• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Intel Core Ultra 9 285k 'Arrow Lake' Discussion/News ("15th gen") on LGA-1851

I’m gobsmacked (I did not believe the earlier rumours about 20a, which had no official source to back them up). They are trying to describe this as a massive win. The only reason to do this, is that there’s a problem with 20a. Probably still too low yield.

Intel has never done this for a desktop CPU series before, I can’t see how it will benefit customers.

Official Intel announcement:

Personally, I don’t think the 9700X will seem like a bad option, compared to Intel’s 15th gen. Especially if running at 105w, where it will still likely use less power than Intel equivalents.

I can see why the CEO considers 18a to be so important now, that 20a isn’t going to be used for anything. They’ve been stuck on 10nm (no EUV) desktop CPUs now for a long time (since late 2021), and 18a desktop CPUs may not be ready until 2026 (when 18a was predicted to be produced in larger volume).
 
Last edited:
another article, similar content.

Bad news for AMD - since this will push up prices for TSMC nodes. Intel's probably happy to swallow the loss of margin on the flagship SKU's if it means a return to leadership in terms of performance and equal or better performance per watt compared to Zen5.
 
Intel is going to have to contend with TSMC 2nm based products in 2026, so I don’t think 18a is going to give them much advantage:

I suppose they’ll be back on track, at least?

TSMC is going to have 1.6nm class products with backside delivery at some point, though (probably 2027).

Maybe they will try to rush out 18a mobile /desktop CPUs in late 2025?
 
Last edited:
Bad news for AMD - since this will push up prices for TSMC nodes. Intel's probably happy to swallow the loss of margin on the flagship SKU's if it means a return to leadership in terms of performance and equal or better performance per watt compared to Zen5.

But surely worse news for Intel...

  • They'll spend at least as much as AMD for wafers, on probably a larger cpu (going by historical comparions) and potentially more per wafer if they want/need N3 to compete
  • They've spent $lots on developing 20A which is now in the bin
  • They've spent a decent chunk of R&D time/effort on either maintaining designs for both 20A and N3 or more likely a late stage switch (sure design libraries will do a lot of the heavy lifting but not all)
  • Same again on QA'ing chips from two 'different' manufacturers/designs.

Yes it hurts AMD but it hurts Intel more, and Intel aren't in a position to be throwing money away like that
 
It sounds like Arrow Lake won't actually be out any time soon then, unless they'd already changed direction a long time ago without telling anyone?
 
Internal node was supposed to have been for the lower SKUs (i3, i5?).
But luckily (well luck and having spent lots of money) they did dual source. Whether that means the lower end stuff will be very tight in supply? Who knows, but probably.

Internally I'm sure the decision was probably months ago.
 
Last edited:
..and all that raptorlake nonsense could have been avoided if theyd stuck to the meteorlake thing for desktop aswell.


but anyway...
decision made, waiting on 265k . there are two x86 cpu makers. i really do not want only one choice from one company. intel chippies have served me well.
each to their own. its a free world. the freedom to choose.
 
Last edited:
I think it’s true that Intel is betting the company’s future on their 18a process.

The obvious question is - why not just delay Arrow Lake, and try to improve 20a yields?
 
I think it’s true that Intel is betting the company’s future on their 18a process.

The obvious question is - why not just delay Arrow Lake, and try to improve 20a yields?
Maybe scrapping 20a and using TSMC takes the pressure off and gives them more time and resources to put into getting 18a right.
 
I hope that the 20a foundaries can be repurposed for the still in development 18a process, or the cost of this switch could be quite high.
 
I wonder if TSMC’s 3nm fab processes are even any better than ‘Intel 3’ or ‘Intel 4’?

Or, just higher yield?

I’m guessing that what we will get with Arrow Lake, will be very similar in (CPU) performance to Luna Lake, except the voltage and clocks will be higher for desktop CPUs…

Lion Cove cores are supposed to have an improved IPC of up to 14% (according to Intel), vs Meteor Lake CPUs, so the CPUs clocked over 5Ghz on 8-P cores should still be faster than the best 14th gen desktop CPUs…

Shouldn’t be difficult to run more efficiently than the 14700K and 14900KS, with air cooling:
 
Last edited:
Back
Top Bottom