Intel processors traditionally are powerful and power hungry. I admit their effort of reducing power consumption. Yet, it is yet close to what ARM can deliver.
Long battery life is one of the critical requirement of IoT. I wonder what processor Intel is going to put out to tackle the market requirement. Well! Unless Intel isn't eyed on the small devices, rather on the hub to empower the small IoT devices.
IoT requires different blends of chips for veriety of applications, it requires working closely with the OEMs. I think only ARM has that kind of flexibility to blend. I don't how it works with Intel (usally it's unidirectional with Intel - OEMs have to use what Intel produces).
>> I think only ARM has that kind of flexibility to blend. I don't how it works with Intel
ARM seems to be a great technology company. I think it. The problem is that it has a really bad business model. It makes pennies while its customers make real money. Yet that model seems to be the reason why it is infectous in the ecosystem as it drives growth and get more companies to adopt its platform.
> ARM seems to be a great technology company. I think it. The problem is that it has a really bad business model. It makes pennies while its customers make real money. Yet that model seems to be the reason why it is infectous in the ecosystem as it drives growth and get more companies to adopt its platform.
Why do you think ARM has a bad business model? Most people would agree ARM has one of the best business models of the industry. Sell many billions of chips and receive a few percent of the price they are sold at. It's practically free money at low risk (eg. no expensive fabs), resulting in great margins. Have you actually looked at the financial statements?
If you think QC mostly uses IP from ARM then think again. QC is doing well precisely because they have very good IP themselves. They design their own CPUs which are competitive with ARM's, they have a good GPU design, and they are the number one in modems.
It seems you're just looking at numbers without understanding anything of the underlying business. That's fine if you use it for investing based on numbers, but don't use those numbers as proof as to which business is better. Higher numbers don't automatically mean better in business. Otherwise you could say the only valid business model is drilling a few holes in the ground and polluting the earth.
>> Otherwise you could say the only valid business model is drilling a few holes in the ground and polluting the earth.
Thanks - that is the summary. I am open to that Wilco because the last I checked, it is 100% legal and even the government gives them subsidies to make that "pollution" process very profitably to the investors.
Selling black gold may still be profitable, but you'd be crazy to invest into dinosaurs today. It's becoming very expensive to get that last carbon out of the ground. Regulations are getting stricter. Subsidies turn into taxes. If you don't feel the wind of change blowing then you'll miss the next big thing.
>> I admit their effort of reducing power consumption. Yet, it is yet close to what ARM can deliver.
I am yet to figure out why Intel will not buy ARM. Regulators may not like that. Ok but I am yet to see the Intel strategy in the low-power budget systems. They have these great papers, but when you benchmark their products with ARM on power, ARM does better. The IoT will be a test on the future of Intel. They have new leadership and can make it happen.
Why would Intel even want to buy ARM? Ironically Intel used to make ARM CPUs which were used in lots of phones but they sold it to Marvell. They strongly believe x86 is far superior in every way. Look at Atom for example, which they said was going to take over the mobile world 6 years ago. It wouldn't be possible to use the internet from ARM, they claimed, you always need x86 for performance and compatibility.
Today Intel's mobile phone market share is a paltry 0.2% due to purchasing "design wins" and giving millions of Atom CPUs away for free... The business unit responsible for Atom is running on a $0.5B loss per quarter as a result. Flogging a dead horse?
So reviving a 25 year old 486 is their best effort at getting back into microcontrollers (after letting the 386/486 microcontroller market die) - seriously, what are these guys smoking?!?
>> Today Intel's mobile phone market share is a paltry 0.2% due to purchasing "design wins" and giving millions of Atom CPUs away for free..
I still think that ARM needs to learn alot about business from Intel. It is not just about technology. You need to think of making money. ARM is not doing that right now. Their model is not very smart to allow partners to make all the money while using your IP
>> I disagree...ARM cores proliferate exponentially...once they dominate they can just increase licensing costs at will
Great point for those that will wait until then. But you never have to wait because no one owns this tech world. I am hoping they jack that $25B valuation to match Qualcomm's $132B. Waiting?? Good luck
I am noit sure ARM wil ever catch up to Qualcom valuation...somewhat different businesses...Qualcom is a wireless machine with some key IP that most of their competitors have to licence...in a high tech world there is no way to predict, perhaps tomorrow someone comes up with another neat idea for parallel processing that kills Arm or someone comes up with a revolutionary patent on wireless communication that will eventuall kill Qualcom ;-)...Kris
My point is now largely on the comparison of ARM and Qualcomm valuation. I am talking about why ARM may consider to get into making things. I know they have more valuable IPs than their current valuation because of their business model
> My point is now largely on the comparison of ARM and Qualcomm valuation. > I am talking about why ARM may consider to get into making things. I know they > have more valuable IPs than their current valuation because of their business model
If ARM started making their own SoCs, it would compete directly with all its customers. ARM has no modem IP and no experience selling chips. ARM's business model works so well because they are not really competing with anybody (apart maybe Imagination Technologies).
>> If ARM started making their own SoCs, it would compete directly with all its customers
Good luck to them. I know Google makes tablets and phones even though they own Android. That is not an excuse. We are looking for ways to jack up earnings and return money to shareholders. They can enjoy the cents they get per phone. Good luck to them.
Can you explain what exactly could ARM learn from Intel, apart from how to make uncompetitive mobile SoCs and huge losses?
I'm sure you already know Intel only makes money from having their own chip manufacturing plants and selling their x86 chips at hugely inflated prices (some at $7000) due to practically being a monopoly. Intel got rid of AMD as a competitor via illegal anti-competitive business practices for which they got fined multiple billions.
ARM simply uses a different business model - and it is working well for them.
>> ARM is already making a lot of money for investors. For example if you invested 10 years ago, you would be sitting on a nice profit of 666%. Compare that to Intel's return of -7%.
That is bad analysis. At the same age, within a decade, Intel returned 5678%. You need to benchmark your time. Also, it does not matter the time of investment, we are looking at total valuation. Look into the future and not the past.
ARM must learn how to make more money like Intel does.
Total valuation is not a useful measure for anything. What matters is growth potential into the future. And the facts point to ARM having a much higher chance to give good returns over the next 10 years than Intel. What Intel did 30 years ago is not relevant today.
What matters today is that Intel's profits are declining, fab costs are increasing fast, TSMC is catching up, Intel missed all opportunities in mobile phones (a miserly 0.2% marketshare after spending many billions over 6 years) and now appears to push a summer student project (Quark) as their IoT chip...
>> Total valuation is not a useful measure for anything. What matters is growth potential into the future.
No problem. I look forward to the future but I want relevance today. I agree with you and I am not defending Intel for anything. I am saying that today, it is has a higher valuation than ARM and when you go through its history, it was more profitably at the age ARM is today.
Good point @wilco1...I was always impressed with ARM and wanted to buy their stock seeing their bright futures...but silly me, I always thought the stock was too expensive...so in my mind ARM valuation was too high ;-)
>> Intel got rid of AMD as a competitor via illegal anti-competitive business practices for which they got fined multiple billions.
They call it the cost of doing business. Until SEC and U.S. law close companies and force them into bankruptcies, that will never change. Catch me, I pay fine. Out of every $10B, I pay fine of $10M, that is good I will say.
>> Important question is - Will Intel like this thin margin product? If they do not, how long this parterneship will last?
I think YES. When you try to compete, you look at market share. If Intel wants to beat ARM, it must forgo margins and focus on market share. That could mean taking losses or even having low margins in its business.
@goafrit: This will entail Intel to drift to new business model. Current Intel structure has very high overheads and this is satisfied with high margin business. When they move to low cost products, they need new set of people at low cost. Will Intel float new organization for this?
Reach of IoT is practically every where, obviously OEMs won't buy $100 chip for a $50 product. A week ago I saw an IoT chip offering at 70 cents. I don't know if the American companies can make money with extreemly low margins, but it's still a big business on the whole.
The best way to handle this is to build some soft IP that can blend for veriety of end applications and enjoy the royalties, obviously ARM can be big winner in this game.
I still belive Intel should come up with X86 soft IP's to be able to compete with ARM in IoT game.
Next level of IoT is handling data from billions of devises, which requires high end processors and networking products. These are low volume high margin products where Intel has obvious advantage but it seems ARM is catching up with A57 range of processor cores.
There is a big IoT opportunity ahead, one has come up with a good business model to win it.
Interesting observations about margins, especially if the "$50 product" you are talking about is a small mobile device rather than a network hub that processes data from large numbrs of small mobile devices.
Those small mobile devices need to move to a much lower price range if there are to be billions of them sold.
Most of the discussion is around Intel not being able to stomach lower margins, but in this space Microsoft is competing against Linux being given away for free! Several RTOS companies have recycled into a Linux value-add model, but in the meantime the tools available to roll your own kernel have gotten unbelievably better. In order to succeed here the Wintel guys would have to use the same model that IBM did in the PC days - Take all of the company rules, stuff them in the trash can, and start over.
It's not always about profit margins and market shares. It's about bringing the technology to the next level, creating new experiences, accelerating the development of new markets. And these are interlinked together. After all, the increasing number of mobile devices and IoT nodes goes with the need for more server capacity.
>> A week ago I saw an IoT chip offering at 70 cents.
That is actually good price. There are chips that cost 5 cent and in the commoditized consumer market, if you get anything more than 50 cent, the stars are aligning. Unless you are supplying to Apple that wants more tighter specs that others, most OEMs will take it
It is a testing time for the Giants of the PC market!! I guess, the situation is not as easy as it was for Intel to excel in the processor market for personal computers. It is late for joining the party as many others are out there already. But I am hopeful to see some efforts being taken. Anybody could try out Intel's Galileo (Arduino compatible development board), which has a Quark X1000 SoC on it? How does it compare with the other popular ones?
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...