By: David Kanter (dkanter.delete@this.realworldtech.com), January 29, 2013 11:51 pm
Room: Moderated Discussions
Richard Cownie (tich.delete@this.pobox.com) on January 29, 2013 6:14 pm wrote:
> bakaneko (nyan.delete@this.hyan.wan) on January 29, 2013 2:18 pm wrote:
>
> > Sorry, I got hotblooded and gave you a wrong
> > answer to a misunderstood question:
> >
> > (there was more in my posts, but meh, too lazy)
>
> Yup, I enjoyed the rant :-)
>
> > The question is why would you replace Intel with
> > a group of ARM based suppliers if the price is
> > the same for similar products?
>
> If we widen the definition of price to "total cost of
> ownership per unit of throughput", then the answer is,
> you wouldn't change. It just seems likely that the
> people who are building low-power quad-core ARM SoC's
> to go into tens of millions of sub-$200 tablets will also
> have a chance at making some neat cheap ultra-dense low-
> power servers.
It's possible, but tablets are quite different animals.
> > Alternatively: why wouldn't they need the high
> > margin (not want, NEED) and could lower the
> > prices?
>
> They wouldn't *need* the kind of margins Intel has because
> a) they have businesses designed to survive on the high
> volumes and razor-thin margins of the smartphone/tablet
> business,
Which is fine, except ARM-based servers won't have high volumes --> you need fat margins. It's one thing to spend $100M developing a chip where you sell 300M units. It's quite different when you are eagerly hoping to sell 100K units.
> b) the ARM ecosystem works by having fab
> expertise developed by foundries, "standard" core designs
> and various other useful IP by ARM plc, a bunch of free
> or cheap software (gcc, linux etc). So the barriers to
> entry are rather low.
Honestly, the trend is away from standard ARM cores. Also, there's a lot of other IP that servers need that isn't in the normal tablet/phone space, like high performance wired networking and PCI-E.
> > And if they need it, why would they have lower
> > production costs and thus could lower the prices?
>
> They wouldn't have lower production costs. Intel has the
> best fabs; and on top of that the foundries want a profit.
> But most (maybe all) ARM SoC vendors would be willing to live
> with much lower margins than Intel currently gets for server cpus, which
> could translate into lower prices.
The right way to think about this is to compare Intel's margins to TSMC+SoC vendor. I think TSMC gross margins are somewhere around 50%. SoC vendors get away with 20-30% gross margins, but only on super high volume products. For a lower volume product, they probably want a lot higher than that. Also, you need to pay for some of the 3rd party IP, e.g. ARM licenses.
That's starting to look like pretty fat margins.
> I don't believe that Intel is selling server cpu's at the marginal
> cost of production. Do you ?
: )
David
> bakaneko (nyan.delete@this.hyan.wan) on January 29, 2013 2:18 pm wrote:
>
> > Sorry, I got hotblooded and gave you a wrong
> > answer to a misunderstood question:
> >
> > (there was more in my posts, but meh, too lazy)
>
> Yup, I enjoyed the rant :-)
>
> > The question is why would you replace Intel with
> > a group of ARM based suppliers if the price is
> > the same for similar products?
>
> If we widen the definition of price to "total cost of
> ownership per unit of throughput", then the answer is,
> you wouldn't change. It just seems likely that the
> people who are building low-power quad-core ARM SoC's
> to go into tens of millions of sub-$200 tablets will also
> have a chance at making some neat cheap ultra-dense low-
> power servers.
It's possible, but tablets are quite different animals.
> > Alternatively: why wouldn't they need the high
> > margin (not want, NEED) and could lower the
> > prices?
>
> They wouldn't *need* the kind of margins Intel has because
> a) they have businesses designed to survive on the high
> volumes and razor-thin margins of the smartphone/tablet
> business,
Which is fine, except ARM-based servers won't have high volumes --> you need fat margins. It's one thing to spend $100M developing a chip where you sell 300M units. It's quite different when you are eagerly hoping to sell 100K units.
> b) the ARM ecosystem works by having fab
> expertise developed by foundries, "standard" core designs
> and various other useful IP by ARM plc, a bunch of free
> or cheap software (gcc, linux etc). So the barriers to
> entry are rather low.
Honestly, the trend is away from standard ARM cores. Also, there's a lot of other IP that servers need that isn't in the normal tablet/phone space, like high performance wired networking and PCI-E.
> > And if they need it, why would they have lower
> > production costs and thus could lower the prices?
>
> They wouldn't have lower production costs. Intel has the
> best fabs; and on top of that the foundries want a profit.
> But most (maybe all) ARM SoC vendors would be willing to live
> with much lower margins than Intel currently gets for server cpus, which
> could translate into lower prices.
The right way to think about this is to compare Intel's margins to TSMC+SoC vendor. I think TSMC gross margins are somewhere around 50%. SoC vendors get away with 20-30% gross margins, but only on super high volume products. For a lower volume product, they probably want a lot higher than that. Also, you need to pay for some of the 3rd party IP, e.g. ARM licenses.
That's starting to look like pretty fat margins.
> I don't believe that Intel is selling server cpu's at the marginal
> cost of production. Do you ?
: )
David
