Array
(
[content] =>
[params] => Array
(
[0] => /forum/threads/intel%E2%80%99s-ceo-%E2%80%98we-are-not-in-the-top-10%E2%80%99-of-leading-chip-companies.23162/page-3
)
[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] => 2021770
[XFI] => 1050270
)
[wordpress] => /var/www/html
)
Given the capital Intel used to have, I believe it should have been able to compete with Nvidia. It just needed the right person to lead that effort. Once again, this reflects PG's management style—particularly in his hiring decisions.I agree with your capital intensity part but the strategy is always long duration and high risk Nvidia took a risk they were maturing CUDA they didn't do it overnight
i agree
It's not possible.Given the capital Intel used to have, I believe it should have been able to compete with Nvidia. It just needed the right person to lead that effort. Once again, this reflects PG's management style—particularly in his hiring decisions.
If Intel truly intended to compete with Nvidia, it should have chosen someone with at least the same calibre as Jensen to lead the efforts.
It's not possible.
Intel does not have a corporate culture that settles for No2 as you need to wait for a long time as AMD did.
Foundry is very capital intensive something AMD doesn't need to worry about.
If AMD can win some shares with their ROCm, Intel certainly could do it when it had enormous capitals than all its peers.
Hence, in my opinion, focusing on AI would be easier than focusing on the foundry business. As CEO, he should prioritize maximizing shareholder value. The AI path is not only easier but also offers greater potential rewards.Foundry is very capital intensive something AMD doesn't need to worry about.
Foundry leads to AI as well they can print their own AI chips plus others as well use foundry as a leverage in AI space in inference if you think you can't beat Nvidia in training.Hence, in my opinion, focusing on AI would be easier than focusing on the foundry business. As CEO, he should prioritize maximizing shareholder value. The AI path is not only easier but also offers greater potential rewards.
At the moment, there are not many clients for 18A. Will they use 18A for Jaguar Shores?Foundry leads to AI as well they can print their own AI chips plus others as well use foundry as a leverage in AI space in inference if you think you can't beat Nvidia in training.
It's most likely 18AP/14A from what I know the reason for TSMC doing GPGPU was because Intel didn't have a node optimized for this stuff.At the moment, there are not many clients for 18A. Will they use 18A for Jaguar Shores?
Pat's biggest design weakness is that he hates anything he perceives as "accelerators", which are really just specialized hardware designs, unless they are attached to x86 CPUs. Examples: Larrabee and the various "engines" integrated into Xeons. Pat's had this view for at least 25 years.Hence, in my opinion, focusing on AI would be easier than focusing on the foundry business. As CEO, he should prioritize maximizing shareholder value. The AI path is not only easier but also offers greater potential rewards.
The excessive fab expansion before there was any form of customer commitment was absolutely a mistake. Gelsinger repeated one of the key mistakes in Intel's previous foundry efforts. He made the assumption that fabless companies would tolerate almost anything to have the latest, greatest technology. In reality, they expect ease of use at least in the same ballpark as TSMC and they need it earlier than the 18A PDK was provided.He chose a high capital intensity, long-duration, high risk, and low-return strategy for Intel. Such decisions resemble those made by governments. With that amount of capital, Intel could have competed with Nvidia—and probably with more visibility. Judging from the current situation, 18A is mainly used for internal products, which is clearly different from what PG initially envisioned. This raises the question of whether 18A could have been done more economically if he had foreseen this scenario.
I read a post by a former Intel employee last year, who argued that the fabs Intel really needed were the ones in Ireland and Arizona. That assessment seems quite accurate given the current situation. If PG had been more prudent, I believe things could have turned out very differently at least in terms of financials.