MRVL - Marvell Technology

[muscleman]

Are you guys worried about Intel?  (I am long Intel.)  Intel has a moat in the semiconductor business because its scale allows it to build cutting-edge fabs.  Its chips will have lower power consumption and/or better performance than everybody else.  If you look at Intel versus AMD... it's not a fair fight.  So Intel taking on all the smartphone SoC manufacturers is going to be scary.  If you look at reviews of Intel's smartphone chip in the Xolo (and there's another phone), performance and power consumption are middle of the pack right now. 

 

http://semiaccurate.com/2012/04/26/how-well-does-intels-new-phone-work-as-a-phone/

 

Eventually the software ecosystem will move away from having to do ARM emulation (emulation is slow)... this will increase the performance of Intel chips on Android phones.  And Intel will start moving its smartphone chips onto fabs that are a step ahead of the competition.  This will not be a fair fight.

 

*Historically...

Intel took huge market share away from the RISC instruction set chip manufacturers.  PowerPC, Cray, SGI, etc.

Intel didn't kill off Xilinx and Altera.  Their software design tools are critical and difficult to build.

Intel didn't kill Nvidia and ATI.  Graphics chips are hard to do.  Intel is doing ok with its integrated graphics on lower-end processors.

Intel has always enjoyed a huge advantage over AMD.

 

What is the upside for INTC?

I see that ARM is a hot stock and it is making the chips for the smartphones. But for 2012 its net income is only $165M. If that is all the profits from the smartphone market, then it doesn't seem to be very attractive, even if INTC can take over that entire market. Do I misunderstand anything here? ::)

 

no it's not all the profits in smartphones. it's a tiny part of it. it's a IP royalty on a series of chips it sells to others. Samsung makes billions selling devices that contain ARM chips. Intel would be selling the actual chips, which have more profit potential than simply selling IP for a royalty.

 

If the phone chip is only $15-$25, how much can INTC make from that? $10 at most? For PC chips selling like $120-$200, I think INTC can make $60-$100 from it. Let's say the phone market expands and there are 3 phones to every PC, and INTC can take over the entire phone chip market, that would still only be a very small boost to INTC's bottom line, isn't it?

[Guest wellmont]

you are correct. it's not a very good replacement for the PC business. this is why intc sells at very low p/e ratio and even then may be overvalued. it's a very well understood issue for intc.

[ItsAValueTrap]

The chip in a smartphone ranges in cost based on how good it is.  Some are $50-70. 

 

A smartphone SOC often has the following features:

CPU

GPU

image processing for still/video camera

MPEG2/4 decode + encode

wireless

...and a bunch of other things

 

ARM licenses designs for the CPU.  They get a royalty from their CPU designs being integrated into the entire smartphone chip.  Sometimes a manufacturer will build their own CPU that works with the ARM instruction set- ARM will get a smaller royalty for that.

 

The overall smartphone SOC market is about $10B/year in revenue.

 

*ARM gets royalties from their designs being used for embedded applications (e.g. in iPods).  It was their bread and butter before the smartphone explosion.

[Ross812]

Are you guys worried about Intel?  (I am long Intel.)  Intel has a moat in the semiconductor business because its scale allows it to build cutting-edge fabs.  Its chips will have lower power consumption and/or better performance than everybody else.  If you look at Intel versus AMD... it's not a fair fight.  So Intel taking on all the smartphone SoC manufacturers is going to be scary.  If you look at reviews of Intel's smartphone chip in the Xolo (and there's another phone), performance and power consumption are middle of the pack right now. 

 

http://semiaccurate.com/2012/04/26/how-well-does-intels-new-phone-work-as-a-phone/

 

Eventually the software ecosystem will move away from having to do ARM emulation (emulation is slow)... this will increase the performance of Intel chips on Android phones.  And Intel will start moving its smartphone chips onto fabs that are a step ahead of the competition.  This will not be a fair fight.

 

*Historically...

Intel took huge market share away from the RISC instruction set chip manufacturers.  PowerPC, Cray, SGI, etc.

Intel didn't kill off Xilinx and Altera.  Their software design tools are critical and difficult to build.

Intel didn't kill Nvidia and ATI.  Graphics chips are hard to do.  Intel is doing ok with its integrated graphics on lower-end processors.

Intel has always enjoyed a huge advantage over AMD.

 

What is the upside for INTC?

I see that ARM is a hot stock and it is making the chips for the smartphones. But for 2012 its net income is only $165M. If that is all the profits from the smartphone market, then it doesn't seem to be very attractive, even if INTC can take over that entire market. Do I misunderstand anything here? ::)

 

no it's not all the profits in smartphones. it's a tiny part of it. it's a IP royalty on a series of chips it sells to others. Samsung makes billions selling devices that contain ARM chips. Intel would be selling the actual chips, which have more profit potential than simply selling IP for a royalty.

 

If the phone chip is only $15-$25, how much can INTC make from that? $10 at most? For PC chips selling like $120-$200, I think INTC can make $60-$100 from it. Let's say the phone market expands and there are 3 phones to every PC, and INTC can take over the entire phone chip market, that would still only be a very small boost to INTC's bottom line, isn't it?

 

Also keep in mind that if INTC can scale down their flagship processors or versions of their flagship cores they will be able to demand a premium price as no other company makes a chip that can compete with their core i series. High end smartphones are very expensive and the cost of a $60-$100 is only 10-15% of the total cost of a $600+ smartphone.

 

Not to mention the tablet and ultrabook market which will ultimately converge at some point in the future.

[ItsAValueTrap]

Also keep in mind that if INTC can scale down their flagship processors or versions of their flagship cores they will be able to demand a premium price as no other company makes a chip that can compete with their core i series. High end smartphones are very expensive and the cost of a $60-$100 is only 10-15% of the total cost of a $600+ smartphone.

Given the same die area, you can make tradeoffs between power consumption and performance.  That's why Intel has different lines with completely different architectures.  The CPU in their smartphone chips is completely different than the CPU in their high-end desktop/server chips.

 

Intel is also using their best fabs to make desktop/server chips.  They aren't using their best fabs for smartphone chips, so they are using older processes compared to some of the ARM-based competition.  Using essentially the same design on a better process improves power consumption.  However, it takes time to convert the design over as parts of it need to be re-engineered.  The smaller parts have different electrical characteristics that need to be taken into account.

 

Not to mention the tablet and ultrabook market which will ultimately converge at some point in the future.

 

I think that those markets will be a key battleground between ARM and Intel.  It's like the middle of the road between Intel's advantages and ARM's advantages (entrenched ecosystem, ARM has an open ecosystem where you can choose the best GPU and wireless, lower cost, etc.).

[Ross812]

Also keep in mind that if INTC can scale down their flagship processors or versions of their flagship cores they will be able to demand a premium price as no other company makes a chip that can compete with their core i series. High end smartphones are very expensive and the cost of a $60-$100 is only 10-15% of the total cost of a $600+ smartphone.

Given the same die area, you can make tradeoffs between power consumption and performance.  That's why Intel has different lines with completely different architectures.  The CPU in their smartphone chips is completely different than the CPU in their high-end desktop/server chips.

 

Intel is also using their best fabs to make desktop/server chips.  They aren't using their best fabs for smartphone chips, so they are using older processes compared to some of the ARM-based competition.  Using essentially the same design on a better process improves power consumption.  However, it takes time to convert the design over as parts of it need to be re-engineered.  The smaller parts have different electrical characteristics that need to be taken into account.

 

Not to mention the tablet and ultrabook market which will ultimately converge at some point in the future.

 

I think that those markets will be a key battleground between ARM and Intel.  It's like the middle of the road between Intel's advantages and ARM's advantages (entrenched ecosystem, ARM has an open ecosystem where you can choose the best GPU and wireless, lower cost, etc.).

 

Atom processors are making headway, but need LTE to become viable for US based smartphones. What I am waiting for is another 50% power reduction in Intel's flagship processors. i7 Haswell mobile processors are down to a 17 W TDP with their Y series down to 13W.  How many more tick-tock iterations are needed until their flagship series is the best choice for a tablet. How many more until they can be used in a phone? I think the atom is a stop gap measure to get the software running on x86 architecture and begin establishing intel with smartphone makers, but the real prize is going to come in 5 years, when an i3 processor can be dropped into an iPhone.

 

[ItsAValueTrap]

That's not going to happen.  Intel is going to have a high-performance architecture and a low power consumption architecture.  They already have different manufacturing processes for low power and high performance.  In manufacturing the chips, there are tradeoffs between power consumption and cost and performance I believe.

 

2- Actually I am thinking that Intel's smartphone/tablet strategy is not the greatest. 

 

a- Vertical integration or not?:  Intel is vertically integrated.  Historically, most of the businesses with this model have lost out (e.g. mainframes, Wang, etc.).  Apple is the exception.

Semiconductor manufacturing is a business that favours economies of scale (within your own niche at least- manufacturing memory is different than manufacturing CPUs).  More volume could make Intel more competitive.  Intel could be making ARM-based smartphone SOCs for companies like Nvidia.  ARM wouldn't care and Intel is already one of ARM's biggest customers.

 

b- Bad execution:  Intel doesn't offer a smartphone SOC with the wireless capabilities that cell phone manufacturers need.  As I understand it, there are various frequency bands used in each country.  So, companies like Qualcomm have many variations on their SOCs to accommodate all the wireless frequencies.

 

Other SOC design companies license the wireless component of their SOC.  Intel has dropped the ball in integrating LTE wireless onto its SOC.  Intel did buy Infineon's Wireless business... but so far it has failed to execute.  It won't have a LTE smartphone chip until early 2014.  So far Intel's approach of vertical integration and a somewhat-closed ecosystem hasn't been working out too well.

 

Intel could open up its ecosystem and make smartphone/tablet SOCs with an Atom CPU and a Nvidia GPU.  Or it can simply manufacture SOCs for Nvidia, which is what Nvidia's CEO suggested.  Last year Intel actually ended up with too much capacity in its fabs.  It should go after more business instead of trying to spread an x86 duopoly everywhere.

[Palantir]

Hmmm....do you guys think a dividend cut for Intel is likely as it goes through this transition? I think this opportunity is interesting....but I'm wondering if INTC would be pressed to preserve its cash flow for the time being.

[ItsAValueTrap]

Intel is massively profitable and has a hard time reinvesting that much cash.

[infinitee00]

Heck, I am confused by this thread. Did I miss the secret merger between INTC and MRVL ?  :D

[Ross812]

That's not going to happen.  Intel is going to have a high-performance architecture and a low power consumption architecture.  They already have different manufacturing processes for low power and high performance.  In manufacturing the chips, there are tradeoffs between power consumption and cost and performance I believe.

 

2- Actually I am thinking that Intel's smartphone/tablet strategy is not the greatest. 

 

a- Vertical integration or not?:  Intel is vertically integrated.  Historically, most of the businesses with this model have lost out (e.g. mainframes, Wang, etc.).  Apple is the exception.

Semiconductor manufacturing is a business that favours economies of scale (within your own niche at least- manufacturing memory is different than manufacturing CPUs).  More volume could make Intel more competitive.  Intel could be making ARM-based smartphone SOCs for companies like Nvidia.  ARM wouldn't care and Intel is already one of ARM's biggest customers.

 

b- Bad execution:  Intel doesn't offer a smartphone SOC with the wireless capabilities that cell phone manufacturers need.  As I understand it, there are various frequency bands used in each country.  So, companies like Qualcomm have many variations on their SOCs to accommodate all the wireless frequencies.

 

Other SOC design companies license the wireless component of their SOC.  Intel has dropped the ball in integrating LTE wireless onto its SOC.  Intel did buy Infineon's Wireless business... but so far it has failed to execute.  It won't have a LTE smartphone chip until early 2014.  So far Intel's approach of vertical integration and a somewhat-closed ecosystem hasn't been working out too well.

 

Intel could open up its ecosystem and make smartphone/tablet SOCs with an Atom CPU and a Nvidia GPU.  Or it can simply manufacture SOCs for Nvidia, which is what Nvidia's CEO suggested.  Last year Intel actually ended up with too much capacity in its fabs.  It should go after more business instead of trying to spread an x86 duopoly everywhere.

 

I agree with most of your points. We just see a different end game for Intel getting into smartphones and tablets. I think they started neglecting the atom because they wanted to push ultra books instead of net books a few generations ago. I think the popularity of tablets caught them off guard and they scrambled to resurrect their neglected low power atom to get into the new growth market. I expect them to have several iterations of the atom, but over time I expect power requirements for their flagship cpu's to drop to a point where they can be used in tablets and eventually even smartphones. Atoms will be in both smartphones and tablets in the next couple years. Atoms will start to be replaced by iX processors in tablets in a few years, and iX will start to replace atoms in phones in another few years. Tic-tock.

 

I agree that intel's execution has been terrible thus far. They are playing catch up in mobile and their wireless integration has been really slow as well. I like Intel because the business is a cash machine. They are still growing the high margin server business, they still make money of the contracting PC business, mobile is slowly being integrated, and there is a free call option if Intel decides to do contract foundry work for other chip manufacturers.

 

Heck, I am confused by this thread. Did I miss the secret merger between INTC and MRVL ?  :D

 

Yikes we have gotten off topic haven't we? Has anyone heard anything about the time frame for the judge making a decision about the crazy settlement against Marvel? 

[beerbaron]

Mai/june eas the court date they stated in their last conf call

[krazeenyc]

http://news.yahoo.com/judge-upholds-1-17-billion-patent-verdict-against-003937450.html

 

Judges Denies Marvell new trial and upholds award. Important to note -- she has not ruled yet on Marvell's request to lower the $$ amt.  I kind of expected her to uphold it. But I expect that in the Federal Circuit Appeals Court in Washington DC it'll be struck down/lowered.

 

[infinitee00]

Yes I agree. Not something I am worried about though.

[krazeenyc]

So... I was able to sell my MRVL position at $13 it was a pretty large position. Today, I waded into the LEAPS -- i bought 20 of the Jan 2015 10 strikes for $2.35 each. I feel like given the potential litigation risk, this is a pretty good price -- but I'm not experienced in dealing with options/leaps.

[krazeenyc]

According to Ben harrington KKR has taken a 5% stake in MRVL -- a prelude to a takeover.

[beerbaron]

Anybody has some explanation on the 7% drop today. I don't see any relevant news...

 

BeerBaron

[rainman]

all semi companies down huge today.

[fareastwarriors]

consolidation continues.

 

https://investor.marvell.com/news-releases/news-release-details/marvell-acquire-inphi-accelerating-growth-and-leadership-cloud

 

 

the transaction consideration will consist of $66 in cash and 2.323 shares of stock of the combined company for each Inphi share. Upon closing of the transaction, Marvell shareholders will own approximately 83% of the combined company and Inphi stockholders will own approximately 17% of the combined company.