Arthur Hanson
Well-known member
Is there any way Intel can recover even part of their past glory or is it to late?
Can Intel recover even part of their past dominance?
- Thread starter Arthur Hanson
- Start date
MKWVentures
Moderator
you dont recover glory. The gain new glory.
Intel was a dominant execution engine in the past with no competitor who could succeed. those days are gone.
yes the brand gives them an advantage.... but that brand has new characteristics now.
just an opinion.
Intel was a dominant execution engine in the past with no competitor who could succeed. those days are gone.
yes the brand gives them an advantage.... but that brand has new characteristics now.
just an opinion.
Intel is still dominant in business computers. I suspect this is due to decades of contracts embedding them into long term work; however, that wont last forever.
Intel's issue is going to continue to be cost and profit. It doesn't do much good to have the majority of the market share if you are bleeding money to do it.
AMD seems content (at present) to continue the march to gobble up as much of a % of the really profitable segments (DC, HEDT) as they can before putting the full court press on OEM (where they will be forced to sell at a much lower margin).
For Intel to return to its days of glory, it must find a way to make a profit. Every layoff depletes Intel's future ability to dominate.
I personally believe that they will need to re-focus their efforts on design and spin off the foundry. Intel can't survive on their own chips alone in foundry. The equipment has gotten too expensive.
Intel's issue is going to continue to be cost and profit. It doesn't do much good to have the majority of the market share if you are bleeding money to do it.
AMD seems content (at present) to continue the march to gobble up as much of a % of the really profitable segments (DC, HEDT) as they can before putting the full court press on OEM (where they will be forced to sell at a much lower margin).
For Intel to return to its days of glory, it must find a way to make a profit. Every layoff depletes Intel's future ability to dominate.
I personally believe that they will need to re-focus their efforts on design and spin off the foundry. Intel can't survive on their own chips alone in foundry. The equipment has gotten too expensive.
siliconbruh999
Well-known member
You know that Foundry is what's keeping them market share and IFS alone is not the only problem design is as well and AMD is benifiting off of Intel's Software Efforts.
AMD is still getting freebie from Intel they are not contributing meaningful to x86.
Not even x86-64 ?
siliconbruh999
Well-known member
i meant the software part AMD is very dependent on Intel Software Contribution there are numerous article where Intel doing OSS stuff is helping AMD as well but AMD doesn't contribute much in the kernel that benefit both
Xebec
Well-known member
I think AMD is even worse than this -- by giving a strong x86 design to China, they locked out both Intel and AMD from a large (or complete) portion of the Chinese market long term. Talk about a poison pill for your competitors growth..
AMD–Chinese joint venture - Wikipedia
AMD has received permission from the US Department of Defense and Department of Commerce to export the Zen 1 core design to China.<a href="https://en.wikipedia.org/wiki/AMD–Chinese_joint_venture#cite_note-CutressWilson-7"><span>[</span>7<span>]</span></a> Due to legal restrictions AMD has set up multiple companies to allow licensing of x86technology to China
..
siliconbruh999
Well-known member
Yeah this is not good for both of them but it was a necessary evil for AMD to be able to sell in China l.
Probably not. Lip-Bu will transform Intel but they will no longer be as dominant. They will certainly be relevant and innovative but not number #1. There is just too much competition from all sides and too much money to be had. Nvidia is the new Intel.
"For the first time in years, AMD's unit share of all x86 client and server CPUs shipped exceeded 25% and now stands at 25.6%, up from 24.2% in the prior quarter and up from 24% in the same quarter a year ago. Intel still commands 74.4%, but it lost some share in certain segments, allowing AMD to hit an important milestone amid a market shrink."
Arthur Hanson
Well-known member
Any thoughts on what the next advance will win the race and what companies stand the best chance of winning the race?
Good question.
Personally I think system companies will continue to make their own silicon. It really is a competitive advantage if you do it right. Apple clearly does it right and now the hyper scalars (Google and Amazon) and car companies (Tesla) are doing it right plus domain specific specific (AI) chips. That QCOM, MediaTek, Broadcom etc... will continue to gain market share that x86 once dominated.
AMD and Intel will continue to thrive but not at the market shares of the past. There really is room for everyone but no one wants a monopoly in the chips race, absolutely.
MKWVentures
Moderator
So do google and amazon make their own chips. Or do they give Inputs to Broadcom who tapeout their chips to TSMC? Historically when I worked on HDD/SSD controllers, the SSD company provided some custom logic but 70%+ was from Broadcom/LSI/Marvell.
Side note: If they have a ASIC with HBM, I assume Google/Amazon buy the HBM and then it is added to the interposer at TSMC?
thoughts?
Back to the question: I think companies that allow systems companies to design chips (Broadcom/ASIC vendors) and TSMC are the long term winners. Memory companies too IF they can keep constrained for the next 10 years
They mostly do their own chips.
In my example systems companies are like Apple, Tesla, Google, etc.... Fabless Systems Companies where silicon is just one piece of their product.
Broadcom is an ASIC vendor and a fabless chip company. The curse of the ASIC business is that the customers work with ASIC companies first then do it themselves so they are training their replacement.
Based on past experience, low power consumption will one day become a core technological competitiveness
siliconbruh999
Well-known member
Node shrink will continue to matter than cause while density benefits have been lowered but Power and performance still continue to scale
the absolute progress will continue for sure, but the major factor to drive Moore's type scaling has been shifting to both bigger chip(s) (hence bigger system) through adv packaging. Perhaps industry needs to pay more attention to shrinking further Vdd, with different architecture of devices. IMEC has set target of 0.25v in their long term research I believe
siliconbruh999
Well-known member
I think Bigger Chips will be challenging cause High-NA will commercialize sooner or later which would cut your die size so it's better to have decentish die and advance packing