Discussion Intel Nova Lake in H2-2026: Discussion Threads

[511]

Do you write / compile your own software? As otherwise don't count on seeing a lot of usage in initial years.

I do it for my hobby and fun.

I have been wondering myself if I would need more, but turns out I have modest needs and wouldn't have anything to slot there.

well 24 is a decentish amount of lanes excluding TB5 i can put two storage direct to my CPU alongside a x16 GPU

That's only if you are buying a new board for AM5 vs LGA 1954

 

[RTX]

What applications benefit from 48 cores vs 24 cores?

https://twitter.com/x/status/2021295727448817669

@Geddagod Did he delete something?

 

[511]

What applications benefit from 48 cores vs 24 cores?

Unironically Cinebench/Blender/Vray rendering applications in general

https://twitter.com/x/status/2021295727448817669

@Geddagod Did he delete something?

lol

 

[RTX]

https://twitter.com/x/status/2024508973093703964

Is Titan lake delayed? 14A in 2029? @Geddagod

 

[Geddagod]

What applications benefit from 48 cores vs 24 cores?

https://twitter.com/x/status/2021295727448817669

@Geddagod Did he delete something?

He was pointing out like a 10% lead or something small in nT for PTL vs some Apple sku (I think it was the base M5) and claimed Intel was destroying or shi**ing on Apple or something to that effect.
Just completely fanboying for Intel, as usual.

https://twitter.com/x/status/2024508973093703964

Is Titan lake delayed? 14A in 2029? @Geddagod

No clue.

 

[Thunder 57]

https://twitter.com/x/status/2024508973093703964

Is Titan lake delayed? 14A in 2029? @Geddagod

As if SiliconFly is accurate.

 

[Geddagod]

As if SiliconFly is accurate.

Yeah, I just like calling out his BS sometimes.
This dude has seen Meng's posts on twitter, there's no way pointing out that Intel originally claimed 14A HVM is in 28' and then more recently at the Cisco AI summit saying it's 29', is the most stupid thing he has ever heard.
I have no great love for semianalysis either, especially after I got egg on my face from this incident quoting him lol

But tons of Intel stock owners are up in arms about that Semianalysis tweet because it paints Intel negatively. Undeservedly up in arms, IMO.

 

[Momoka_]

As if SiliconFly is accurate.

SiliconFly's claim checks out, especially with the construction pause in Ohio caused by Intel's funding shortage.

 

[poke01]

SiliconFly's claim checks out, especially with the construction pause in Ohio caused by Intel's funding shortage.

for Sillicon Fly to actually make a good point, that shows how bad Dylans post was

 

[Geddagod]

SiliconFly's claim checks out, especially with the construction pause in Ohio caused by Intel's funding shortage.

Which claim?

for Sillicon Fly to actually make a good point, that shows how bad Dylans post was

No I completely disagree. Every node jump that Intel has committed recently has faced problems, and Intel 18A is pretty much a N3 class node, which is basically what Dylan's post was.
And yes, if LBT is talking about 14A in the video and not 14A-E, it would be a delay based on what Intel themselves said in previous earnings call. I urge anyone dubious of this to go back and check it for themselves if they want, LBT claimed risk production and volume production a year earlier before for 14A.
And other than saying it was 14A, LBT's comments weren't even specifically in the context of external customers or node sub improvements, so the context surrounding what he is saying doesn't imply it's 14A-E either.
So the best hope is that LBT misspoke at the conference.

 

[DKR]

So the best hope is that LBT misspoke at the conference.

Stacy Rasgon (Bernstein) said in his post Q4'25 earnings analyst note that Intel management confirmed that LBT misspoke about 14A timelines in the Q4'25 earning calls (i.e. risk production in '27 as mentioned in ER call is not correct). So 14A is risk production in '28 and volume in '29 is the current correct timeline.

https://twitter.com/x/status/2014654567171162419

Just for context this what was said in IFS Direct Connect 2025. Note that both 14A & 14A-E are clubbed together in the same timeline.