TR Forums

https://finance.yahoo.com/news/intel-pr ... 04969.html

"We're going to be pretty pragmatic about if and when we should be making stuff inside or making outside, and making sure that we have optionality to build internally, mix and match inside and outside, or go outside in its entirety if we need to," Chief Executive Bob Swan said on a call with investors.

you heard it here first, folks -- Intel to go fabless!

Where's chuckula to point out the revenues, gross margin, and profit?

Sounds like Intel is going "Hey, it worked for AMD..."

cegras wrote:

Where's chuckula to point out the revenues, gross margin, and profit?

Usually he also posts a stock price chart but that doesn’t work this time.

So 10nm was delayed and delayed and delayed and eventually only used for a small number of mobile products.

And now 7nm is delayed by a year over original expectations.

This is not looking like a nice pattern.

K-L-Waster wrote:

So 10nm was delayed and delayed and delayed and eventually only used for a small number of mobile products.

And now 7nm is delayed by a year over original expectations.

This is not looking like a nice pattern.

When GloFo threw in the towel, I speculated that Intel would be next.

It seemed like a radical idea at the time, but this industry is driven by economies of scale. Intel surrendered the lead in economies of scale when they declined to participate in the smartphone market, preferring to defend PC profit margins. The fools running Intel thought their only competitor was AMD.

https://www.oregonlive.com/silicon-fore ... oyees.html

I wonder if the 2015 layoffs of over 1,000 senior engineers had anything to do with Intel’s failures these past five years.

Interesting management shakeup

https://wccftech.com/intel-restructurin ... velopment/

They are putting a woman in charge of making 7nm work. Given dr Lisa su’s success, maybe that’s the ticket to success.

I don’t mean that as flippantly as it might sound. Effective management requires setting your ego aside and letting others shine and it’s about facilitating productive collaboration. On average, the ladies might be better at that.

blastdoor wrote:

Interesting management shakeup
I don’t mean that as flippantly as it might sound. Effective management requires setting your ego aside and letting others shine and it’s about facilitating productive collaboration. On average, the ladies might be better at that.

Or, you get a Carly Fiorina. Could go either way.

ludi wrote:

blastdoor wrote:

Interesting management shakeup
I don’t mean that as flippantly as it might sound. Effective management requires setting your ego aside and letting others shine and it’s about facilitating productive collaboration. On average, the ladies might be better at that.

Or, you get a Carly Fiorina. Could go either way.

Ha — good point!

blastdoor wrote:

ludi wrote:

blastdoor wrote:

Interesting management shakeup
I don’t mean that as flippantly as it might sound. Effective management requires setting your ego aside and letting others shine and it’s about facilitating productive collaboration. On average, the ladies might be better at that.

Or, you get a Carly Fiorina. Could go either way.

Ha — good point!

Carly Fiorina - Destroyer of Technology Businesses

Thinking more about what future Intel has, given their repeated process face plants, I wonder if the thing for them to do is to give up on being an independent manufacturer and enter into some kind of joint venture with a foundry. Intel could spin off their fabs into this joint venture, maintaining partial ownership while the rest of the ownership goes to the partner.

TSMC would probably be the ideal partner, I just don't know if TSMC would give Intel the best terms. Sammy is the other obvious option, and they might give Intel better terms, given they're a bit more desperate.

Of course, another option would be for Intel senior management to pull its head out of its arse and F'ing FIGHT TO WIN. They could still do it if they tried -- they still have a lot of money, lots of IP, lots of smart engineers. I'm just not optimistic they're willing to try.

Any other thoughts on the future of Intel?

I figure they'll get their manufacturing issues sorted eventually. They're still the 500 lb gorilla in semiconductors.

just brew it! wrote:

I figure they'll get their manufacturing issues sorted eventually. They're still the 500 lb gorilla in semiconductors.

I don't know, I think maybe people's view of Intel is lagging reality. We all "grew up" with Intel as this untouchable leader far out in front of everybody. Back in 2005 they were able to bounce back from their terrible choices with Itanic and Nutbust, but they could do that because of their manufacturing lead, and they had that lead because of their economies of scale lead.

Economies of scale really is everything in semiconductors. Even though AMD edged in front of Intel for a while back in 2005, AMD's disadvantage in terms of scale meant they were always running up hill and into the wind.

Now Intel is the one running up hill and into the wind. TSMC has the economies of scale from the smartphone market providing a huge wind at their backs.

If Intel wants to get back in this game they need to do something big and they need to do it fast. They need to figure out how to match TSMC in terms of economies of scale. I don't see how they do that without partnering.

Sadly, though, I fear Intel is still too complacent. Fighting back is hard, going fabless is easy, so I bet they go fabless.

I'm not convinced they're complacent. If they were complacent they wouldn't have had 10nm and 7nm in the works. The issue here isn't that they weren't trying to move forward: the issue is their attempts failed.

What I could believe is they have a huge case of "not invented here" syndrome. As in, anything they came up with in-house must inherently be better than anything the competition comes up with simply because they're Intel.

There is definitely an echo of 80's and 90's era IBM in Intel's current predicament.

K-L-Waster wrote:

There is definitely an echo of 80's and 90's era IBM in Intel's current predicament.

And we all know where that led. On the upside, office workers finally got to wear non-white dress shirts.

blastdoor wrote:

Now Intel is the one running up hill and into the wind. TSMC has the economies of scale from the smartphone market providing a huge wind at their backs.

Something needs to power all the back-end systems that provide the services those smartphones rely on. And until pretty recently, AMD was not a credible player in the enterprise market. Intel has economies of scale too.

Captain Ned wrote:

K-L-Waster wrote:

There is definitely an echo of 80's and 90's era IBM in Intel's current predicament.

And we all know where that led. On the upside, office workers finally got to wear non-white dress shirts.

I still get people asking me if I need to wear a tie to the office.

(Assuming the office ever opens back up, I plan to go back to wearing jeans and Hawaiian print shirts to work, as opposed to my "work at home" shorts and a t-shirt...)

just brew it! wrote:

I still get people asking me if I need to wear a tie to the office.

60+ years later, the Thomas Watson meme still sticks hard to IBM.

For the young'uns. Circa the 1960's, IBM was the stereotypical blue suit, white shirt, red tie workplace. IBM men made FBI men look like fashion plates.

just brew it! wrote:

blastdoor wrote:

Now Intel is the one running up hill and into the wind. TSMC has the economies of scale from the smartphone market providing a huge wind at their backs.

Something needs to power all the back-end systems that provide the services those smartphones rely on. And until pretty recently, AMD was not a credible player in the enterprise market. Intel has economies of scale too.

Replace "smartphone" with "PC" and "Intel" with "IBM" and it's a new verse of the same song.

The market for x86 CPUs is approximately 200 million units per year. The market for iPhones alone is about 200 million units per year -- the entire smartphone market is over a billion units per year.

Meanwhile, the costs of moving to new processes continues on the same exponential growth curve they've always been on.

https://www.extremetech.com/computing/2 ... ocess-node

“3nm will cost $4 billion to $5 billion in process development, and the fab cost for 40,000 wafers per month will be $15 billion to $20 billion.”

also

even if these new transistor types can be refined and brought to market: design cost. The price of a 3nm chip is expected to range from between $500M to $1.5B, with the latter figure reserved for a high-end GPU from Nvidia.

Intel *can* afford to pay that amount - they have the profits to fund it. But will that be justifiable to the bean counters and Wall Street? Or will the bean counters conclude that it's cheaper to just outsource production to TSMC and use those profits to pay dividends?

The thing is, I think the bean counters are wrong. Intel isn't going to be able to hang on the way IBM has, just gradually shrinking while paying out dividends checks. Intel has nothing like the vendor lock-in that IBM *still* enjoys with its mainframe business. Nobody cares if their AWS Linux server is running Intel or AMD chips, and soon few may care if it's running ARM. Intel is far more vulnerable than they seem to realize. Losing their manufacturing edge means they compete only on design, and they don't seem to have much of a design edge.

So bottom line... I agree that Intel has the money to stay in the race. But shortsighted beancounter math would say that they don't have the economies of scale to make the investment worthwhile. Yet the beancounters aren't appreciating how fast and far Intel can fall if they are forced to compete on design alone. Intel is at huge risk right now and I'm so skeptical their management knows it or knows what to do about it.

blastdoor wrote:

just brew it! wrote:

blastdoor wrote:

Now Intel is the one running up hill and into the wind. TSMC has the economies of scale from the smartphone market providing a huge wind at their backs.

Something needs to power all the back-end systems that provide the services those smartphones rely on. And until pretty recently, AMD was not a credible player in the enterprise market. Intel has economies of scale too.

Replace "smartphone" with "PC" and "Intel" with "IBM" and it's a new verse of the same song.

The market for x86 CPUs is approximately 200 million units per year. The market for iPhones alone is about 200 million units per year -- the entire smartphone market is over a billion units per year.

I'm pretty sure that 200 million figure is desktops and laptops only...

Edit: Gartner says 261 million "PCs" shipped last year:
https://www.gartner.com/en/newsroom/pre ... r-the-year

just brew it! wrote:

blastdoor wrote:

just brew it! wrote:

Something needs to power all the back-end systems that provide the services those smartphones rely on. And until pretty recently, AMD was not a credible player in the enterprise market. Intel has economies of scale too.

Replace "smartphone" with "PC" and "Intel" with "IBM" and it's a new verse of the same song.

The market for x86 CPUs is approximately 200 million units per year. The market for iPhones alone is about 200 million units per year -- the entire smartphone market is over a billion units per year.

I'm pretty sure that 200 million figure is desktops and laptops only...

Edit: Gartner says 261 million "PCs" shipped last year:
https://www.gartner.com/en/newsroom/pre ... r-the-year

So a slightly larger drop in the bucket compared to the nearly 1.4 billion smartphones (https://www.canalys.com/newsroom/canaly ... et-q4-2019).

Not really fair to refer to "within an order of magnitude" as "a drop in the bucket". Also keep in mind that desktop/server chips are physically much larger, so in terms of wafer volume they're probably comparable (or Intel may even be ahead).

just brew it! wrote:

Not really fair to refer to "within an order of magnitude" as "a drop in the bucket". Also keep in mind that desktop/server chips are physically much larger, so in terms of wafer volume they're probably comparable (or Intel may even be ahead).

Also remember that the only reason Smartphones and ultrabooks are viable as primary devices is because most of the heavy lifting happens server-side in the vast majority of apps. The more popular mobile devices get, the more data center CPU cores there need to be.

Intel's lead in the data center isn't unassailable, but it's not going to collapse in a quarter or two.

K-L-Waster wrote:

just brew it! wrote:

Not really fair to refer to "within an order of magnitude" as "a drop in the bucket". Also keep in mind that desktop/server chips are physically much larger, so in terms of wafer volume they're probably comparable (or Intel may even be ahead).

Also remember that the only reason Smartphones and ultrabooks are viable as primary devices is because most of the heavy lifting happens server-side in the vast majority of apps. The more popular mobile devices get, the more data center CPU cores there need to be.

Intel's lead in the data center isn't unassailable, but it's not going to collapse in a quarter or two.

I certainly am not arguing it will collapse in a quarter or two. More like a year or two.

blastdoor wrote:

I certainly am not arguing it will collapse in a quarter or two. More like a year or two.

LOL... even if it does end up killing them (unlikely), it'll take a lot longer than that.

Look at how long AMD survived between the release of Bulldozer and Zen.

just brew it! wrote:

blastdoor wrote:

I certainly am not arguing it will collapse in a quarter or two. More like a year or two.

LOL... even if it does end up killing them (unlikely), it'll take a lot longer than that.

Look at how long AMD survived between the release of Bulldozer and Zen.

THE LEAD will collapse. Not intel as a company.

blastdoor wrote:

just brew it! wrote:

blastdoor wrote:

I certainly am not arguing it will collapse in a quarter or two. More like a year or two.

LOL... even if it does end up killing them (unlikely), it'll take a lot longer than that.

Look at how long AMD survived between the release of Bulldozer and Zen.

THE LEAD will collapse. Not intel as a company.

I don't think that's gonna happen in just a year or two either. I doubt AMD has enough fab capacity contracted at TSMC to take the lead away from them.

just brew it! wrote:

blastdoor wrote:

just brew it! wrote:

LOL... even if it does end up killing them (unlikely), it'll take a lot longer than that.

Look at how long AMD survived between the release of Bulldozer and Zen.

THE LEAD will collapse. Not intel as a company.

I don't think that's gonna happen in just a year or two either. I doubt AMD has enough fab capacity contracted at TSMC to take the lead away from them.

Well, Huawei was booted and they used a fair bit of capacity.
Also, AMD doesn't use TSMC's cutting edge, they are 1 year behind (which is still ahead of poor sad Intel).
So.... I dunno. I'll bet they aren't capacity constrained.

Intel is doomed:

https://www.anandtech.com/show/15981/in ... -play-with

Feels kinda retro to me; GPU vendors were doing the gaudy retail packaging thing 15+ years ago. I guess Intel is longing for the days when AMD didn't pose as much of a threat.