Intel's 7nm is broken, company announces delay until 2022, 2023

[williamcll]

OH nonononono

[mr moose]

21 minutes ago, williamcll said:

OH nonononono

Watch it over the next week. It probably isn't as bad as it looks. 

 

 

[porina]

1 hour ago, Arika S said:

and now for Intel, they can't exactly just switch to another fab, they have the figure it out themselves (as they should). TSMC would have to be running at near capacity these days, no way they could accommodate Intel even if they wanted to.

Intel have just said they will look to use other fabs. We just don't know which products, at which fabs, on which processes. Depending on when they made that decision will affect when products will start to go to external fabs, since they need to be designed for the process which takes time. Still, taking off the internal pressure will probably help them out a bit there.

 

TSMC just got a bunch of potential capacity freed up with the ongoing war the US gov are having with Huawei, but we don't know who will gobble that up yet.

 

1 hour ago, Arika S said:

which i think is something people forget when talking about how AMD took over the lead, they didn't have to worry about the fab aspect of it which is a huge burden they don't have to deal with.

That was a decision made for survival at the time. Their finances then were not great. Also it remains a gamble. Today, it has worked out really well for them, but what if TSMC had the similar problems Intel had? There is no way they could know that at the time.

 

16 minutes ago, mr moose said:

Watch it over the next week. It probably isn't as bad as it looks. 

Agreed, the stock market tends to (over) react to good or bad news initially, and it'll shift towards a new normal as they take a breather. That doesn't mean Intel wont take a hit, but the question is how much?

 

It isn't unusual for heads to roll at the top if there is a big drop in stock price. This one isn't good, but might not be bad enough to trigger shareholders to ask for blood. I also wonder what Intel's management plan is. As I understand, Bob Swan is only interim CEO after his predecessor was less than gracefully required to leave.

[xg32]

43 minutes ago, mr moose said:

Watch it over the next week. It probably isn't as bad as it looks. 

 

 

if it stays under 58 it likely isn't going anywhere, i was already going out of my way to take it out of any of the tech funds i own, amd's on a huge breakout though, probably going to 90, that's the big part of this news. intc is at best a dead stock.

[RejZoR]

54 minutes ago, mr moose said:

Watch it over the next week. It probably isn't as bad as it looks. 

 

 

It always drops on drama and then goes back to nearly the same point it was before. The wonders of stock exchange where everything is speculated on bullshit most of the time.

[5x5]

2 hours ago, RejZoR said:

It always drops on drama and then goes back to nearly the same point it was before. The wonders of stock exchange where everything is speculated on bullshit most of the time.

This isn't just drama, this is a massive delay. Mark my words, you will NOT see anything but 14nm+++++ up until the second half of 2022 AND weay even get a Broadwell-style cancellation of the gen after Rocket Lake. Rocket Lake will NOT have PCIe 4.0 from what I have heard and it will be a back port to 14nm as 10nm is proving disastrous on desktop and server. 10nm development is getting killed off completely and Intel's 7nm is going the same way 10nm went all those years ago. If intel don't make drastic concessions, 7nm will be a flop as, by the time it's usable for anything, they'll be up against TSMC 3nm

[Jito463]

I wonder if Intel might - instead of utilizing another fab for their production - attempt to license the technology to use in their own facilities.  For example, pay a licensing fee to TSMC to use their 7nm process, but build the actual products in their own fab.  Intel gets off the 14nm++++++++++++++++++++++++++++++++ bandwagon, and TSMC gets a chunk of change without having to add any additional production capability.

[5x5]

1 minute ago, Jito463 said:

I wonder if Intel might - instead of utilizing another fab for their production - attempt to license the technology to use in their own facilities.  For example, pay a licensing fee to TSMC to use their 7nm process, but build the actual products in their own fab.  Intel gets off the 14nm++++++++++++++++++++++++++++++++ bandwagon, and TSMC gets a chunk of change without having to add any additional production capability.

That's a possibility but unlikely. TSMC can milk a lot more from intel directly as they have the money to pay up and the dire need for fab space.

[Mateyyy]

That sucks, not just for Intel but also potentially for the consumers. Hopefully Intel gets their shit together before AMD could start mimicking Intel's super incremental "upgrades", because competition is key.

If Intel are still on the same 14nm process with the upcoming move to DDR5, I see AMD absolutely smoking Intel in desktop market share with AM5.

[KaitouX]

6 minutes ago, Mateyyy said:

That sucks, not just for Intel but also potentially for the consumers. Hopefully Intel gets their shit together before AMD could start mimicking Intel's super incremental "upgrades", because competition is key.

If Intel are still on the same 14nm process with the upcoming move to DDR5, I see AMD absolutely smoking Intel in desktop market share with AM5.

Considering how much time AMD needs to get their market and mind share high enough to be able to do that, I don't think AMD will be slowing down the performance or increasing the prices too much. But i do hope Intel gets back to it in the next 3 years or so, it certainly would have some effect on pricing if Intel went completely "quiet" like AMD did with Bulldozer.

[BigDamn]

Intel trading below $50 right now, and AMD over $66. Never thought I'd see the day...

[D13H4RD]

3 minutes ago, BigDamn said:

Intel trading below $50 right now, and AMD over $66. Never thought I'd see the day...

Green team seems to be on a bit of a roll this year although they have dipped a bit 

[5x5]

18 minutes ago, Mateyyy said:

That sucks, not just for Intel but also potentially for the consumers. Hopefully Intel gets their shit together before AMD could start mimicking Intel's super incremental "upgrades", because competition is key.

If Intel are still on the same 14nm process with the upcoming move to DDR5, I see AMD absolutely smoking Intel in desktop market share with AM5.

AMD won't stagnate as much. They can't afford that. But intel are rapidly losing desktop maker share. I think AMD has about 30-35% concurrently.

[Lord Vile]

6 minutes ago, 5x5 said:

AMD won't stagnate as much. They can't afford that. But intel are rapidly losing desktop maker share. I think AMD has about 30-35% concurrently.

It's under 30

[Master Disaster]

Jesus.

 

Isn't it about time Intel start looking at abandoning Core entirely and calling in Jim the Magician to design them an all new architecture? The sunk cost fallacy very much is in effect here.

[porina]

 

7 minutes ago, Master Disaster said:

Isn't it about time Intel start looking at abandoning Core entirely and calling in Jim the Magician to design them an all new architecture? The sunk cost fallacy very much is in effect here.

Intel's problem isn't designing a CPU. It's making the things. Two separate areas. I know desktop is still stuck on Skylake, but they have the first post-skylake generation (Sunny Cove) in Ice Lake laptop CPUs over a year ago. It's successor Tiger Lake is close to launch, and likely to get another architecture generation update (Willow Cove). Sunny Cove is a significant uplift in IPC over Skylake, the problems with it mainly come down to the 10nm problems. Changing the architecture without improving the process wont help much.

[rogerwilco91]

5 hours ago, porina said:

Intel have just said they will look to use other fabs. We just don't know which products, at which fabs, on which processes. Depending on when they made that decision will affect when products will start to go to external fabs, since they need to be designed for the process which takes time. Still, taking off the internal pressure will probably help them out a bit there.

 

TSMC just got a bunch of potential capacity freed up with the ongoing war the US gov are having with Huawei, but we don't know who will gobble that up yet.

 

That was a decision made for survival at the time. Their finances then were not great. Also it remains a gamble. Today, it has worked out really well for them, but what if TSMC had the similar problems Intel had? There is no way they could know that at the time.

 

Agreed, the stock market tends to (over) react to good or bad news initially, and it'll shift towards a new normal as they take a breather. That doesn't mean Intel wont take a hit, but the question is how much?

 

It isn't unusual for heads to roll at the top if there is a big drop in stock price. This one isn't good, but might not be bad enough to trigger shareholders to ask for blood. I also wonder what Intel's management plan is. As I understand, Bob Swan is only interim CEO after his predecessor was less than gracefully required to leave.

He became the full time CEO not long after.
 

 

[Master Disaster]

3 minutes ago, porina said:

 

Intel's problem isn't designing a CPU. It's making the things. Two separate areas. I know desktop is still stuck on Skylake, but they have the first post-skylake generation (Sunny Cove) in Ice Lake laptop CPUs over a year ago. It's successor Tiger Lake is close to launch, and likely to get another architecture generation update (Willow Cove). Sunny Cove is a significant uplift in IPC over Skylake, the problems with it mainly come down to the 10nm problems. Changing the architecture without improving the process wont help much.

Right, but starting over with a new design would mean retooling the fabs and starting again with a new process.

[porina]

5 minutes ago, rogerwilco91 said:

He became the full time CEO not long after.

Doh! Missed that.

 

2 minutes ago, Master Disaster said:

Right, but starting over with a new design would mean retooling the fabs and starting again with a new process.

Resetting their fabs is no trivial matter. They have previously stated they will decouple the design from fab more going forwards, which potentially ties in with them exploring external fabs. I'm sure they're still tinkering on the edges of 10nm but their focus is on their (now officially delayed) 7nm process. They say they know what is wrong and can fix it, but we'll have to wait and see if that explodes like 10nm did.

 

I suppose we can look at it in 3 parts, the logical architecture design (not a problem), the fabrication process (a big problem), and in between is the implementation of that design for that process. It's not like they have a drop down in CPU design software where you pick 14nm, 12nm, 10nm, 7nm and it does a % scale on it like you were printing onto different paper sizes. The design rules for the specific process have to be implemented which is quite time consuming. Like Rocket Lake is expected to appear next year with some post-Skylake architecture but built on 14nm.

 

What we might see them doing going forward is implementing the design on more than one process to cover their options and reduce risk. For example, they might design a near future CPU for their own 10nm, their own 7nm, and TSMC 5/7nm to give some coverage. I picked TSMC purely as an example, they could also go to Samsung for example. Probably not desperate enough to look at GoFlo unless it is for low end stuff. Speculation is that the external fabs might not be for CPU necessarily, but any of the other silicon products Intel are involved with.

[porina]

3 minutes ago, gabrielcarvfer said:

Oof. Are you sure on that one?

Look at the IPC of Sunny Cove as used in Ice Lake mobile. It provides tangible uplift over Skylake. I'd love to see it on desktop, but doesn't seem like we're getting it until Rocket Lake next year, even then on 14nm. They have Willow Cove as a successor in the pipeline. They're not standing still on the design side. For the majority here it isn't very visible since we mostly care about desktop, which seems to be Intel's lowest priority right now.

 

3 minutes ago, gabrielcarvfer said:

Pretty weird they didn't have multiple options to go with from the start. I guess shareholders were upset with the idea of spending more on R&D to maintain the monopoly Intel virtually had.

They probably thought they could either get 10nm working "good enough" and/or hold out for 7nm to return to the fight. Given the 1st still hasn't happened and isn't likely to, and the 2nd is now also delayed, they're finally looking at wider backup plans for the future. We don't know how long ago they started down that route, but it isn't a quick one and we're probably looking well into next year before we see anything new produced externally, if at all.

[Master Disaster]

30 minutes ago, porina said:

Look at the IPC of Sunny Cove as used in Ice Lake mobile. It provides tangible uplift over Skylake. I'd love to see it on desktop, but doesn't seem like we're getting it until Rocket Lake next year, even then on 14nm. They have Willow Cove as a successor in the pipeline. They're not standing still on the design side. For the majority here it isn't very visible since we mostly care about desktop, which seems to be Intel's lowest priority right now.

 

They probably thought they could either get 10nm working "good enough" and/or hold out for 7nm to return to the fight. Given the 1st still hasn't happened and isn't likely to, and the 2nd is now also delayed, they're finally looking at wider backup plans for the future. We don't know how long ago they started down that route, but it isn't a quick one and we're probably looking well into next year before we see anything new produced externally, if at all.

Meanwhile AMD is currently designing/fabbing 5nm and looking even beyond that.

 

This isn't how it was supposed to happen tbh, we all wanted AMD to give Intel a good kicking but at this point AMD landed the first punch and Intel has spent the rest of the fight lying on the ground punching themselves in the face. When one company gets too far ahead it only ever hurts consumers. I just hope AMD doesn't start doing the shady shit Intel used to.

[SansVarnic]

17 hours ago, iBabySlapper said:

Sounds like an awfully dangerous line to take in a double whammie of a presendential election and a market crash, but it's not my place to comment really.

 

I thing about the market is knowing when to quit not when to be squemish. I weigh the stock before I buy it along with economy and politics. My last major purchase was just before Trump was elected (all markets fall before a major election) and I am sitting pretty. My AMD stock I purchased for $10 a share before they introduced their new line of cpus is also doing quite nicely.

Also, with all investing diversifying is the how to go. Not all my eggs sit in Intel or AMD.  

17 hours ago, mr moose said:

most long term investments will have periods of loss,  the thing is balancing inflation against long term growth against fee's and tax against dividends. 

 

Absolutely, I make sure to try and plan for market drops. Selling because it falls doesn't mean much if it bounces back higher afterwards.

 

It is too difficult to sell and then buy quickly to catch that gain. I would rather not get the attention of the SEC, sometime it's better to take the short loss and get a short gain than appear to have "an edge" that will get you in trouble.

 

 

I think we are derailing so I'll stop here.

[mcbaes72]

Maybe they could double their prices to make up for lost revenue.

 

Oh wait, nvm.

[straight_stewie]

At this point Intel might do better to figure out what's coming after 5nm (which could very well be the last node shrink anyone does), and head straight for that.

It seems like they are going to be behind for the rest of the traditional processor technology manufacturing game.

[CTR640]

While this is bad news and probably for us as customers, Intel deserves it. They've been way too cocky, too arrogant, too ignorant, too proud, too greedy and too lazy it's time to get in deep trouble to wake up and start getting serious.

 

I was just wondering about something right now: what about Jim Keller? Does it has something to do with all this shit?

Jim left for personal or unknown reason so suddenly, not sure if there is a conflict between those all.