Asianometry: Intel's Reign of Terror

[Xebec]

An excellent historical piece from Asianometry:

Some interesting tidbits:

- Some detail about Andy Grove's retirement (and his leadership) - "he was tired".
- Craig Barretts tenure and his (mostly failed) strategy around growing Intel into (tele)communications (perhaps this failure made Otellini more cautious later?)
- The Andy/Craig focus on stock buybacks - in 2005, more buybacks than investments in Fab R&D
- A quote of Craig Barrett.. quoting Gordon Moore re: you don't save your way out of a recession, you invest your way
- How the transition to Otellini happened, and why he was selected. Some disucssion on missing the iPhone.

- Other Items of Interest:
- The 1990s runup in revenue, including the creation of alternate product lines: Celeron, Xeon
- Tugs of war between Compaq and Intel, and Microsoft and Intel (Gates vs. Grove ).
- Intel's "bad business" behavior. Funny: Intel found not guilty of popping VIA's balloons at a trade show: https://www.zdnet.com/article/intel-executives-cleared-of-balloon-popping-charges/ (Not as funny) - the anti-trust suits.

There's a lot more content here, but while I've seen retrospectives on early days and later days, I found this unique to me covering the 'middle days' of Intel.

 

[hist78]

An excellent historical piece from Asianometry:

Some interesting tidbits:

- Some detail about Andy Grove's retirement (and his leadership) - "he was tired".
- Craig Barretts tenure and his (mostly failed) strategy around growing Intel into (tele)communications (perhaps this failure made Otellini more cautious later?)
- The Andy/Craig focus on stock buybacks - in 2005, more buybacks than investments in Fab R&D
- A quote of Craig Barrett.. quoting Gordon Moore re: you don't save your way out of a recession, you invest your way
- How the transition to Otellini happened, and why he was selected. Some disucssion on missing the iPhone.

- Other Items of Interest:
- The 1990s runup in revenue, including the creation of alternate product lines: Celeron, Xeon
- Tugs of war between Compaq and Intel, and Microsoft and Intel (Gates vs. Grove ).
- Intel's "bad business" behavior. Funny: Intel found not guilty of popping VIA's balloons at a trade show: https://www.zdnet.com/article/intel-executives-cleared-of-balloon-popping-charges/ (Not as funny) - the anti-trust suits.

There's a lot more content here, but while I've seen retrospectives on early days and later days, I found this unique to me covering the 'middle days' of Intel.

My thoughts after watching this video:

Intel's motto: Compete against our own customers/partners whenever that's possible.

It was then and it is now.

 

[cheesehead]

Intel's motto: Compete against our own customers/partners whenever that's possible.

I don´t necessarily agree. In my opinion this behavior is merely the result of their actual motto: "maximize my profit". Not only did they show little loyalty to their partners, but also squeezed out competition to make sure every available dollar on the market would land in their pockets.

While this may considered a desirable situation in capitalism, I think things went wrong during the peak of Intel market power in the mid 2000s. This is when they realized their market stopped growing rapidly and started stock buy back programs to maximize shareholder value.
If they had focused on customer value instead, their shareholder value might look entirely different today...

 

[tooLongInEDA]

I don´t necessarily agree. In my opinion this behavior is merely the result of their actual motto: "maximize my profit". Not only did they show little loyalty to their partners, but also squeezed out competition to make sure every available dollar on the market would land in their pockets.

While this may considered a desirable situation in capitalism, I think things went wrong during the peak of Intel market power in the mid 2000s. This is when they realized their market stopped growing rapidly and started stock buy back programs to maximize shareholder value.
If they had focused on customer value instead, their shareholder value might look entirely different today...

That is perhaps the key point from what I thought was an excellent video. Essentially that Intel didn't build trust and create lasting win-win relationships. So that when market infelctions and troubles did eventually come there wasn't a lot of goodwill and loyalty to fall back on.

This is all rather odd to me as my experiences in ASIC and EDA are that successful relationships need close cooperation and teamwork at all levels and a very high degree of trust. Perhaps Intel persisted longer with a more separated component supplier type mindset.

There's also an element of zero sum game thinking about Intel's historic approach (as claimed by Asianometry). They certainly went all out to maximise their share of the pie (PC industry profits in this case) far more than relying on the size of the pie (PC market) increasing. Straying off target, I do wonder if Apple hasn't taken exactly the same approach to the mobile phone market. Fine while it's working. But it's a long way down if and when that stops.

 

[Artificer60]

That is perhaps the key point from what I thought was an excellent video. Essentially that Intel didn't build trust and create lasting win-win relationships. So that when market infelctions and troubles did eventually come there wasn't a lot of goodwill and loyalty to fall back on.

This is all rather odd to me as my experiences in ASIC and EDA are that successful relationships need close cooperation and teamwork at all levels and a very high degree of trust. Perhaps Intel persisted longer with a more separated component supplier type mindset.

There's also an element of zero sum game thinking about Intel's historic approach (as claimed by Asianometry). They certainly went all out to maximise their share of the pie (PC industry profits in this case) far more than relying on the size of the pie (PC market) increasing. Straying off target, I do wonder if Apple hasn't taken exactly the same approach to the mobile phone market. Fine while it's working. But it's a long way down if and when that stops.

Assuming they choose an outside candidate, Intel's new CEO could go a long way to change that part of the culture. Someone with some sort of experience in both manufacturing and manufacturing customer designs would be optimal in my mind. With a stand alone CEO Intel products should be able to regulate themselves. The CEO could focus on fixing the foundry business.

 

[siliconbruh999]

Assuming they choose an outside candidate, Intel's new CEO could go a long way to change that part of the culture. Someone with some sort of experience in both manufacturing and manufacturing customer designs would be optimal in my mind. With a stand alone CEO Intel products should be able to regulate themselves. The CEO could focus on fixing the foundry business.

I don't have hope on Product CEO due to the mid designs we got she will be taking credit for Pat's products like he took credit for Swan's

 

[XYang2023]

I don't have hope on Product CEO due to the mid designs we got she will be taking credit for Pat's products like he took credit for Swan's

I think a bit more honesty might also be needed. Under her management, she has at least made the Arc GPU group more relevant. In contrast, within the DCAI group, the focus seems to be more on Gaudi rather than on GPUs—just an observation from an outside perspective.

What could she have done differently with the Arrow Lake launch? I wonder if the restructuring after the August earnings announcement had anything to do with the Arrow Lake launch.

Also, there were reports of products to address AMD's 3d-cache:

Intel's Patent Details Meteor Lake's 'Adamantine' L4 Cache

Intel's Adamantine L4 cache tile will rival AMD's 3D V-Cache tile.

www.tomshardware.com

Was the cancellation of such products due to PG's efforts of "portfolio trimming/simplification")?

 

[siliconbruh999]

I think a bit more honesty might also be needed. Under her management, she has at least made the Arc GPU group more relevant. In contrast, within the DCAI group, the focus seems to be more on Gaudi rather than on GPUs—just an observation from an outside perspective.

Yeah i don't have issue at her at communication but at the Product level due to the technical skill required imo

What could she have done differently with the Arrow Lake launch? I wonder if the restructuring after the August earnings announcement had anything to do with the Arrow Lake launch.

Also, there were reports of products to address AMD's 3d-cache:

Intel's Patent Details Meteor Lake's 'Adamantine' L4 Cache

Intel's Adamantine L4 cache tile will rival AMD's 3D V-Cache tile.

www.tomshardware.com

Was the cancellation of such products due to PG's efforts of "portfolio trimming/simplification")?

yes. Pat was overconfident many times imo

 

[XYang2023]

Yeah i don't have issue at her at communication but at the Product level due to the technical skill required imo

yes. Pat was overconfident many times imo

It seems PG focused all his attention on the Foundry and CHIPS Act while, at the same time, not taking the competition seriously. He appears to be a highly subjective person, though I’m not sure if that observation is entirely accurate.

 

[Barnsley]

I don´t necessarily agree. In my opinion this behavior is merely the result of their actual motto: "maximize my profit". Not only did they show little loyalty to their partners, but also squeezed out competition to make sure every available dollar on the market would land in their pockets.

While this may considered a desirable situation in capitalism, I think things went wrong during the peak of Intel market power in the mid 2000s. This is when they realized their market stopped growing rapidly and started stock buy back programs to maximize shareholder value.
If they had focused on customer value instead, their shareholder value might look entirely different today...

Isnt a CEO No.1 priority shareholder value?

All those resources wasting pumping the share price.

Ah well!

 

[siliconbruh999]

Isnt a CEO No.1 priority shareholder value?

All those resources wasting pumping the share price.

Ah well!

There are different way to maximize shareholders value the best method is to invest in your tech and grow it this directly affects shareholder value other one is share buyback and dividends you need to be careful with these cause it can get you in a bad situation (Intel is the prime example).

 

[XYang2023]

There are different way to maximize shareholders value the best method is to invest in your tech and grow it this directly affects shareholder value other one is share buyback and dividends you need to be careful with these cause it can get you in a bad situation (Intel is the prime example).

I don’t believe the approach Pat took was the right way to maximize shareholder returns. He overextended the company. Instead, he could have:
1. Streamlined the company from the start.
2. Limited fab build-outs to Ireland and Arizona.
3. Pooled resources into GPU development to compete with Nvidia, even if the initial products were inferior, rather than having no presence in the market.
4. Exploring opportunities outside the x86 architecture, e.g. acquiring Nuvia,
5. Considered collaborating with Japan on next-generation fab build-outs instead of focusing on Ohio.

Essentially, the goal should be to maximize the value of every dollar the company has, de-risk the business, and focus on growth. I don’t think he has achieved any of these objectives.

I think, at the moment, Intel is focusing on better utilization of its capital, stabilizing its current business, and launching GPUs/accelerators to capture a share of the growth market.

 

[XYang2023]

One way Intel could change its culture is by learning from the current natural disaster in LA. It should terminate any DEI programs, streamline unnecessary middle management layers, and promote meritocracy and fairness.

 

[siliconbruh999]

One way Intel could change its culture is by learning from the current natural disaster in LA. It should terminate any DEI programs, streamline unnecessary middle management layers, and promote meritocracy and fairness.

DEI is an issue it had destroyed Video Game industry already in the west

I don’t believe the approach Pat took was the right way to maximize shareholder returns. He overextended the company. Instead, he could have:
1. Streamlined the company from the start.
2. Limited fab build-outs to Ireland and Arizona.
3. Pooled resources into GPU development to compete with Nvidia, even if the initial products were inferior, rather than having no presence in the market.
4. Exploring opportunities outside the x86 architecture, e.g. acquiring Nuvia,
5. Considered collaborating with Japan on next-generation fab build-outs instead of focusing on Ohio.

Definitely I would say x86 doesn't need replacement rn so it can be scheduled later and on Ohio I don't see the issue but see issues with Poland/Israel/Germany as for Japan with fab buildout they are already opening a tech center but overall I agree with these points

Essentially, the goal should be to maximize the value of every dollar the company has, de-risk the business, and focus on growth. I don’t think he has achieved any of these objectives.

I think, at the moment, Intel is focusing on better utilization of its capital, stabilizing its current business, and launching GPUs/accelerators to capture a share of the growth market.

He de Risked the FAB and client Business though but didn't achieve growth and piled more issues

 

[osnium]

One way Intel could change its culture is by learning from the current natural disaster in LA. It should terminate any DEI programs, streamline unnecessary middle management layers, and promote meritocracy and fairness.

One can imagine that the Trump administration could change the Chips act grants terms to force Intel to terminate DEI programs.

 

[Brady]

I don’t believe the approach Pat took was the right way to maximize shareholder returns. He overextended the company. Instead, he could have:

3. Pooled resources into GPU development to compete with Nvidia, even if the initial products were inferior, rather than having no presence in the market.

I think, at the moment, Intel is focusing on better utilization of its capital, stabilizing its current business, and launching GPUs/accelerators to capture a share of the growth market.

Question regarding GPUs: did Gelsinger not do just that with Falcon Shores? It was a mixed “APU” like AMD MI300, but it’s now just a GPU, correct? From what I understand this is why the Gaudi line probably won’t see a Gaudi 4 as Intel is (finally) merging their Gaudi and Xe IP into a BFGPU for the data center.

Is my understanding wrong?

 

[XYang2023]

Question regarding GPUs: did Gelsinger not do just that with Falcon Shores? It was a mixed “APU” like AMD MI300, but it’s now just a GPU, correct? From what I understand this is why the Gaudi line probably won’t see a Gaudi 4 as Intel is (finally) merging their Gaudi and Xe IP into a BFGPU for the data center.

Is my understanding wrong?

My understanding is that Falcon Shores is a platform. Some IPs from Gaudi, such as communication, might be merged into the platform. The platform is an XPU platform that allows for:

CPU+GPU (HPC?)
GPU only (training)
Gaudi only (inferencing)
And other combinations

All of them use oneAPI.

 

[KevinK]

One way Intel could change its culture is by learning from the current natural disaster in LA. It should terminate any DEI programs, streamline unnecessary middle management layers, and promote meritocracy and fairness.

Maybe Intel's problem is the lack of DEI at the top ? Intel‘s greatest successes came under the leadership of a couple of white Jewish guys, and downfall came under a succession of leadership by white Christian men. Intel is currently getting their bottom handed to them by a couple companies, one being led by an Asian woman and the other by an Asian man, both of which have white Jewish men as CTOs. The optics at the top kind of contradict your thesis

Sorry - just being the provocateur here

 

[XYang2023]

Maybe Intel's problem is the lack of DEI at the top - Intel‘s greatest successes came under the leadership of a couple of white Jewish guys, and downfall came under a succession of leadership by white Christian men. Intel is currently getting their bottom handed to them by a couple companies, one being led by an Asian woman and the other under an Asian man, both of which have white Jewish men as CTOs. The optics at the top kinds of contradict your thesis

My understanding of DEI is that the system enforces artificial quotas. I mentioned Intel should instead promote meritocracy and fairness. If Taiwanese managers can better manage fabs, then I think Intel should consider Taiwanese managers alongside all other candidates for their fabs. They should be judged on an individual basis rather than through a quota system. The objective should be to compete effectively in the marketplace at the lowest cost.

 

[Brady]

My understanding of DEI is that the system enforces artificial quotas.

No good DEI system does this.

A good DEI system says “we acknowledge that we have systemically overlooked certain groups of people. This is artificially reducing our talent pool”.

Of course, the ever-insightful reactionaries always scream “quotas! White replacement!”, and poorly implemented DEI initiatives reinforcement this sentiment.

It’s truly sad that a good thing is twisted. But then, the scourge of identity politics has done this as a whole.