https://semiaccurate.com/2018/05/08/contra-revenue-comes-back-in-a-big-way-at-intel/
Looks like Intel's infamous "contra revenue" program is back. Thoughts?
Looks like Intel's infamous "contra revenue" program is back. Thoughts?
[SemiAccurate] Contra-revenue comes back in a big way at Intel
- Thread starter JDG1980
- Start date
Ashraf had a good piece on it back in 2016:
https://www.fool.com/investing/gene...s-contra-revenue-strategy-was-a-huge-was.aspx
https://www.fool.com/investing/gene...s-contra-revenue-strategy-was-a-huge-was.aspx
Basically Intel shoehorned high end chips into low to mid range platforms (tablets) which required expensive components to work and Intel subsidized the cost to the companies who marketed products using Intel's chips. It cost Intel and shareholders billions and they never made the money back because the products did not sell well. The goal was to establish long term relations with vendors producing tablets and to get Intel instruction sets working on Android. The hope was they could make up the cost in the long term, didn't work.
So unlike them APU laptops with single channel ram and bad components you get a good balanced system for cheaper then similar?!?!
So wait, are we against that and if yes than why?
Also isn't that exactly what AMD did with the custom made APUs for all the consoles?
I don't know if it was ever proven but everybody was saying that they were barely making any money from them.
So wait, are we against that and if yes than why?
Also isn't that exactly what AMD did with the custom made APUs for all the consoles?
I don't know if it was ever proven but everybody was saying that they were barely making any money from them.
Yes, we are against it, for the same reason we are against any dumping policy: it's meant to drive out competition through sheer force of capital, and when the storm clears and competition is in disarray the consumer ends up paying the bill anyway. Who do you think pays for that contra revenue, Intel shareholders or Intel consumers?
That having been said, I'm not holding my breath for the new contra-revenue scheme coming back based on the SA article quoted above.
That having been said, I'm not holding my breath for the new contra-revenue scheme coming back based on the SA article quoted above.
It was more atom related and specifically atom vs arm, than big core vs amd.
Aka a software ecosystem was developed that did not need x86. This software ecosystem allowed non x86 cpus. These non x86 cpus (arm cpus) had much cheaper motherboards, and other parts of the builds of materials for more stuff was on the cpu die, and the stuff that was not on the cpu die was cheaper to make off the motherboard, and the motherboard did not have to be as thick.
Aka everything that was not x86 and not arm cpus, but was platform related was much cheaper on the arm platform than the x86 platform. And this made the total cost of ownership to the oems cheaper for arm vs x86. Futhermore the OEM had the flexibility to choose partners in a way that gave the OEM freedom and Intel hated this for you get market pressure downward for producer surplus and it is a race to the bottom, while increasing the OEM's consumer surplus.
So Intel's strategy was to subsidies the platform and other things in order to be cost competitive, take billions of dollars losses now for the hope of billions of dollars of profit later. The problem of why this did not work is that the system is multifactoral it was not just price that allowed ARM to be chosen rather than Intel it was also the freedom that ARM allowed. Intel has to have a massive advantage in something in order to mitigate this freedom angle.
Aka a software ecosystem was developed that did not need x86. This software ecosystem allowed non x86 cpus. These non x86 cpus (arm cpus) had much cheaper motherboards, and other parts of the builds of materials for more stuff was on the cpu die, and the stuff that was not on the cpu die was cheaper to make off the motherboard, and the motherboard did not have to be as thick.
Aka everything that was not x86 and not arm cpus, but was platform related was much cheaper on the arm platform than the x86 platform. And this made the total cost of ownership to the oems cheaper for arm vs x86. Futhermore the OEM had the flexibility to choose partners in a way that gave the OEM freedom and Intel hated this for you get market pressure downward for producer surplus and it is a race to the bottom, while increasing the OEM's consumer surplus.
So Intel's strategy was to subsidies the platform and other things in order to be cost competitive, take billions of dollars losses now for the hope of billions of dollars of profit later. The problem of why this did not work is that the system is multifactoral it was not just price that allowed ARM to be chosen rather than Intel it was also the freedom that ARM allowed. Intel has to have a massive advantage in something in order to mitigate this freedom angle.
Ah ok thanks, that makes sense actually.
Still ARM is going the more cores road for years now which means that individual cores can't be made much faster/better and there will come a time where ARM just won't be able to cut it anymore,intel has to keep up with developing tablets -even at a big loss now- because they have to be ready when the time comes,for example 3dxpoint is a part of that, the dimm modules will be able to be used as main ram with no "extra" "real" ram so it would cut costs in such small devices.
Still ARM is going the more cores road for years now which means that individual cores can't be made much faster/better and there will come a time where ARM just won't be able to cut it anymore,intel has to keep up with developing tablets -even at a big loss now- because they have to be ready when the time comes,for example 3dxpoint is a part of that, the dimm modules will be able to be used as main ram with no "extra" "real" ram so it would cut costs in such small devices.
AMD is making money on the consoles because of the volume. The margins themselves are very low, single digit%. Like say they sold the APU for $100 but it cost $92 for GloFo to fab it.
Just to make it obvious, Contra-Revenue (at least during the Atom era) was like OEMs got the chip for free AND got a subsidy because the platform costs even with a free chip wasn't enough to be competitive with ARM. I have a hard time believing this is true because it got really expensive for Intel although I do think OEMs with Chromebooks are getting a very good deal.
Just to make it obvious, Contra-Revenue (at least during the Atom era) was like OEMs got the chip for free AND got a subsidy because the platform costs even with a free chip wasn't enough to be competitive with ARM. I have a hard time believing this is true because it got really expensive for Intel although I do think OEMs with Chromebooks are getting a very good deal.
I highly doubt they regressed backwards to low single digit %. They started at low double digit and grew to >15% and expected to hit 20%. Even with half gen consoles I highly doubt AMD would accept a lower margin than what they grew.
https://www.kitguru.net/components/...s-is-thriving-margins-set-to-grow-to-over-20/
https://www.kitguru.net/components/...s-is-thriving-margins-set-to-grow-to-over-20/
The original contra revenue program used Atom chips, not "high end".
The problem was that Intel wanted to capture a significant portion of the Tablet pie, and that meant getting their chips into $100-150 Tablets. I think Intel realized that with the pace of developments on the ARM devices they'd have a hard time capturing high end so cost became important.
Officially the contra revenue program was a $20 subsidy(offsetting the cost of the chip) to develop Tablets based on their platforms because the PCBs that used Atom chips used components and methodologies that were supposedly $20 more than comparable Atom PCBs. It was supposed to halve on the Airmont(14nm) generation and fully disappear with Broxton
Unofficially they were also pouring money into making reference platforms that basically fit the needs of the many smaller OEMs, which is why we saw so many Atom Tablets come in such a short time from unknown Chinese brands. What many of us understood was Intel was basically making a Tablet for you, and you'd only need resources for rest, like marketing and support. Which was controversial in itself because they weren't just giving subsidies for the chip, but giving you the whole device.
This slide talks about cost reductions that were going to come to Broxton(instead changed to Apollo Lake): https://images.anandtech.com/doci/10635/intel_apollo_lake_atom_1.png
The problem was that Intel wanted to capture a significant portion of the Tablet pie, and that meant getting their chips into $100-150 Tablets. I think Intel realized that with the pace of developments on the ARM devices they'd have a hard time capturing high end so cost became important.
Officially the contra revenue program was a $20 subsidy(offsetting the cost of the chip) to develop Tablets based on their platforms because the PCBs that used Atom chips used components and methodologies that were supposedly $20 more than comparable Atom PCBs. It was supposed to halve on the Airmont(14nm) generation and fully disappear with Broxton
Unofficially they were also pouring money into making reference platforms that basically fit the needs of the many smaller OEMs, which is why we saw so many Atom Tablets come in such a short time from unknown Chinese brands. What many of us understood was Intel was basically making a Tablet for you, and you'd only need resources for rest, like marketing and support. Which was controversial in itself because they weren't just giving subsidies for the chip, but giving you the whole device.
This slide talks about cost reductions that were going to come to Broxton(instead changed to Apollo Lake): https://images.anandtech.com/doci/10635/intel_apollo_lake_atom_1.png
Interesting. So what is Intel throwing their money at this time? If they want to throw their money and subsidize an 8 core gaming chip for us, I'd be cool with that. I'm not kidding. If they want to really stick it to AMD, they can give us enthusiasts a new 8 core coffee lake and Z390 combo for half price. That makes sure we don't end up with AMD stuff in our box. If they have any other plans for how to buy influence with their money, I'm not interested.
Let me Guess,
Name me one segment, or customer, that they got pissed off so bad it needs this contra revenue?
I dont see DC need any of this. Intel already does very well with all cloud customer and has plenty of margin to play with, with no immediate threat.
PC, Laptop? They are milking it, and it is not like AMD has made another Athlon moment yet.
One possible target is Apple. Which is threatening the transition to ARM. They had their whole product lineup planned, Mac Pro, iMac, MacBook, all aligned with what Intel was suppose to deliver it to them last year. And not only is it not coming this year. They will have to wait for another year. Talk about a 2-3 year delay in product development for Apple.
Another one is Intel brute force entry into GPU. Assuming Intel need 20% more die size because of Software / Drivers / Hardware inefficiency compared to its competitor, Intel will also need to sell it cheaper then its competitor.
But do we need contra revenue if it is a new segment and there is nothing to hide?
Modem, Intel is trying to break into the 4/5G modem space. And its current price just isn't attractive enough to compete with Qualcomm.
Name me one segment, or customer, that they got pissed off so bad it needs this contra revenue?
I dont see DC need any of this. Intel already does very well with all cloud customer and has plenty of margin to play with, with no immediate threat.
PC, Laptop? They are milking it, and it is not like AMD has made another Athlon moment yet.
One possible target is Apple. Which is threatening the transition to ARM. They had their whole product lineup planned, Mac Pro, iMac, MacBook, all aligned with what Intel was suppose to deliver it to them last year. And not only is it not coming this year. They will have to wait for another year. Talk about a 2-3 year delay in product development for Apple.
Another one is Intel brute force entry into GPU. Assuming Intel need 20% more die size because of Software / Drivers / Hardware inefficiency compared to its competitor, Intel will also need to sell it cheaper then its competitor.
But do we need contra revenue if it is a new segment and there is nothing to hide?
Modem, Intel is trying to break into the 4/5G modem space. And its current price just isn't attractive enough to compete with Qualcomm.
Seeing how nVIDIA is selling well the Switch and how the next ARM core from them are dangerously great, they will allow it... but not alone. Maybe GloFo (or maybe is time to add Samsung to the equation?) will help to take some profit loss in order to cancel nVIDIA and TSMC advances.
BTW... Intel should continued with Broxton... was the X86 counterpart of Qualcomm Kryo... that along better GPU cores, would make a decent deal.
BTW... Intel should continued with Broxton... was the X86 counterpart of Qualcomm Kryo... that along better GPU cores, would make a decent deal.
In principle, I would normally be against contra revenue. But in the case of Atom, it was beneficial to the consumer and fostered competition because they were trying to introduce an alternative to ARM. Some of the early cheap atom tablets were actually quite decent for the price, and I am sorry they basically disappeared from the market. It is not like intel had control of the market and were using contra revenue to drive out competitors, it was quite the opposite. It turned out to be a terrible business decision though, because they dumped billions of dollars and ended up abandoning the market anyway.
BTW, did anyone actually get beyond the paywall and see what Charlie was saying? The part that could be viewed free actually said nothing. Like some other posters, unless confirmed by other sources, I am very skeptical that Intel would get involved in contra revenue again.
BTW, did anyone actually get beyond the paywall and see what Charlie was saying? The part that could be viewed free actually said nothing. Like some other posters, unless confirmed by other sources, I am very skeptical that Intel would get involved in contra revenue again.
The giving away $20 per platform can be viewed as anti-competitive, but since it was supposed to go away, and they already put all that effort in along with the cash, they should have continued it.
The only problem is Intel viewed Broxton as a "hero device", something that would have been top notch in performance. In reality, it would have been #2 next to SD835, which probably became the motive to cancel Broxton.
By the way, Intel didn't discontinue development on x86 Android. It continued on the small IoT devices and automotive Atom chips, like the E3900.
The only problem is Intel viewed Broxton as a "hero device", something that would have been top notch in performance. In reality, it would have been #2 next to SD835, which probably became the motive to cancel Broxton.
By the way, Intel didn't discontinue development on x86 Android. It continued on the small IoT devices and automotive Atom chips, like the E3900.
| Thread starter | Similar threads | Forum | Replies | Date |
|---|---|---|---|---|
|
[SemiAccurate] AMD to differentiate cores | CPUs and Overclocking | 107 | |
| H | News AMD meets EPS, misses on revenue | CPUs and Overclocking | 45 | |
| H | News Intel beats EPS but misses on revenue in 4Q17 | CPUs and Overclocking | 60 | |
| J | Intel Q3: Revenue up 19% | CPUs and Overclocking | 21 | |
|
[SemiAccurate] Intel kills off the 10nm process | CPUs and Overclocking | 276 |
Similar threads
-
-
-
News Intel beats EPS but misses on revenue in 4Q17
- Started by Hitman928
- Replies: 60
-
-
TRENDING THREADS
-
[Rumor, Tweaktown] AMD to launch next-gen Navi graphics cards at E3
- Started by Hitman928
- Replies: 560
-
Question Decent, cheap 1TB sata ssds for raid storage
- Started by imported_dgingeri
- Replies: 3
-
-
-