Not too far from your one month approximation, I wondered if launch will correlate with IDF (late August just a tad over a month) putting us in Q4 territory.
You're right about the date, for some reason I got early August and early September mixed up. So the launch is about a month and a half out. But even then early September is not Q4, it's squarely Q3.
I am curious, where all intel has a role to play in Internet of Things. IoT is a low power processing market, largely dominated by ARM and its partners. $559 million revenue of Intel in IoT is larger than ARM's complete revenue per quarter!! Can someone explain this?
ARM doesn't make anything. Their money comes from licensing, so for each $10 ARM chip that sells, they get a small percentage. For each $10 intel chip that sells, intel gets $10.
Better to compare profits than revenue for such things.
The IoT from Intel is devided in several business devisions. You have to see it bigger then just consumer devices. See more specific needs to gateway networking, sensors etc, all these types of modules that can be used to collect data, share data, connect globally are now assembled in the IoT.
What Intel defines as IoT is not what you and most people think of when hearing IoT. The bulk of this segment is what used to be called "embedded", they just renamed it because IoT is fashionable. So slot machines and cars and ATMs and commercial displays and w/e they don't include in Client ( can't remember if set top boxes are in Client for sure but i think they aren't in IoT).
Tick-Tock is moving to 2.5 years instead of 2 year. In 2016H2 Intel are introducing a 3rd 14nm product: Kaby Lake. 10nm manufacturing (Cannonlake) has been delayed until 2017H2.
I've been saying this for some time, but people just hated to hear it. Experts have been saying that 14nm was very hard, as we've been seeing. They've also been saying that 10nm would be extremely hard, and that 7nm will be extraordinarily hard.
They are also saying that 5nm may be impossible, despite that Intel has shown it on their roadmap. I assume it's there for marketing and investor reasons. They aren't going to want to see it end just two more bumps down the road.
I'm hopeful about EUV. GAA (Gate-all-around) may also improve things. With multigates, you don't have to make the transistors themselves smaller to get more logic per square meter.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
15 Comments
Back to Article
K_Space - Wednesday, July 15, 2015 - link
So Skylake Q4 2015?dragonsqrrl - Thursday, July 16, 2015 - link
Skylake is launching in less than month.K_Space - Friday, July 17, 2015 - link
Not too far from your one month approximation, I wondered if launch will correlate with IDF (late August just a tad over a month) putting us in Q4 territory.dragonsqrrl - Friday, July 17, 2015 - link
You're right about the date, for some reason I got early August and early September mixed up. So the launch is about a month and a half out. But even then early September is not Q4, it's squarely Q3.K_Space - Friday, July 17, 2015 - link
Damn right it is Q3! And there we have why I don't work for intel.karthik.hegde - Thursday, July 16, 2015 - link
I am curious, where all intel has a role to play in Internet of Things. IoT is a low power processing market, largely dominated by ARM and its partners. $559 million revenue of Intel in IoT is larger than ARM's complete revenue per quarter!! Can someone explain this?Thermogenic - Thursday, July 16, 2015 - link
ARM doesn't make anything. Their money comes from licensing, so for each $10 ARM chip that sells, they get a small percentage. For each $10 intel chip that sells, intel gets $10.Better to compare profits than revenue for such things.
karthik.hegde - Thursday, July 16, 2015 - link
That's true, operating margin of ARM is high, although revenue is less.I'm wondering where Intel fits in IoT. They are losing in billions trying to enter mobile market. How do they have a foot hold in IoT?
duploxxx - Thursday, July 16, 2015 - link
The IoT from Intel is devided in several business devisions. You have to see it bigger then just consumer devices. See more specific needs to gateway networking, sensors etc, all these types of modules that can be used to collect data, share data, connect globally are now assembled in the IoT.example:
http://www.theverge.com/2015/1/6/7505423/ces-2015-...
http://www.intel.com/content/www/us/en/internet-of...
jjj - Thursday, July 16, 2015 - link
What Intel defines as IoT is not what you and most people think of when hearing IoT.The bulk of this segment is what used to be called "embedded", they just renamed it because IoT is fashionable. So slot machines and cars and ATMs and commercial displays and w/e they don't include in Client ( can't remember if set top boxes are in Client for sure but i think they aren't in IoT).
jibberegg - Thursday, July 16, 2015 - link
Tick-Tock is moving to 2.5 years instead of 2 year. In 2016H2 Intel are introducing a 3rd 14nm product: Kaby Lake. 10nm manufacturing (Cannonlake) has been delayed until 2017H2.ant6n - Thursday, July 16, 2015 - link
Weird that the article mentioned none of this.dragonsqrrl - Thursday, July 16, 2015 - link
There's a whole article dedicated to it:http://www.anandtech.com/show/9447/intel-10nm-and-...
melgross - Thursday, July 16, 2015 - link
I've been saying this for some time, but people just hated to hear it. Experts have been saying that 14nm was very hard, as we've been seeing. They've also been saying that 10nm would be extremely hard, and that 7nm will be extraordinarily hard.They are also saying that 5nm may be impossible, despite that Intel has shown it on their roadmap. I assume it's there for marketing and investor reasons. They aren't going to want to see it end just two more bumps down the road.
I guess we'll continue thin in the next article.
ant6n - Thursday, July 16, 2015 - link
I'm hopeful about EUV.GAA (Gate-all-around) may also improve things.
With multigates, you don't have to make the transistors themselves smaller to get more logic per square meter.