Out of ARM's way, Brit chip juggernaut runs over analysts again Last year was a bumper 12 months for the chip artists at ARM: the company's revenues jumped in December above analysts' predictions as tablets and phones found their way into so many stockings, almost all of them packing ARM-designed processors. Bloomberg pegged fourth-quarter sales at £152.4m, based on averaging 18 different …
ARM do two things: they license their architecture to others to create their own custom CPU designs that are compatible with other ARM devices (e.g. Apple's A6 CPU, which is based on the A15 architecture, but much lower power), or they sell their own off the shelf designs for a fee per chip made.
"I'm guessing"
It actually says it in the article: ARM licenses its processor core designs to manufacturers and charges a royalty on each chip shipped using those designs.
I understand there are a few options for customers if you don't want a full-blown licence to customise your cores prior to fabrication.
C.
While it cheers me to see a British company doing so well like this, there is a part of me that worries how long someone bringing in hundreds of millions of dollars a year can hold off someone like Intel, bringing in tens of billions.
With that sort of disparity if resources, do you think ARM can stay out in front?
Intel wouldn't want to sell stuff so cheaply so price competition is difficult for them. Also, if they get into trouble, they have the nuclear option (if it isn't built in already) of perpetual licensing.
Intel might go in for predatory pricing just to drive ARM out if ARM starts cutting into desktop/server sales, but there is a reason why an atom costs about as much as a hard-disk, which may well have an ARM chip in addition to its precision electro-mechanical systems.
This is all assuming that intel can get their power consumption down to compete in ARM's market, and Samsung et al, think its wise to go with a cheaper intel product over a chip they can control.
Another point not quite brushed on here is the fact that other companies can make use of ARM as part of their own products. As an example nVidia combining its graphics with the ARM for a tegra processor. Or the new license deal AMD has made with ARM to utilize their chips. Who knows where that will lead, but quite possibly it'll be a sweet spot where AMD will challenge the ATOM.
Why does ARM need to compete, when they can get existing competitors to do it for them.
Absolutely. ARM is the Have It Your Way (tm) chip designer. You can licence entire SoC designs, or licence the ISA and build it in to your designs. Or licence anything else you like about ARM for that matter. You can then go to any chip-baker you like to build them for you be that TSMC, Global Foundries or Warburtons.
With Intel it's all or nothing, you can't licence x86, you buy entire CPUs from them or nothing at all. This inflexibility will be their undoing, not to mention their ludicrous pricing and addiction to 60%+ margins.
A number of people are pointing towards Intel's future designs with their potential to best ARM on performance and equal them with power consumption but it doesn't matter how good they are. Intel are late to the party, and in the the rapidly declining PC market it's hard to see how they can supply chips for $10 and still maintain their position as as a industry bellwether.
>Intel might go in for predatory pricing just to drive ARM
If Intel could compete on power - otherwise, you can offer a phone maker an Atom for free - it still can't use it.
Intels big problem is that if they did make cheap desktop/server chips to keep ARM out of the market - they would be undercutting their own chip business.
"This is all assuming that intel can get their power consumption down to compete in ARM's market, and Samsung et al, think its wise to go with a cheaper intel product over a chip they can control." They already have. Try comparing a Motorola Razr M (ARM processor) with a Motorola Razr i (identical, except with an Atom SoC). You won't find any significant difference in performance or battery life. In fact, I own a Razr i and I get more than 2 days of battery life with light usage, whereas with my old Samsung Galaxy S I would be happy to get more than 1. As far as anecdotal evidence can go, I think that shows that Intel has reached a level where they definitely can compete with ARM phone and tablet SoCs.
The thing is, ARM with all its customers (the ARM ecosystem) is much bigger than Intel in terms of R&D, chip fabs, design capacity, you name it. The ARM ecosystem gets to amortise its design costs across everyone participating (for which ARM takes a small cut). Intel must meet all its development costs from only its own sales.
"If Intel decides to licence core designs for a reasonable fee then it might be game over for ARM."
Intel stopped giving their chips numbers so they could even control the name you are allowed to call your compatible chip.
Does that tell you something of the level of control that Intel think is "appropriate" for their core product designs? They b**ched about having to license the 286 because the DoD did not like having a single IC supplier for core systems. That won't be happening again any time soon.
You will pry the IP for Intel's core processor designs out of the cold dead hands of their CEO and the members of the board.
"You will pry the IP for Intel's core processor designs out of the cold dead hands of their CEO and the members of the board"
And what, it's their IP to hang on to. Not for nothing but it ARM were half as smart as Intel they'd make quite a lot more than 160m while we're at it.
But the numbers are small. Sales in the quarter of 152M?
(iNTEL) Net income for the last three months of 2012 fell 27% to $2.47bn (£1.54bn), although the figure beat analysts' forecasts. Revenue fell 3% to $13.5bn.
See the difference? I'm all for ARM kicking ass - I would love a brit CPU and technology core to be world class in size and scale, but I am frequently surprised how small the ARM stuff is money wise.
"But the numbers are small"
Only when you compare apples and oranges.
Intel is a chip-baker, chip-set and motherboard maker (not to mention flash and other non-microprocessor activities).
Take ARM's numbers, then add the revenues and profits that its licensees make from ARM-based designs (Samsung, Apple, etc.). According to ARM:
Cortex Processors Licenses
Cortex-A 117
Cortex-R 28
Cortex-M 162
Classic ARM Processors Licenses
ARM11 Family 79
ARM9 Family 273
ARM7 Family 171
Source: http://www.arm.com/products/processors/licensees.php
Very, very different business model though.
If you added up the money brought in by Samsung's chip division (exynos is licensed from ARM, at the instruction set level IIRC), Qualcomm's chips (likewise snapdragon), nVidia's (tegra), Marvell, Broadcomm, Freescale, Ti etc etc you'd get a fuller picture.
ARM are a chip-design firm, not a full on design/manufacture/sales organisation.
What we need to make real money in the UK from ARM, is a UK-based implementor-manufacturer. Should we make 2013 the year of the British Mobile Phone?
"What we need to make real money in the UK from ARM, is a UK-based implementor-manufacturer. Should we make 2013 the year of the British Mobile Phone?"
Well, in a way we have. The Brcm2835 SoC has an Arm (designed in UK) and a GPU (designed in UK), and the two were bolted together in the UK.
Then used on the Raspberry Pi, also designed in the UK (and now mostly manufactured in the UK)
The chip is of course, like almost all other chips, manufactured in the Far East. I don't know of any chip fabs in the UK who could take it on.
Whether it makes a huge amount of money is another story.
And the business is niche. Lets say Qualcomm are building a SoC and need a core. They license one of ARM's designs because they know it works and it saves them doing it themselves.
Lets say their total licence and royalty payment to ARM for that chip's fabrication run ends up at £1 per chip.
Well, Qualcomm also have to design the rest of the SoC that the ARM core is mounted on. Let's say that's £2. They might have to do some integration work for it's final application, another £1. Then they have to go and build a fabrication facility to make the chip and ship it, another £1.
Total chip cost = £5. Let's say they stuff an extra 50p or £1 on in profit for themselves. So Qualcomm's revenue per chip will on paper be 5.5x - 6x that of ARM's, but they have a shedload of extra expenses, and couldn't just pour the profit into R&D on their own range of cores, because they also have to do R&D on the fabrication processes, broader system design, etc. In my example above Qualcomm are making maybe 9-16% profit on each chip sold.
ARM by contrast (assuming this is a well-worn chip design that has passed the break-even point) are pocketing their £1 as pure profit that they can plough into their next gen product. And if someone comes along wanting to use that design, they can just roll out the plans at little to no cost (beyond lawyers fees to sort out the contract). A huge % of their revenue is profit.
If someone wants Qualcomm to make that SoC again, they have to go and physically make another one, with a 9-16% return on investment.
*Disclaimer, I have absolutely no idea what ARM's per-chip royalties look like, nor what the ratio in terms of ARM licensing cost compared to the other elements like SoC design/integration or fabrication run up to. My 1:2:1:1 ratio is plucked from thin air to make a point about why ARM generate far less revenue than the people who use their designs but are nevertheless wildly successful.
Intel certainly has the cash to do so and with privately held companies the first thing you'd be likely to hear would be that the deal has been done. However, there is probably nothing that Intel could offer the owners of ARM that its licence-holders couldn't match or better. Even if Intel did win a probably ruinous bidding war it's almost certain that at Britain and the European Commission would raise the spectres of "national interest" (always a good one and if it works for yoghurt then it will certainly work for high-tech) or threat to competition.
Why would we want to get into fabrication? It's low yield profit wise, highly toxic environmentally, power consuming and there is vast competition from cheaper more entrenched companies in the US and Far East.
You'd be better off burning 3 billion pounds in a park and throwing a party where tickets are a tenner to watch 3 billion pounds to go up in smoke.
I think you might reasonably consider the huge international ARM sub-economy as a juggurnaut, but I don't think that word can be used to describe ARM holdings. Whilst ARM holdings are highly profitable, in terms of profit margin, they are a tiny (if technologically critical) part of the whole ARM sub-economy, let alone the whole micro-processor market.
Of course it's this very point, that the ARM IPRs are so cheap to license and produce that makes it so very difficult for Intel to compete. Short of Intel virtually giving away technology (with consequent issues about anti-competitive behaviour), it's difficult to see how the company would approach what it managed in the PC/server market. Even if they did, the returns would be tiny compared to what they were used to.
"Why doesn't Intel just buy ARM?"
Quite simply because the won't be a competition authority in the World that would allow such a thing due to very obvious monopoly issues. Of course the way is open for another company to buy ARM (Appled have been mentioned in the past), but that might prove tricky as any threat to the independence and customer neutrality of ARM could undermine the entire basis of its business as customers started exploring strategic alternatives.
A few years back my HSBC financial advisor told me my holding of £150 worth was "uneconomic to hold in my portfolio" and seemed dismissive of "this small company". I disagreed and moved away from HSBC private banking (having found almost all their advice was to buy turkeys). Just flogged those ARM for about £5k (should have taken my own advice an bought more way back...). Having sold I expect the price will rocket again although the p/e ratio is dire.
However my current broker agreed it was time to take the profit and said "what if Apple stopped using ARM?" so destroying all confidence in him!
After buying mine at just 42p about 6 years ago (after a ridiculous drop from £1.20 due to a single profit warning), I watched them go back up to £1.50, then sink back down and wobble around the pound mark for several years. I decided to sell at £1.02 in 2009, and thought I'd made a tidy profit, but since that day they haven't stopped going up since. They were £9.33 yesterday, if I hung on I could have been minted.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
