What would you do if you are the CEO of Intel?
- Thread starter hist78
- Start date
Neither group will have expert advice on semiconductor manufacturing or product design strategy. In my experience, the main thing these consulting companies offer is financial analysis, and perhaps some (IMO) clueless market analysis. If you want expert consulting on the semiconductor industry you need to hire semiconductor industry experts. How would a consulting company executive judge the expertise of a semi expert? By just reading their resume? Looking for advanced degrees? Impressive sounding titles? As the saying goes, it takes one to know one.
Last edited:
I was being a smart alec as I dont think they provide anything constructive , they are though a drain on businesses.
Meh, I used to work for Avanade (an Accenture sister company). They are not 'experts,' just a bunch of MBAs who charge you $1,000 per hour to tell you to lay off more people.
CC Wei in an interview:
"A company will perish because of the CEO's arrogance, so I often remind myself not to forget who I am," echoing the theme of the speech, "Be proactive and humble."
CC Wei said that when a person forgets who they are, they become unable to listen to other opinions. "The perspective becomes narrower, and then the CEO will inevitably lead the company into trouble, expanding recklessly." He believes that companies in a business crisis often turn to McKinsey & Company. By this stage, the company is on the verge of extinction. "The final step is to sit back and await the death."
hist78
Well-known member
Intel execs admit they were caught off guard by business downturn, triggering historic cuts.
www.oregonlive.com
Intel execs admit they were caught off guard by business downturn, triggering historic cuts
“We should have had better insights into where the business was going."
www.oregonlive.com
tooLongInEDA
Moderator
I understand that Intel is not exactly short of MBAs on its full time staff. Agree with both of you that they're more likely part of the problem than the solution.
mattyenright
New member
I would first spend the initial several weeks crying. Once I gathered enough emotional confidence I would subsequently change the company into a conglomerate separating the design, cloud and manufacturing areas into three distinct companies. Each with there own CEO and culture.
I would recommend the newly formed CEO of Intel manufacturing to form a research and development data sharing partnership with Samsung foundries. Reducing the cost of node development for both companies. Alongside form an additional department whose purpose is to construct equipment to automate further functionality of foundries to reduce the impact of high wages. Thereby improving the competitiveness relative to TSMC.
I would recommend the newly formed CEO of Intel manufacturing to form a research and development data sharing partnership with Samsung foundries. Reducing the cost of node development for both companies. Alongside form an additional department whose purpose is to construct equipment to automate further functionality of foundries to reduce the impact of high wages. Thereby improving the competitiveness relative to TSMC.
Everyone seems obsessed with killing off the working man it seems.
Today's events make me think Intel can't be saved unless it goes private with a new board.
Multiyear turnarounds are extremely hard in the best of times when your audience is focused on the quarterly.
The answer is “retire”. 
tooLongInEDA
Moderator
You prompted me to go back and re-read this thread right from the start and apply the low pass filter.
There are two contributions from March 2015 (over 9 years ago) that really foresaw the critical (in my view) issue when no one else (I include myself here) seemed to. Both @astilo and @Brianhayes pointed out that Intel ultimately needed to split design and manufacturing.
Some people argue this has effectively been done now. I don't buy that. It's not really separate until they are completely separate companies.
Of course, we'll run into all the current objections about Taiwan/national security/ the CHIPS Act that never existed in 2015 (plus the eternal "too big to fail") - and which had Intel got ahead of the curve and acted earlier might not have appeared.
hist78
Well-known member
This is the text that started at the very beginning of this thread. Sadly, there are more troubles added to the Intel's problem list since March 13, 2015.
"Intel is facing serious challenges in the recent years such as.
· Stagnant revenue
· Diminishing PC/Laptops market
· No strong mobile processor product offering
· Losing huge amount of money from mobile product line.
· Escalating costs of leading edge process and manufacturing cost
· High gross margin that might not be sustainable because Intel is not the dominant player in those new market that Intel is trying
· Well-funded and well positioned competitors
· Foundry business hasn't generated big impact and revenue
Above is not the complete list and you can add your own. What would you do if you are the CEO of Intel and if you have only two years to fix it before the board asking you to spend more time with your family?"
What would you do if you are the CEO of Intel?
Intel is facing serious challenges in the recent years such as: · Stagnant revenue · Diminishing PC/Laptops market · No strong mobile processor product offering · Losing huge amount of money from mobile product line. · Escalating costs of leading edge...
semiwiki.com
DaRAGingLunatic
Member
Well for starters, how about providing what it is their customers want.
Let’s take for instance AMD’s X3D cache tech.
AMD’s x3D cache has allowed those processors which are generally less performant in workloads than intel, to catch up and beat intel in gaming situations and other situations. Consumers want them for the performance gains and the consumer is paying more for them.
But Intel doesn’t want to do that as “It’s a small market, but we will do that in servers”.
Amd is also using that tech in servers as well. To be competitive in the server space, now Intel is finally going to do it. But only to server chips.
By not doing it for its other chips, intel isn’t competing and therefore there is no competition at all, so what will the consumer do? Buy the better, faster product.
It’s just like what Rory Reed did to AMD. “Our cpu’s are good enough”. That didn’t stop intel from producing faster, better cpu’s which consumers bought in droves, because they were better than the “good enough” cpu’s of AMD.
What does any consumer want? Better, faster products, at the lowest prices they can get them.
Intel needs to start competing asap. In all areas. I mean even look at Altera and what being bought by Intel has done to their revenue.
Why does it take Intel so long to build a Fab compared to TSMC? Why does it take Intel so long to ramp up production?
These need to be looked at.
Take a look at arrow lake chips. You have the cpu, the gpu, the iod, the npu.
A monolithic die has been cut into smaller bits and stuck together. The smaller sizes mean you can get more bits on a wafer and increase the use of a wafer, bringing increased profits. Does it actually save costs compared to ensuring that all the little bits are properly aligned and in place and can connect together and talk seamlessly and with low latencies, or does doing all that require extra manpower, extra thinking, extra logistics, extra interposer, extra robotics machinery, extra testing, extra sorting, extra electricity, yada yada yada. Is it better? Is it faster? Is it actually more competitive and is it actually worth it?
All the little parts are also produced on different nodes. Why is that? They want to maximise profits.
Intel needs some hard questions to ask itself. Like would it be better to get rid of the waste of space npu, and just use that space for a larger GPU that can also do the same thing the npu can do. What is it that my customers actually want. Would they prefer a larger gpu?
Knowing intel, they’d cut the NPU and just try maximise profits leaving the gpu the same size.
That’s all Intel does constantly. Not giving the customer what they want whilst aiming instead for profit.
They will produce a cpu capable of avx512, and then burn off that capability so the customer cant use it, because only the top teired highest cost product that people pay the most for will have it allowed and not burned off.
They do the same with server parts all the time.
I’ll tell you what it makes people do. It makes people hate the damn company
I would find out who has these stupid ideas and fire them.
Next, i would have Intel produce less parts.
You don’t need 100 different server parts.
Intel does need fabs though. Not only that but if it wants big customers, it needs very big fabs.
If it is not building brand new big fabs all the time, then it will fall behind.
I mean, Intel could very well stop fab production and just switch to TSMC. But then they will always be on the same node as the competition…
Let’s take for instance AMD’s X3D cache tech.
AMD’s x3D cache has allowed those processors which are generally less performant in workloads than intel, to catch up and beat intel in gaming situations and other situations. Consumers want them for the performance gains and the consumer is paying more for them.
But Intel doesn’t want to do that as “It’s a small market, but we will do that in servers”.
Amd is also using that tech in servers as well. To be competitive in the server space, now Intel is finally going to do it. But only to server chips.
By not doing it for its other chips, intel isn’t competing and therefore there is no competition at all, so what will the consumer do? Buy the better, faster product.
It’s just like what Rory Reed did to AMD. “Our cpu’s are good enough”. That didn’t stop intel from producing faster, better cpu’s which consumers bought in droves, because they were better than the “good enough” cpu’s of AMD.
What does any consumer want? Better, faster products, at the lowest prices they can get them.
Intel needs to start competing asap. In all areas. I mean even look at Altera and what being bought by Intel has done to their revenue.
Why does it take Intel so long to build a Fab compared to TSMC? Why does it take Intel so long to ramp up production?
These need to be looked at.
Take a look at arrow lake chips. You have the cpu, the gpu, the iod, the npu.
A monolithic die has been cut into smaller bits and stuck together. The smaller sizes mean you can get more bits on a wafer and increase the use of a wafer, bringing increased profits. Does it actually save costs compared to ensuring that all the little bits are properly aligned and in place and can connect together and talk seamlessly and with low latencies, or does doing all that require extra manpower, extra thinking, extra logistics, extra interposer, extra robotics machinery, extra testing, extra sorting, extra electricity, yada yada yada. Is it better? Is it faster? Is it actually more competitive and is it actually worth it?
All the little parts are also produced on different nodes. Why is that? They want to maximise profits.
Intel needs some hard questions to ask itself. Like would it be better to get rid of the waste of space npu, and just use that space for a larger GPU that can also do the same thing the npu can do. What is it that my customers actually want. Would they prefer a larger gpu?
Knowing intel, they’d cut the NPU and just try maximise profits leaving the gpu the same size.
That’s all Intel does constantly. Not giving the customer what they want whilst aiming instead for profit.
They will produce a cpu capable of avx512, and then burn off that capability so the customer cant use it, because only the top teired highest cost product that people pay the most for will have it allowed and not burned off.
They do the same with server parts all the time.
I’ll tell you what it makes people do. It makes people hate the damn company
I would find out who has these stupid ideas and fire them.
Next, i would have Intel produce less parts.
You don’t need 100 different server parts.
Intel does need fabs though. Not only that but if it wants big customers, it needs very big fabs.
If it is not building brand new big fabs all the time, then it will fall behind.
I mean, Intel could very well stop fab production and just switch to TSMC. But then they will always be on the same node as the competition…
DaRAGingLunatic
Member
Ok lets see.
What is intel currently doing with its 22nm, 14nm++++, 10nm++… 22nm is making interposers I think.
How about stop producing meteor lake on intel 4 and start using those wafers for cache. Haha.
Probably the simplest solution. Start sticking cache to those intel chips and then they will become a bit more competitive.
What is intel currently doing with its 22nm, 14nm++++, 10nm++… 22nm is making interposers I think.
How about stop producing meteor lake on intel 4 and start using those wafers for cache. Haha.
Probably the simplest solution. Start sticking cache to those intel chips and then they will become a bit more competitive.