Intel Broadwell BDW-H delayed May 2015

[phillyman36]

Just saw this.


http://wccftech.com/intel-broadwell-enthusiasts-delayed-2015/

 

[inf64]

wcctech, makes mistake even in very 1st sentence of their "news article"

A couple of days back, we informed you about direct word from Intel management about the release schedule of the 14nm ‘tock’ version of Haswell, the Intel Broadwell.

They don't even know what is tick and what is tock. Sad.

 

[Fjodor2001]

But what's important is the original source, which is Chinese VR-Zone, and they are usually correct.

Anyway, here's the release plan:

 

[Fjodor2001]

If this is correct you have to wonder how come Broadwell-H will be released so much later than Y/U? Is the intention to use it as a mid life kicker?

Also, no date set for the release of Broadwell-K in the leaked plan, right? Or is it included in the last H-row? It says "[...] 4+3e LGA platform"?

 

[Arachnotronic]

Think about the markets that Intel serves. Thin and light notebooks are the most important, with big DTR machines and desktops less so. You milk the 22nm Haswell (which has no real CPU competition in those areas anyway) in the areas where you're so far ahead it doesn't matter and then you bring the shiny 14nm goodness to the areas that need it most.

This makes perfect business sense, especially now that the enthusiasts will be served by the -E line of products.

 

[bullzz]

giving priority to U/Y makes a lot of sense. H can wait

 

[Techhog]

Safe to say that Broadwell-K is well over a year away then. Probably late Q3/early Q4 2015

 

[bullzz]

wait. just noticed U2+3 is also RTS 2015. so apple is not going to refresh its macbook current products?

 

[Ajay]

Safe to say that Broadwell-K is well over a year away then. Probably late Q3/early Q4 2015

Makes me wonder if the rumor that Broadwell-K will be cancelled and Skylake LGA taking it's place is true. Seems dumb to come out with BW-K unless Skylake is being pushed out as well.

 

[Ajay]

Think about the markets that Intel serves. Thin and light notebooks are the most important, with big DTR machines and desktops less so. You milk the 22nm Haswell (which has no real CPU competition in those areas anyway) in the areas where you're so far ahead it doesn't matter and then you bring the shiny 14nm goodness to the areas that need it most.

This makes perfect business sense, especially now that the enthusiasts will be served by the -E line of products.

So now we see the affects of no competition from AMD. We are not being sucked dry by high prices**, but are being bored by the slower pace of product releases!

** Although delays will push more enthusiasts to the higher priced HEDT (server derived) CPUs and motherboards, which is in effect a price increase.

 

[Fjodor2001]

giving priority to U/Y makes a lot of sense. H can wait

Yes, but priority for what reason? Are they having yield issues, so they prioritize producing U/Y chips?

Another observation is that all the high frequency parts will not be released until May 2015. That's a lot later than the U/Y chips. I would not be surprised if that's because they are having problems to produce high frequency chips @ 14 nm, and are hoping to have perfected the process tech by then...

 

[phillyman36]

It said Broadwell Delayed 2015. What I heard(in my mind) Skylake 2016 lol

 

[ShintaiDK]

So now we see the affects of no competition from AMD. We are not being sucked dry by high prices**, but are being bored by the slower pace of product releases!

** Although delays will push more enthusiasts to the higher priced HEDT (server derived) CPUs and motherboards, which is in effect a price increase.

Thats still BS. Intel already for years only competed with itself. If there is nothing new worth buying, people dont buy. Hence lower revenue, lower profit and so forth. There is no static demand, its dynamic.

 

[Fjodor2001]

So now we see the affects of no competition from AMD. We are not being sucked dry by high prices**, but are being bored by the slower pace of product releases!

** Although delays will push more enthusiasts to the higher priced HEDT (server derived) CPUs and motherboards, which is in effect a price increase.

Yes. Intel's new strategy is as follows:

1. Prioritize mobile to make use of process tech advantage to compete with ARM/Qualcomm/etc, where volume is going towards. That's Intel's only hope of entering that market segment. But it's questionable if Intel will succeed since they've had mobile in focus for years, but failed to succeed in that segment.

2. Intentionally delay mainstream desktop CPUs more than necessary to force enthusiasts to HEDT platform, where margins are higher.

The release plan in this thread makes it all clear.

 

[Techhog]

Makes me wonder if the rumor that Broadwell-K will be cancelled and Skylake LGA taking it's place is true. Seems dumb to come out with BW-K unless Skylake is being pushed out as well.

Well, two possibilities: Skaylake is late 2016 for desktops, or the "K" CPUs will now be a generation behind like the "E" CPUs.

 

[witeken]

Think about the markets that Intel serves. Thin and light notebooks are the most important, with big DTR machines and desktops less so. You milk the 22nm Haswell (which has no real CPU competition in those areas anyway) in the areas where you're so far ahead it doesn't matter and then you bring the shiny 14nm goodness to the areas that need it most.

This makes perfect business sense, especially now that the enthusiasts will be served by the -E line of products.

But 14nm brings huge cost/transistor improvements. And the fabs are now in volume production, why would they wait so much longer? And Skylake should be released only a few months later.

 

[witeken]

Well, two possibilities: Skylake is late 2016 for desktops, or the "K" CPUs will now be a generation behind like the "E" CPUs.

Skylake is a 2015 product, certainly not late 2016.

Source 1.
Source 2:

 

[Techhog]

Skylake is a 2015 product, certainly not late 2016.

Source 1.
Source 2:

That roadmap also calls Broadwell-H a 2014 product. We don't know what they'll do with Skylake, but given the extreme slowdown in node advancement, it would probably make more sense for Intel to really stagger releases to maximize profits.

 

[Ajay]

Thats still BS. Intel already for years only competed with itself. If there is nothing new worth buying, people dont buy. Hence lower revenue, lower profit and so forth. There is no static demand, its dynamic.

WTH? I made no assumption about the market being static or dynamic.

Anyway: 1) If there was more competition on for Destktop/DTR x86 CPUs then product release cycle wouldn't be slowed down. It's taken a while for Intel management to decide to capitalize on this, that's all. 2) More competition in x86 could even lead to expanded sales because there would be more excitement in the market. There would likely be an increase in performance from one release to another - creating a higher churn rate.

So yes, it is a dynamic market. And, no, you aren't able to read my mind.

 

[Ajay]

That roadmap also calls Broadwell-H a 2014 product. We don't know what they'll do with Skylake, but given the extreme slowdown in node advancement, it would probably make more sense for Intel to really stagger releases to maximize profits.

It is more of a product slowdown than a slowdown in node development. And even then it's not clear whether this is a one off or not yet. If there is a real slow down in node introduction, then we will know for sure by the next node (10nm), IMHO.

 

[ShintaiDK]

Yes. Intel's new strategy is as follows:

2. Intentionally delay mainstream desktop CPUs more than necessary to force enthusiasts to HEDT platform, where margins are higher.

The release plan in this thread makes it all clear.

People buying LGA2011-x for desktop is simply staggering. I mean its a multi billion$ segment! And you cant be an enthusiast if you dont spend all you can on bling bling. The more money spend, the bigger enthusiast.

Might have a slight touch of sarcasm.

 

[ShintaiDK]

WTH? I made no assumption about the market being static or dynamic.

Anyway: 1) If there was more competition on for Destktop/DTR x86 CPUs then product release cycle wouldn't be slowed down. It's taken a while for Intel management to decide to capitalize on this, that's all. 2) More competition in x86 could even lead to expanded sales because there would be more excitement in the market. There would likely be an increase in performance from one release to another - creating a higher churn rate.

So yes, it is a dynamic market. And, no, you aren't able to read my mind.

I think you confuse the past and now. Desktop was a majority market back then. Today its the most insignificant segment in priority terms there is. And a lot people seems to be in denial of this and blaming/excuse everything else. The same goes for dGPUs for that matter.

 

[witeken]

That roadmap also calls Broadwell-H a 2014 product. We don't know what they'll do with Skylake, but given the extreme slowdown in node advancement, it would probably make more sense for Intel to really stagger releases to maximize profits.

This roadmap is from May. Need yet another source? Intel will start volume production of 10nm in 2015 , so Cannonlake will be a mid to late 2016 product.

And as I said earlier, going to a lower node doesn't only benefit the consumer, but margins increase as well, and those nodes is what allows Intel to rapidly gain market share in the tablet and smartphone markets, so there's no slowing down.

 

[Fjodor2001]

And as I said earlier, going to a lower node doesn't only benefit the consumer, but margins increase as well, and those nodes is what allows Intel to rapidly gain market share in the tablet and smartphone markets, so there's no slowing down.

Not as simple as that. They have to get ROI for their invested R&D money. So the sales window of each CPU generation cannot be too short.

If you spend $1 billion on R&D for a CPU generation, and only have a sales window allowing you to earn $0.5 billion until the next CPU generation is released, you will still make a loss.

 

[witeken]

Two years is plenty for Intel (~100B$ revenue).