Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/good-news-in-servers-for-qcom-arm-not-so-good-for-intel.7249/
        )

    [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] => 2021370
            [XFI] => 1050270
        )

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

Good news in servers for QCOM/ARM, not so good for Intel

I have seen many companies give public support for ARM servers over the years, FaceBook, Google, Oracle, Amazon, etc... Hopefully with their new server architecture it will actually happen. Intel needs some competition here. This is what ARM CTO said at the most recent ARMTechCon and I hope it is true for the greater good of the power grids of the world:

“I think that we are on the cusp now of having this plethora of ARM-based servers to actually transform what happens in that marketplace,” Muller said. “It is not just about servers, because when you look at it, it is about everything from network infrastructure to wireless access and it is where the ARM business model, with its partnerships, allows you to have very diverse pieces of silicon being put into very different products that scale all the way from scale out servers on one end through to very specialized, dedicated networking and wireless components on the other. Wireless infrastructure is a market segment where we are going to have 20 percent, 40 percent, 80 percent market share over the coming five years. So for us, it is very exciting to see ARM 64-bit technology pervading out from what is our traditional from sensors out to smartphones and really going out and pushing all the way into servers.”

View attachment 16479


I will be at the Linley Data Center Conference next week and will write what I learn....
 
Last edited:
Apparently Google is planning to publicly give its support for ARM-based QCOM servers to be used in its datacenters. This may only be for a fraction of those servers, but nonetheless a crack in the wall of the Intel server fortress. ARMH and QCOM stock are up on the news and INTC is down

Qualcomm/ARM outperform, Intel underperforms after Google server report - Alphabet Inc. (NASDAQ:GOOG) | Seeking Alpha

Google support is conditional. The original article states this: " It adds Google and Qualcomm have cooperated on design work, and that the former will commit to using the latter's chips if they meet performance goals."

So it is still work in progress. So far I have not seen any credible claims/reports about ARM server chips outperforming Intel chips. Then of course there is a price. But I believe I read that for server applications the chip price is not as critical as other factors (like performance per watt).
 
Like Google, others are also willing to lend support to ARM based servers and Qualcomm can definitely make a big dent there if performance criteria is met. Since Intel is a big existing giant in the server business, naturally it will take time for others to establish credibility. But it will be an interesting fight between Intel and others in server business, Intel must work on strategies to safeguard its forte.
 
With 99% of the server market, Intel can only lose share over time. That was the rap on Qualcomm, before MediaTek, Samsung, and others developed some 3G/4G modem chops. On the other hand, Intel has always been ruthless, and sometimes "questionable", in protecting market share. Maybe they'll just buy Qualcomm, and solve their current handset problems, and their future server threat, simultaneously?
 
Even if Intel doesn't buy Qualcomm (I don't think Qualcomm would join Intel either), if you see from market perspective, a collaboration between them would have been a great compliment for each other. Qualcomm to get into server and Intel to strengthen handsets. But does that happen in reality between arch rivals? Moreover their business model is very different, ARM being in the middle for Qualcomm. So, I don't think Intel will buy Qualcomm. Intel will do everything to keep their market share in servers. Qualcomm is on slippery ground though on handset business considering many other competitors there.
 
Maybe they'll just buy Qualcomm.

I vaguely remember some other big firm saying that somewhere in the lost pages of history ....

A collaboration between them would have been a great compliment for each other.

Been tried. Intel built 6 million chips for Qualcomm, then walked away - and left the door open for ARM.


To me, this is revisiting the telecom infrastructure boom right around the turn of the last century. What's a "server"? Intel does have a massive lead in application servers, but when you move into telecom infrastructure it's not that clear and there is a lot more competition. That is where Qualcomm and all the other ARM players are likely focusing.

It's not surprising that Google and others have some specific workloads where ARM chips could help. There is also another threat looming for Intel on that - open architectures, such as RISC-V. Again, if you step out of application servers (where software compatibility is still important) and into workload-optimized platforms, there could be some very interesting purpose-built solutions.
 
Back
Top