Brian Krzanich just killed the PC

[WaitingForNehalem]

They've been saying this nonsense for years now. Good luck competing against TI, NXP, Infineon, STMicro...you know the companies that have been doing this stuff for decades. It's going to be hard switching from fat margin $5000 Xeons to $0.30 micro-controllers...

 

[moonbogg]

They can quit after my next upgrade. No one will need a better desktop CPU than skylake-E...like ever. That will be the last real chip ever made. That chip will be closest to the top when archaeologists of the future dig up this world's remains for study. "Ah, look here. Skylake-E was as far as they got".

 

[Phynaz]

They've been saying this nonsense for years now. Good luck competing against TI, NXP, Infineon, STMicro...you know the companies that have been doing this stuff for decades. It's going to be hard switching from fat margin $5000 Xeons to $0.30 micro-controllers...

They want to sell more Xeons - that's the cloud piece.

 

[poofyhairguy]

I''m sure smart oven etc makers would be so over giving Intel their 60% chip margins instead of some priced to free ARM solution, assuming consumers even want that sort of end product to begin with.

That is what I don't get.

I mean, Intel failed in the mobile market that was filled with $700 devices (aka devices that could absorb their margins). Instead Apple and Samsung decided to eat those margins by themselves.

IOT is about making almost all devices internet connected, basically everything gets a screen less smartphone shoved inside. But consumers still care about price and non IOT version will still be available so cost will matter more to IOT adoption than it did mobile phone adaption (which was a device that justified it's own purchase by itself).

I can't see how any rational company can look at IOT and see:

1. A lot of margin

2. A clear competitive advantage for X86

There will be a lot of volume sure, as each device needs its own SoC. But each SoC will make very minimal margin.

Is this Intel announcing that they plan to move to a high volume low margin business? And if so why didn't they do that years ago to compete in phones? It seems whatever would help them in phones would help in IOT, I don't even see the point of considering it a separate product strategy.

It almost feels like they are lying to themselves and hopes no one calls then out on it.

 

[PingSpike]

That is what I don't get.

I mean, Intel failed in the mobile market that was filled with $700 devices (aka devices that could absorb their margins). Instead Apple and Samsung decided to eat those margins by themselves.

IOT is about making almost all devices internet connected, basically everything gets a screen less smartphone shoved inside. But consumers still care about price and non IOT version will still be available so cost will matter more to IOT adoption than it did mobile phone adaption (which was a device that justified it's own purchase by itself).

I can't see how any rational company can look at IOT and see:

1. A lot of margin

2. A clear competitive advantage for X86

There will be a lot of volume sure, as each device needs its own SoC. But each SoC will make very minimal margin.

Is this Intel announcing that they plan to move to a high volume low margin business? And if so why didn't they do that years ago to compete in phones? It seems whatever would help them in phones would help in IOT, I don't even see the point of considering it a separate product strategy.

It almost feels like they are lying to themselves and hopes no one calls then out on it.

As you said, x86 has no advantage here. They don't need windows on their smart toilet. They don't need performance on their smart toilet. They don't need what Intel has. If they even have a plan, I'd imagine its going to be to dump product with $20 bills attached to try and snatch up the market and gain a monopoly position whereupon I guess they'll start jacking up the price to get those fat margins back. Assuming the market even exists (it doesn't IMO, IoT is manufacturers looking to get rid of what they have not build what the market asked for) and that plan worked, it will probably be just as janky and fragmented as Smart TVs. There won't be any vendor lockin because LG Smart toaster OS 3 will just be left broken to die on the vine with a thousand other variants. When Intel turns the screws they'll be left on the same vine since no one has a buy in and no one needs a quad core Xeon on their Smart Fridge ad blaster.

 

[sm625]

can't wait for someone to hack my thermostat and crank the heat up to 10000...

There isnt much motive to drive your thermostat to max. What they will do is hack your thermostat to deduce when you're not home.

 

[RussianSensation]

This post nailed it.

I don't know if Sandy Bridge's time is up now, but for me it was up when 22nm IB was released. At least 22nm IB gave a decent power reduction for same performance over SB.

But this 14nm Skylake launch just has me puzzled (business-wise, not technology-wise). Its like 14nm gave Skylake no benefit over 22nm Haswell. All the performance gains are architectural, designed in by some hard working engineers.

In the meantime, 14nm looks to be basically 100% cost-reduction focused (areal shrinkage maximized, electrical parametrics be damned) versus 22nm and 32nm.

So here is my half-baked assessment at this point - Whether or not Sandy Bridge is made obsolete by Skylake, the progression of diminishing returns with Intel's Tick-Tock tells me, as a consumer, that I may as well buy Skylake now because there is pretty much no damn good reason to hold out for any future Intel processor(s) as they are also just as likely to deliver, at best, a solid 5% performance bump in 1, 2, or 3 years time.

Intel's Skylake is telling us that this is not as good as it is going to get, but at the same time they are communicating that 10nm and 7nm are going to give us CPUs that might be, if we are lucky, another 10-20% over what Skylake is giving us today.

Buy one now, log out of your Anandtech forums account, and don't worry about the semiconductor world for at least 5 years. Save yourselves hundreds, if not thousands, of hours to go do something more productive with your lives in the meantime. Maybe 5nm will come along in 2022 and give us a grand total of 25% improvement in performance/watt over Skylake, then its time to consider an upgrade.

If we assume 15% increase in IPC for Icelake in 2018, and 15% increase in IPC for next major Intel architecture in 2021, that's 32% IPC over the next 6 years since Skylake launched in 2015. Let's assume the 2021 CPU can overclock to 5.3Ghz over a 4.8Ghz Skylake, that's another 10.4%. That means in a very optimistic scenario, single threaded performance may grow 46% from Fall 2015 to Fall 2021. It might actually be less since Intel is now changing the strategy and focus on non-traditional PC segments.

"What might be called the “new” Intel will be built upon five pillars, Krzanich said:

The cloud—including servers, data centers, and virtualization
- Connected “things,” such as sensors, autonomous vehicles, or PCs
- An evolving memory business, from 3D XPoint memory to advances in server and data center infrastructure
- Connectivity, specifically 5G networking
- Manufacturing and the underlying fab technology.

About 40 percent of Intel’s revenue and 60 percent of its profit margin already come from outside the PC, Krzanich said last week, when the company began publicly signalling its new focus. “It’s time to make this transition and to push the company over all the way to that strategy and that strategic direction,” Krzanich said then. “That’s why we wanted to do it now.”

I have a feeling Icelake will not get 15% increase in IPC over Skylake and neither will the 2021 CPU over Icelake. I am predicting 10-12% per each new major architecture (2018, 2021), maybe less. The die size savings will be used to increase the useless iGPU component instead and we might eventually get a 6-core mainstream chip, possibly with Cannonlake or Icelake but that assumes Zen/Zen+ are competitive.

 

[cytg111]

Yea IDC was spot on.
So, summa summarum;
- Intel is betting the farm on IoT and Servers
- AMD is betting the farm on VR
Should we place our bets? I think Intel will be crawling back on those words (I hope).

 

[dave_the_nerd]

- Intel is betting the farm on IoT and Servers.

I kind of object to this characterization.

Intel would be a poorly run company if, with as much clout and money as they have, they didn't hedge their bets and continue to (at least try to) offer compelling products in a variety of categories as long as those product lines are still making money.

 

[Arachnotronic]

Yea IDC was spot on.
So, summa summarum;
- Intel is betting the farm on IoT and Servers
- AMD is betting the farm on VR
Should we place our bets? I think Intel will be crawling back on those words (I hope).

AMD is betting the farm on servers.

 

[Cogman]

So, honestly I think that intel is really just opening itself up to competition in the consumer PC sector. They aren't going to pull out of it, but they may pull back developement. That, I think, will open the path for AMD or some ARM producer to start putting out PC CPUs and devices with comparable performance to current intel offerings.

It won't happen this year, or the next year, probably around the next 5 years.

What needs to change is some other OS, besides windows, needs to take over the consumer PC market. Windows could, but people expect backwards compatibility with all their applications.

 

[Arachnotronic]

So, honestly I think that intel is really just opening itself up to competition in the consumer PC sector. They aren't going to pull out of it, but they may pull back developement. That, I think, will open the path for AMD or some ARM producer to start putting out PC CPUs and devices with comparable performance to current intel offerings.

It won't happen this year, or the next year, probably around the next 5 years.

Intel isn't going to invest less in the core IPs aimed at PCs (graphics, CPU, memory controller, ISP, etc.).

 

[cytg111]

AMD is betting the farm on servers.

Yes, forgot about Zen (yea right..)

Anyway I was reading this and thought it made sense;

http://arstechnica.co.uk/gadgets/2016/04/amd-focusing-on-vr-mid-range-polaris/

AMD is betting on VR, and it's betting hard. Compelling content for VR isn't quite here yet—"we don't have something that's so compelling that you just can't wait to get home, or take days off sick, to play," says Taylor—but when it does, and when a regular, mainstream consumer goes to an electronics store buy a PC to get that VR experience, AMD wants it to be powered by its GPU at an accessible price. It's a big gamble, one that depends on numerous stakeholders each playing their part. That's not to mention that just because Nvidia is rumoured to be launching a high-end part now, it can't release a high-powered mainstream part in the future—or simply just drop the price of its high-end cards.

(edit : is problary at part for V&G section)

 

[Exophase]

I don't know why Intel seems so confident about their IoT position. They're going up against a ton of microcontroller vendors that have been integrating stuff for this purpose for a long time and have a lot of market establishment. Their manufacturing advantage doesn't apply as much to devices of this nature either. Their current microcontroller-class CPU design (Quark) is a joke.

They do have the radio technology, but the wifi stuff isn't that hard to come by and AFAIK other microcontroller manufacturers have been offering packages with integrated wifi. I don't as of yet expect a big boom of every tiny device connecting directly to the internet, but I guess we'll see.

 

[Arachnotronic]

I don't know why Intel seems so confident about their IoT position. They're going up against a ton of microcontroller vendors that have been integrating stuff for this purpose for a long time and have a lot of market establishment. Their manufacturing advantage doesn't apply as much to devices of this nature either. Their current microcontroller-class CPU design (Quark) is a joke.

They do have the radio technology, but the wifi stuff isn't that hard to come by and AFAIK other microcontroller manufacturers have been offering packages with integrated wifi. I don't as of yet expect a big boom of every tiny device connecting directly to the internet, but I guess we'll see.

I think that people are misunderstanding what Intel is talking about when they say IoT. They aren't going after micro-controllers, that's not their bag.

They are talking about things like Point of Sale terminals, In-Vehicle Infotainment, etc. Anything that's not a PC where they can stick an Atom or a Core processor.

Their revenues in this business are quite substantial, well north of $2 billion per year. To put this into perspective, Intel's IoT group does more annual revenue than Microchip (pre-Atmel buy) does or, before Microchip bought it, Atmel did. Margins in this business are also quite fat.

 

[frozentundra123456]

Yea, but the problem I see in the IoT is that intel is too entrenched in x86. Their core line has good performance, but really, except for the ability to run x86, ARM has cheaper and at least good enough if not better chips than intel for almost anything that does not require x86. Plus they are willing to take much smaller margins.

As for AMD if Zen is good, they could take a decent server market and if they can get HBM APUs at a reasonable price, they could do well in that area, as well as dgpus. The advantage AMD has, is that they dont have to gain much to get a significant increase in market share and maybe even make a profit.

VR could be a profitable area as well, but I think it will remain a niche, unless they can bring the price down and somehow eliminate the headsets.

 

[Exophase]

I think that people are misunderstanding what Intel is talking about when they say IoT. They aren't going after micro-controllers, that's not their bag.

So what's the point of Quark D1000 and D2000? Those are as bog standard as you get when it comes to low end microcontrollers.

They are talking about things like Point of Sale terminals, In-Vehicle Infotainment, etc. Anything that's not a PC where they can stick an Atom or a Core processor.

Has the in-vehicle stuff seen a lot of growth? It seems like every last ARM SoC maker that failed to make it big in phones transitioned here (nVidia, TI, Freescale) and they all claim they're doing well in it.

I always figured Point of Sale devices these days were basically PC hardware running Windows anyway.

 

[Qwertilot]

If you're into 'bigger' IoT style things don't forget robotics as a medium term sort of target (and perhaps ultimately a truly vast market.).

Self driving cars and things too of course.

 

[CHADBOGA]

Krzanich may well turn out to be Intel's Hector Ruiz.

 

[Arachnotronic]

Krzanich may well turn out to be Intel's Hector Ruiz.

Not that bad, Ruiz is a special kind of awful, but this guy is pretty out of his depth, IMO. He's basically a Rory Read. Looking forward to seeing him replaced ASAP. Terrible CEO, probably worse than Craig Barrett.

 

[HiroThreading]

Not that bad, Ruiz is a special kind of awful, but this guy is pretty out of his depth, IMO. He's basically a Rory Read. Looking forward to seeing him replaced ASAP. Terrible CEO, probably worse than Craig Barrett.

Agreed.

A real shame that they basically forced Pat Gelsinger out because of Larrabee. I think he would have made a great CEO at Intel.

 

[HiroThreading]

Consumers don't decide what they want or need, they're mindless drones. People don't need PC's any more, because Apple, Samsung etc. are telling them they don't need PC's (and have been doing so for years).

People want what's "in" and being marketed to them. Nobody wanted the iPhone before they knew about it, but Apple made people want it. Now Intel is trying to make people want the Internet of Things.

Such hubris...

Suppose we go with your hypothesis, I suppose that your entire choice or interest in gaming and hardware is entirely synthetic, and you -- as well as everyone on this forum -- are a mindless drone?

Most studies into consumer behaviour have debunked this myth. It can certainly apply in generalities or greater societal institutions (school, church, work and etc), but not at this kind of small scale.

 

[Arachnotronic]

Agreed.

A real shame that they basically forced Pat Gelsinger out because of Larrabee. I think he would have made a great CEO at Intel.

I like your thinking.

 

[Systems analyst]

ARM provided the IP for the original digital mobile phone, and they have led the field ever since; Intel did not respond, until it was too late, and even then, they were half-hearted about it; mobile computing is converting to ARM SoCs, and Android or iOS; the end result is the pricing-pressure on PCs, and falling sales of Wintel.

Intel and Microsoft seem to be in a deadly embrace 'like two spent swimmers that do cling together, and choke their art.' It may be too late to save that alliance, and by extension, to save Windows and Intel PC business.

It just seems to be the trend in computing; you could 'blame' ARM or Intel or Microsoft, but the market makes the decision as to which survives.

 

[master_shake_]

There isnt much motive to drive your thermostat to max. What they will do is hack your thermostat to deduce when you're not home.

also bad.