panther_Seraphin
The foundries adopting EUV earlier than TSMC is a gamble and if it plays out correctly it will be big.

It wont give them a foundry lead in terms of density or node size but in THEORY it could both reduce costs due to less amount of times going through the patterns and in turn reduce failed dies due to the repeated exposures.

That would be a big thing when going against AMD/nVidia and being able to charge 5/10/15% less per die but still keeping the same profit margins.




18A and beyond just need to come good on the yield sides and Intel will be back in the game foundry wise vs TSMC. Then its the core designers to finally get off their ass and refresh the uArch with something competitive in a per core basis in both power and efficency.
Posted on Reply
#5
N/A
Scrap it already, just like the 20A. Time to move on. What is the point, when 14A is bound to replace it soon after.
Posted on Reply
#6
kondamin
N/AScrap it already, just like the 20A. Time to move on. What is the point, when 14A is bound to replace it soon after.
Considering they are fake nodes scrapping one if it works is as pointless as just using one.
Posted on Reply
#7
tfp
DavenSplit off IFS
License x86 like ARM

Until then STFU Intel!
Lol, this will really make them profitable.
Posted on Reply
#8
RUSerious
N/AScrap it already, just like the 20A. Time to move on. What is the point, when 14A is bound to replace it soon after.
Because foundries must implement each node in order to get to the next one. There is a long lead time development plan in place for node improvements. If it turns out that a node is terrible, then it's only used for low volume. The caveat is that screwing up a node isn't something a foundry wants to happen for than once in a very rare while (or, really, never - but it happens). Even terrible nodes teach the process engineers something about the next node.
Posted on Reply
Mar 28th, 2025 08:05 CDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts