Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/intel-cancels-its-mainstream-future-xeon.24020/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2030770
            [XFI] => 1060170
        )

    [wordpress] => /var/www/html
)

Intel Cancels its Mainstream Future Xeon

siliconbruh999

Well-known member
After weeks of speculation, we are breaking news here and on ServeTheHome that Intel has made a titanic roadmap shift, with an entire next-gen Intel Diamond Rapids Xeon line being removed from the roadmap. Here, we are going into a bit about the “why” we have heard from various partners and then the market impacts. If you saw this week’s AMD Financial Analyst Day 2025, and the claim that AMD had a path to 50% server market share, this is likely to help that cause.



We have removed Diamond Rapids 8CH from our roadmap. We’re simplifying the Diamond Rapids platform with a focus on 16 Channel processors and extending its benefits down the stack to support a range of unique customers and their use cases. (Source: Intel Spokesperson to STH)
 
Last edited:
Patrick has been covering servers for a while if this is true than this is Intel doing Sucide for the DC Business I got overexcited i am sorry
 
Last edited:
This seems like a very strange decision. Are Intel conceding the server market to AMD? I'm sure their margins have shrunk now. But what will they do instead in servers? Maybe sell their stockpiles of prior cpus?

Color me confused.

ETA now that I've read the report, it seems more nuanced than what I first thought. Still somewhat confused though.
 
Last edited:
"Extending it's benefits down the stack" tells me that they're just going to use the 16 channel version and 'down bin it' for the whole range. That must mean they don't see the cost difference of developing multiple solutions as worth it vs what is expected to sell.

- They can easily just disable memory channels to serve some other markets
- They expect demand to be drive primarily by AI applications - where bandwidth is everything.
- The performance uplift for 'lower channel versions' isn't good enough to make new SKUs; (IIRC - there was some wording by Intel that DR wasn't as competitive as they hoped?)
 
The 8 channel Xeons were the mainstream of Intel's server market, *including* AI servers that are offered by Dell/Lenovo, etc. The reason is because 12 channel is overkill for a lot of customers. You can't just "bin them down" for the same reason that you can't "bin down" i7 dies and use them for i3 products. Those products exist for a reason.
 
Its good that the roadmap changes and simpifications are now coming out. .... Ice lake is still selling. Sapphire and emerald are still highest volume sales. Sierra forest and granite are low volume and will ramp over the next 2 years. and it has been reported that Diamond rapids doesnt give Intel leadership. and yes Intel is losing share. Simplifying and not releasing ovelaping SKUs with previous generations seems very smart.

All this is before looking at the Cash flow 18A cost issues in Fab 52.

Good move by LBT (IMO). more to come
 
If what they want to do is processors segmentation by memory capacity/throughput, it's very smart choice. Intel already has a bunch of 8-ch processors(Sierra forest and Granite rapids, which does not use Intel 18A).
 
Back
Top